Differentiated Understanding

Epizode

• China’s internet ecosystem, manufacturing base, batteries, EVs, robotics, and semiconductor becoming an AI-enabled industrial system 01.06.2026 50min

  In this episode of Differentiated Understanding, I spoke with THE TP Huang, an independent China tech analyst known for his work on fintech, EVs, batteries, AI, semiconductors, and the broader China industrial ecosystem.The conversation traces China’s technology evolution from the early internet era to the present. TP argues that China’s internet ecosystem was shaped by a combination of censorship, protectionism, local engineering talent, and intense competition. That created powerful domestic champions such as Tencent, Alibaba, Huawei, Baidu, and ByteDance, which later became the foundation for super apps, payments, e-commerce, cloud infrastructure, and AI.The discussion then moves into China’s shift from software and internet platforms into hard tech: EVs, batteries, robotics, drones, semiconductor supply chains, and AI-enabled industrial systems. TP emphasizes that China’s technology companies are unusually willing to enter each other’s markets. Xiaomi moved from phones to chips and EVs; Huawei moved from telecom to semiconductors, AI chips, and autos; BYD moved from batteries to cars, solar, transit, chips, and potentially robotics.A major theme of the episode is that China’s AI story is not only about large language models. It is also about the physical stack around AI: batteries, sensors, motors, chips, power systems, critical minerals, factories, and real-world deployment. TP argues that this manufacturing and supply-chain density may become a major advantage in embodied AI and robotics, especially as real-world robot data becomes more valuable.Follow TP Huang here on X or Substack here To find the previous episodes of Differentiated Understanding, see here.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.Season two will host a series of guests from early-stage investing, as well as builders, researchers, founders, and product managers. For more information on the podcast series, see here.Chapters00:00 The Evolution of China’s Tech Landscape05:58 China’s Internet and Tech Sovereignty09:01 Investment Trends in China’s Tech Sector11:04 The Role of Government in AI Development20:00 The Intersection of EVs and Robotics26:07 China’s Competitive Edge in EVs and Robotics36:18 Global Strategies of Chinese EV Companies42:31 Advancements in AI and Robotics in China48:31 China’s Digital Infrastructure and AI Adoption57:38 Underappreciated Developments in China’s Tech Landscape01:00:00 Non-Consensus Views on China’s Economic HealthAI Generated Transcript (for reference only)Grace Shao (00:00)Hello everyone, welcome back to another episode of Differentiated Understanding. I am your host, Grace Shao. As many of you know, I also write the newsletter AI Proem, which is AI PROEM on Substack, so do give that a follow.Today we’re doing something special. We’re doing an audio-only version. I’m joined by TP Huang, an independent China tech analyst who writes about the intersection of fintech, EVs, batteries, AI, and broader China industrial policy. He has built a large following on X and Substack by combining data, supply-chain detail, and geopolitics to explain where China tech is actually heading.In this conversation, I want to use TP’s lens to understand the bigger China tech landscape: how China moved from internet platforms and payments into EVs, batteries, robotics, and now AI-enabled industrial systems. And since he quite literally said, “I can talk about anything China tech,” when I reached out, this conversation may follow the themes that I prepared, or really just go anywhere it naturally takes us. Very excited to have him on. Welcome, TP.Grace Shao (00:02)Hi, TP. Thank you so much for joining us today. I just did your intro before talking to you. And I told everyone that when I emailed you and reached out, I said, here are some topics I want to talk about. Is that okay? And you quite literally said, “We can talk about anything China tech.” So the conversation today could cover quite a lot of bases. I’m so excited to hear from you and have you kind of dissect a lot of your knowledge for us. And, you know, I’ve been a big fan of following your Twitter, your X, for a long time. Anyhow, thank you so much for joining us today.TP (00:33)I’m just really glad to be here, Grace.Grace Shao (00:37)Yeah. So you’re a mysterious man. Give us some color on your background and why you are so knowledgeable about China’s tech ecosystem, because you’ve really been covering everything from robotics to LLMs to the internet era. You cover them all, including hardware and chips and everything.TP (00:56)Yeah, so it’s kind of interesting that my actual background is not very technical in that area because I’ve been working mostly in the finance sector, or fintech sector slash crypto, for most of my working life. And I did spend a year recently working in an AI firm, so that was something different. But now I’m back to doing more crypto kind of stuff. So my background, I guess now, is a lot more AI-related.But a lot of the interest I had back in the day was in the renewable space and climate change and things like that. So that really got me started following solar panels, wind turbines, and then EVs. I first read about BYD back in 2008, like a lot of other people. And then as EVs were really taking off in China, that’s when I thought, okay, I really need to understand the full tech stack behind it. So that kind of got me into the entire battery supply chain, a lot of the upstream stuff, and then chips.The chips part became such a big deal because of AI. So then we had the October surprise back in 2022. That’s when I decided, okay, I’m really going to try to understand how the semiconductor manufacturing part of it works also. And thankfully, I was able to be connected to a lot of people. That allowed me to really understand a lot more.So I don’t profess to be an industry insider or anything like that. I’m just talking to other people who are working in the industry for some knowledge and writing about it. And then with AI, I actually worked on my own, no, not on my own. I worked with an AI startup, and one of the projects we did was actually for an AI toy. So I had experience running what I would consider to be AI robotics efforts. So I have a lot of real-time experience with embodied AI and also just using large language models. That’s kind of how I got into all this stuff in the first place.Grace Shao (03:26)It’s really cool because you have experience across the whole array. One personal question is: what drives you to really continue writing? Because you do write prolifically on Twitter. You have these hot takes, you put things together, and I think you’re quite widely followed by anyone who covers China tech. So what makes you want to share things publicly?TP (03:49)Yeah, I guess it’s more like a personality kind of thing, where I really just enjoy writing. And I think there’s something missing in the information space about what is going on in China.Last summer I was in China for a month, and I plan to be in China again for a month this summer, and I just saw a lot of really cool stuff. I think it’s good for the world as a whole to understand what’s going on in China, for Americans and for all Westerners to understand what’s going on in China, so that we are better informed in understanding how people can work with China and what kind of things people who want to compete against China need to know. But as a whole, I think it’s better to get proper information out there.And because China is a different language, and most people in China post in their own internet ecosystem on Weibo or WeChat, people don’t really read this stuff. So they get their sources from very bad sources on the English internet. A lot of them are just missing the nuance of what’s actually going on inside China. So because there is this vacuum, I just felt I’m obligated to actually do something about it, to help everyone understand better.Grace Shao (05:31)That’s awesome. It’s part of why I write AI Proem too. Well, okay, let’s get into the real stuff today. You’ve been following China’s tech for a while, like you said. Help us understand, just with the sentiment shift, how you view the early internet era to today’s success in hard tech and AI. What really has propelled China’s success in the tech sector in the last 10 to 20 years?TP (05:58)Yeah, so I think if we look back on things, China made a pretty big bet on developing its tech sovereignty back in the early 2000s and 2010s. It put a lot of policy in there under censorship reasons. It said, we’re blocking, we don’t want Google or whoever wants to enter China to actually censor the search results so that it fits our local law. And then what actually ended up happening was it became more of a protectionism kind of thing. So China was protecting the local tech champions at the same time that it was pouring a lot of money into these firms.So it allowed firms like Tencent and obviously Huawei and Alibaba to grow up. Later on, China also developed ByteDance. And if you look at how things are around the world, most countries, most leading Western countries that could have possibly developed their own tech ecosystem, like European countries or Japan, didn’t do it. The only other country that has a pretty robust local tech ecosystem or tech champion is Korea with Naver.And if you go to Korea, you notice that if you’re using Google Maps, it’s almost unusable. You kind of have to use Naver. So I think there’s a clear correlation between blocking US tech and some level of protectionism to having a local tech ecosystem being developed. And obviously it requires good local engineers also, so that they can take advantage of that. But China had all the ingredients for it.So even though it started maybe a decade after the US in developing this ecosystem, it was able to develop it because it didn’t have to face this immense competition from the US right away.And I think also there’s a lot of, you know, we talk about involution in China. I think there were stories of how when Uber tried to enter the Chinese market, because they had to face all these local Chinese companies that were working under 996-type hours, they were eventually pushed out of the market. So I think those are really the interesting parts of how the China tech scene developed in the 2010s.Grace Shao (09:01)Does that kind of feed into what we’re seeing now? Because right now it’s a completely different set of technology, yet in many ways it is building off the digital infrastructure that we just talked about, that got built out in the last 10 years or so.TP (09:17)Yeah. I think as a whole, if you go to China, even the internet ecosystem works entirely differently from America. In America, for the longest time, we had a search-oriented internet. You use Google, you use a lot of Google products, or you use social media. Whereas in China, because Baidu was never that great, people kind of advanced right away toward these mega apps like WeChat and Alipay.And as part of the movement on these fronts, you have these giant ecosystems developing where they not only have their own super apps, they also have their own e-commerce networks, their own payment systems, and they all got enough resources to eventually build their own cloud infrastructure and now develop into the AI world. So some of the biggest players in China when it comes to AI are the usual tech giants like Alibaba and ByteDance.Grace Shao (10:35)Yeah. So okay, let’s move on from that, from that big holistic overview of China’s internet space and tech sector. So much of the investor focus right now is still through the old internet platforms, like we mentioned, because of the natural progression of how they also become the major players in AI.But what kind of breakthroughs and capital moved from apps and payments into EVs, batteries, robotics, AI, and hardware? Are we seeing that these hyperscalers or big tech companies are also the major players in these other technologies that we’re talking about? Are they the main investors and backers, or is that a completely different ecosystem?TP (11:17)Yeah, China is kind of interesting to me in that a lot of the players are so uber-competitive that they are willing to get into other people’s spaces. So we saw Xiaomi move from the phone into developing their own pretty advanced AI team. They have their own chip design, and most notably they have their own EV division, which is doing really well.We saw Huawei start off in telecom and then move into the entire semiconductor ecosystem, and also their AI chips, and also into the auto division. We saw BYD start off as this battery company, and then it got into all these areas. It got into cars, it got into solar panels, it got into public transit, it got into the chipmaking side of things, and now it’s also looking to get into robotics with humanoid robots.Whereas you don’t really see that as much in America, where it’s mostly a typical thing I used to listen to on Wall Street, this entire idea of capacity discipline. Which is basically: how do we reduce competition so that we can get a higher margin? Whereas the Chinese marketplace seems to be one where everyone’s trying to squeeze in at the same time and just fight it out until whoever has the best cost controls ends up winning.From that point of view, I think this is why for some time people saw that the Chinese stock market hadn’t been growing as much as the US stock market, because there’s just so much competition inside China. So a lot of the funding for these efforts inside China actually had to be backed by the government, these big funds and things like that. And also these things, they are willing to put money into areas of lower initial returns.A lot of the car factories, maybe they’re not the best investment if you’re looking for a 100% return. Maybe it’s not the best for that. But because it provides local jobs and things like that, the government is willing to put some money into it. And we saw that right now with semiconductors also, and also the data center build-outs. So that is how, over time, the entire Chinese manufacturing ecosystem kind of got built out.America is trying to do a little bit of that right now with AI data centers and trying to do that with the tariff wars. But fundamentally, the market in the US is about squeezing out competition and lowering capacity in order to charge more. Whereas the Chinese system is about how to scale up production and lower the cost of production in order to have higher margins. So it kind of works differently.Grace Shao (14:43)Yeah. So on top of government help and actually putting money into sectors that often have lower initial returns, sectors that are not so sexy in the beginning, let’s talk about DeepSeek.I think it’s been interesting because we know DeepSeek and many of the other Chinese labs weren’t getting a lot of capital until maybe 2022 or 2023. However, now they’re obviously being pushed front and center as the main economic drivers. Not only are they being looked at as very sexy investments from the private side, but the government funds are also looking to put cash behind this.How do you view the relationship between government policy, government mandate, and the AI labs in China? That’s part one of the question. Part two is, if DeepSeek and a lot of these Chinese labs permanently price their models at, say, one-thirtieth of the American labs’ prices, what’s the thinking on that? And what’s the sustainable business model for them looking forward?TP (15:44)Yeah, so I think it took a while for China to really catch on to this entire large language model thing, because a lot of the Chinese AI, when I looked at it back in the early 2020s, was aimed at embodied AI. So in terms of smart manufacturing, how to improve the grids, drones, robotics, and also EVs, things like that.Whereas a lot of the US funding for AI was, I guess, kind of abstract. You want to develop the best models, and then we will find the use cases for them. But once it took off, I think there was kind of a light-bulb switch inside the Chinese sector that we can’t just let this go, we have to catch up. The way Chinese people think about things is like, we have to get in on these opportunities.So in the beginning, with Chinese large language model development, I think it was mostly the big tech companies like Baidu that were kind of leading the efforts. But over time, more recently, I think you find that it’s the startups that have done some unique research that have done the best, like DeepSeek, obviously Kimi, and Z.AI.And obviously some of the big tech companies are still quite successful, like ByteDance. They have a very good AI product. And Alibaba, with the Qwen stuff, is also very well developed. But you do see that the Chinese government, ever since the DeepSeek moment, has been investing more in funding to make sure that the domestic AI startups are able to get the funding they need to compete.In the most recent example, DeepSeek, they actually got paired up with Huawei, or maybe they came together somehow. But you can see just in the V4 release recently that there was a lot of integration work between Huawei and DeepSeek. The DeepSeek models are deeply integrated, so that you can use the Ascend chips from Huawei to better train and run the models. And this is part of China’s overall strategy of being self-sufficient in both the hardware and software side of things for AI.So even though it’s probably easier to just buy NVIDIA chips, the risk of getting cut off by the US government is pretty high. So it’s in China’s long-term interest to have its own ecosystem across the board.No other country has that. China has not only the chips and the software, but also the entire AI data center build-out ecosystem. There has been a lot of investment or money put into AI build-out-related stocks recently, like optical modules, optical transceiver suppliers, fiber cable suppliers, PCBs, power chips, and things like that.So there is a lot of investment across China, not just in the software part of it, but also in the hardware integration part of it. And at the end of it, it’s all supported by the Chinese government in some way because they want to make sure that they have a domestic supply chain, so they can’t just get cut off at any point.Grace Shao (20:42)So you’re basically in the camp of what Jensen was saying: export controls are not working. In effect, they are cutting American suppliers or vendors out of China, and in that case, actually pushing China to become more and more self-sufficient.TP (20:57)Yeah. I mean, for a long time there, Jensen and the good people behind SMCI were trying to get as many NVIDIA chips to China through backdoors, or through Asian and Southeast Asian data centers, as they could, right? So that the Chinese AI suppliers remain hooked onto the NVIDIA ecosystem. But you can see that by sometime late last year, the Chinese government was actively blocking these things from happening because they really wanted the domestic AI players to use the local ecosystem.Grace Shao (21:39)But is it actually being replaced right now? Or do you think in the short term, medium term, long term kind of thing? The long-term strategy is self-sufficiency. Short term, it doesn’t seem like it’s realistic yet, right?TP (21:51)Yeah, so this is the interesting part. For much of 2023 and 2024, what the Chinese players were doing was that a lot of them were importing the permitted versions, like H800 and H20s from NVIDIA, through official channels.And then there was a lot of smuggling of chips into China at the same time, and the Chinese government was allowing this. So whenever they were building AI data centers, they would have the data centers that use domestic chips and ones that don’t use domestic chips.So what would happen is, let’s say Alibaba was looking to access NVIDIA compute and it doesn’t want to get sanctioned by the US government. So what it would do is, it buys some NVIDIA H20s, puts them in a data center, and also leases compute that runs on NVIDIA from one of the state-built or local government-built data centers that smuggled in chips, because it didn’t want to get in trouble by buying them if it’s not allowed to.Another thing that these firms started doing that’s entirely illegal, again, is actually just setting up companies offshore that would buy these NVIDIA chips and then build data centers in the rest of Asia, places like Japan, Thailand, and Malaysia. And then they would lease the compute for these NVIDIA chips from these data centers.And that’s still going on right now. The Chinese government is allowing that because domestic firms like ByteDance would just say to the Chinese government, we need this ability to use American chips in order to not be left behind. Because if you talk to the AI developers in China, they don’t enjoy using Ascend libraries for training. They don’t mind using them to run inference, but for training, they still prefer to use NVIDIA chips.So there is an effort right now to also get the training part of it up to par. And that’s kind of what the DeepSeek work with the Huawei team in recent months has been about. It’s kind of interesting to see how much better the integration has made the Ascend chips run training and inference on the DeepSeek models.There has also recently been a Qwen model called 3.7 that came out. And they also released their own AI chip called Chengwu MA90.TP (25:10)And part of the interesting thing about that is not only did Alibaba have the self-designed chip, because it was designed internally, it used its own internal AI models to write the kernels for the chip. And it had some really good results. I think going forward, a lot more of these domestic chips will actually be able to at least do part of the training also.Grace Shao (25:39)That’s really interesting. And it actually echoes some of the stuff I’ve heard on the ground as well. So like I said in the beginning of our conversation, I don’t want today’s conversation to only focus on China’s LLM and model space. I want to double-click on something you mentioned at the very beginning of this answer. You said China actually started with its capital focus and technological focus on EVs and embodied AI.What’s interesting is that that side of things didn’t really pick up in the US or in the West, per se, until more recently. So did the EVs come first, or did robotics come first? Or did they kind of converge and come at the same time, and there’s synergy there?TP (26:23)Yeah, so when it comes to the EV and robotics story, I tend to think of it as something that started because China was doing all the manufacturing of consumer electronics. And that’s how it was able to then develop these OEMs in the smartphone space, like Xiaomi, Huawei, Vivo, Oppo, and Honor. Basically, they developed this entire workforce inside China that was very good at dealing with supply chains and also integrating things together and doing manufacturing.I personally had an experience with this about a year ago, where we were trying to make this AI toy, and I got on a call with a Chinese factory. I won’t say which one. But basically, about five minutes into it, I realized America was in trouble because we had all these American engineers who are decently smart people. And the sales lady at the Chinese factory just knew way more about how hardware works and should work than any of us did.It was a very humbling experience just to see how there’s a lot of process knowledge involved in this. There’s a lot of experience involved in this stuff, right? And my cousin actually works in Shenzhen.Grace Shao (27:46)It was like learning from experience instead of PhDs, right?TP (28:03)They developed their own automated device that tests blood samples to see what kind of disease you might have, something like that. And what I realized talking to him was that this entire supply chain in China around Shenzhen or around Hangzhou is very deep.Because of that, a lot of the modern tech that we see with embodied AI comes from this basic understanding of supply chain, software-hardware integration, and also electrical platforms. What are the commonalities between drones, robotics, cars, and EVs, right?First, you need to have this battery underneath. You need to have electrical platforms. You need to have PCBs. You need to have cooling systems involved. You need to have control chips. You need to have power management chips. You need to have main control chips for the actual device. You need to have AI chips.All this stuff, in the beginning, Chinese suppliers were sourcing from abroad. Over time, due to export controls, they started doing this domestic substitution. They’re still the biggest importer of chips globally, but a lot of that stuff is coming in-house now.So if you do a teardown of a DJI drone, you’ll probably find memory chips from CXMT and YMTC. You’ll probably find CMOS chips for the camera modules from maybe OmniVision or something like that. And the battery is obviously going to be domestic. And all the stuff that we saw with drones and with EVs, we’re now seeing with humanoid robots and other kinds of robots, because at the end of it, a lot of the basic concept is very similar.You need to have some kind of a brain for the embodied AI machinery. And then it needs to have some kind of battery source to actually do the functionalities. And then it needs to move using some kind of motors, and then it needs to be able to absorb information from its surroundings with these sensors.That is why China has such a large ecosystem, because it has a good upstream supplier network and a lot of people working on this stuff. Whereas if you come to America, there’s just not a lot of that talent around.So if you want to develop an AI robot, you have to do everything in-house and figure it out. Because if you can imagine, if you don’t develop in-house and you contact a supplier in China, you can’t really iterate things quickly because you’re working with someone over there who doesn’t speak English and also doesn’t work the same hours you do. So the turnaround time is just much slower.Whereas if you have an idea in China for an AI robot that you want to build and sell to the market, you can get it produced in a month. That would be crazy for any kind of AI startup in America to do.Grace Shao (31:59)Yeah. In fact, I think there are a lot of robotics companies right now with founders who are literally tweeting about this thing: we must move to Shenzhen. Or I know of companies that actually get their hardware completely end-to-end, basically buying from OEMs from Shenzhen and slapping on a tag elsewhere.But I want to ask, why did China ultimately come out on top in EVs? Because from what you just mentioned, technically wouldn’t countries like South Korea have an edge? They have car manufacturers, they have chips, they have memory chips, especially when you just talked about brains. It’s not like the brains that we’re talking about right now are AI brains yet.So what made China actually come out on top with EVs and robots? Was it again this narrative around government push, because the country needs clean air? Was it because of innovation? Was it because of renewables and everything coming together? How do we understand this?TP (33:01)Well, I think Korea itself is actually a country with a lot of industrial policy also. So I wouldn’t necessarily say that the Koreans were less aggressive about government support than the Chinese were.I would say that if you look at just the human capital side of things, we’re looking at a magnitude difference in the number of engineers coming out of South Korea and China. So that’s something not easily made up.If you have 10,000 battery engineers from China every year, and let’s say you have 1,000 from Korea, the 10,000 are going to crush the 1,000 over time. And you can kind of see that. Back in the late 2010s, the Koreans were ahead of China in battery technology. But because Chinese industries were moving so fast and the supply chain was moving so fast, China has been ahead of Korean battery makers for several years now. And the gap is only expanding as we move toward more advanced solid-state batteries, or lower-cost sodium-ion batteries.Batteries are such an important part of the modern electrical transition that it’s kind of mind-boggling that China controls so much of the entire ecosystem. People keep talking about TSMC, or Taiwan having some percentage of manufacturing for chips, which by the way is not true. But Taiwan only has a small part of the entire ecosystem. Korea only has a small part of the semiconductor ecosystem, right? America has a huge percentage of the semiconductor ecosystem.But if you look at things like rare earths, critical minerals, and batteries, China actually probably controls 80% to 90% of these ecosystems. So even the Korean battery makers rely on the Chinese supply chain for a lot of their inputs now. And there’s just no way to get around it because the Chinese process knowledge, cost advantage, and engineering advantage are very hard for a smaller country like Korea to overcome.Grace Shao (35:47)Interesting. Yeah. So how should we understand these companies’ international strategies? Because I think you’ve written about it before. Like you said, they are major exporters. How do the battery companies and EV companies position themselves globally? Are they quite aggressive? Are they suppliers along the supply chain? Are they building up consumer brands? How do we understand that?TP (36:19)Well, it’s different with different people. I think because the domestic market is so aggressive and so competitive, companies like BYD had to go abroad to get higher margins on their products. That’s kind of forced a strategy where they’ve aggressively expanded. Things especially picked up in the past few months because of the Iran war, where there’s also a lot of demand for these EV products abroad.And as a result of that, it helps what I call China Inc. As you see more of these high-tech EVs abroad, as you see more of these DJI drones and Chinese AI models abroad, there is a generally higher view of Chinese products now from much of the Global South. And as a result of that, Chinese firms are also having greater success selling their products.I think one of the interesting things recently is just to see how much the Chinese automakers’ market share in Europe has already surpassed the Koreans and is catching up to the Japanese. Just looking at that, it gives me the impression that the Chinese automakers, and just China Inc. as a whole, have gained a reputation for quality in a very short period of time. And you can only do that if the automakers themselves are making a real effort to build their brands and promote their products in these markets.And I think they’re getting paid off because my guess is that BYD’s automotive sales have much higher margins on stuff sold outside China than inside China.Grace Shao (38:43)I see. So it’s still like a pricing strategy, or winning on pricing, you’re saying.TP (38:50)I think in China it’s more of a pricing strategy, but abroad you see them actually marking things pretty high. So maybe there is a pricing part of it, but if you listen to Stella Li, Executive Vice President of BYD and President of BYD Americas, talk about the new models that they launched in Europe, they’re very much trying to frame it as a luxury brand, with the Denza model brands.She would say that this is technology that does not have any competitor or equal in Europe. We’re just way ahead of the Europeans here. We’re going to build the fastest charging network that you’ve ever seen. You can charge your car in five minutes, for example.It’s kind of interesting because BYD can sell its cars at a much higher price outside China than inside China. Inside China, it might have to sell its cars at a discount to Tesla cars. Outside China, it might sell them at the same price as a Tesla car. So yeah, I find that interesting.Grace Shao (40:04)That’s very interesting. And I’m kind of playing devil’s advocate purposely. Anecdotally, I’ve obviously been in a lot of BYD cars when traveling in China. They are actually really, really sleekly designed. And like you said, in China, for some reason, they’re positioned more as not a luxury car at all.But even in Hong Kong, I’m seeing more and more Zeekr cars and BYD cars taking the roads, and they’re definitely replacing previous Audi and Volvo owners. It’s very interesting that that’s the trend. Outside of mainland China, the reputation of these Chinese EVs is almost more premium than they are in China.TP (40:49)Yeah. And one of the reasons BYD wanted to do well in Japan and Germany was that it thought that once it started selling well in Japan and Germany and got approved by those automotive nations, people inside China, especially suburbanites in Shanghai, would then accept BYD as quality products. It is kind of interesting that a lot of times the Chinese can’t really accept that we have quality products unless it’s also being accepted abroad. It is kind of interesting how that works.Grace Shao (41:21)Psychology, I guess.TP (41:31)Yeah.Grace Shao (41:37)I guess it’s a little bit of a psychological play on this as well. I do like your framing on China Inc. And I think recently we’ve seen that even in the consumer space. It was so interesting that Luckin Coffee bought Blue Bottle coffee, and you’re getting more and more of these kinds of purchases, like SHEIN buying out Everlane, etc.But I want to bring it back. I want to bring it back to AI.You said earlier that China’s mastery of hardware manufacturing has given it an edge in scaling humanoid and service robots. But how do we understand where we are with world models and the actual next stage of embodied AI and physical AI right now? Because like what we just discussed, China’s manufacturers are very experienced in building out the robots, drones, and various forms of robot mechanics. But where are we with actually injecting that with AI?TP (43:02)Yeah, so I’ve been in touch with the guys behind the China Research Collective, and they are actually inside China, so I’ve had some discussions with them about this. They’re telling me that because China has this hyper-competitive local market for jobs, a lot of young people are having trouble getting the jobs they wanted. So they’re willing to help these AI companies collect data on doing things to help these world models.It’s kind of interesting because you need a certain amount of data so that the robots can simulate human movement and then do the tasks. But at a certain point, if you have a child, you know that it takes them a long time to be able to walk around and then run, because they need to first feel and touch everything and learn everything over a year or so. During this time, their muscles develop and their muscle memory develops so that at a certain point they no longer need to think about how they walk. They can just walk. They no longer need to think about what they can or cannot eat, because they already put that stuff in their mouths to test it out.Longer term, I think once you have enough robots in China, they will just be able to improve exponentially in their capabilities because they will be able to fast-track all this, what I call reinforcement learning in the real world. If you try grabbing an object a million times, eventually you’ll figure out the best way to grab it. And once a robot learns how to grab it, that gets shared amongst all the robots of that family.So I think as you see the Chinese robotics rollout speed up, this is when you see this decisive edge in the world models. We already saw this with drones, right? The Chinese drones are just so much better at moving around and doing stuff because they had so much more data than anyone else.We’re seeing it now in EVs, where the Chinese self-driving cars are really good because they’ve had a lot of data out there, where people are just using autonomous features to do all the work. And you’re seeing that BYD today is having this entire unveiling where it’s talking about its path toward L3 and L4 autonomous driving.The more data it has, the better it’s going to get. That data becomes an advantage going forward. In the future, whoever has the most robots out there in the real world, and has all that data, can then train their robots faster. That’s why it’s kind of a big deal right now that BYD says it’s going to have 20,000 robots in its factories this year, because then it has all this data on using robots in a factory setting. That’s going to improve the performance of the world models by leaps and bounds.Grace Shao (46:48)Mm-hmm. Because the biggest bottleneck right now is just not having enough 3D data. And collecting that kind of 3D data is extremely challenging without, like you said, real, actual physical deployment. That’s fascinating.TP (47:10)Yeah. I also want to point out one other big difference between the Chinese players and the foreign players outside China, which is that China has this entire critical mineral supply chain. That is foundational to the rare earth magnets, for example, needed for the different robots and EVs, and for the motors, and also the materials needed to build the humanoid robots themselves, like magnesium. It produces about 80% of the world’s magnesium, and magnesium alloy is considered to be the main material that you want to use for humanoid robots.Grace Shao (47:57)I want to tie it back to what we also talked about earlier. Does the very strong digital infrastructure layer, just from fintech, IoT, and 5G, now contribute to China’s very quick adoption and diffusion of AI in the real economy? And how do you view this kind of positive cycle versus in other economies, where sometimes the digital infrastructure maybe just isn’t there yet and seems to need time to build up as well?TP (48:31)Yeah, I actually think this is one area where America might have a leg up on China, because the American big tech companies tend to also be the biggest cloud service providers. The Chinese ones are a little smaller. So right now, you only see the competition between the US and China because they’re the only two countries that have this data center and AI infrastructure advantage over the rest of the world.The biggest players in China, like ByteDance with their entire AI cloud infrastructure and their entire AI app ecosystem, are also the ones that are able to deploy their apps globally the fastest. In America, ChatGPT/OpenAI has this commanding position not because it has an ecosystem, but just because it was the first to do it. It had a first-mover advantage.But if you look at the players outside of ChatGPT, it’s Google slash Gemini that probably has the largest market share, because it has this big data center hardware, this AI infrastructure advantage over other players. And also it has this app system that people can use the AI features in.In China right now, personally, I don’t get to use the AI apps in China all that much, but I do have a Chinese phone, and I use ByteDance’s Doubao app, and it’s really good. So that has allowed ByteDance to have the best video generation model out there, called Seedance 2.0.Grace Shao (50:29)Mm-hmm. And they really leverage and lean into their data advantage as well. Obviously, if you own TikTok and Douyin, you have the most amount of video data in the world.TP (50:43)And not just that, they also have CapCut.Grace Shao (50:58)They do, which is the editing tool. I actually use it to edit our videos here on AI Proem. It’s great. I kind of want to wrap it up soon.I want to ask you a forward-looking question. If we connect the dots from your fintech days covering the digital economy to what we just touched on, EVs, robotics, hardware, everything, where do you think China’s digital economy goes over the next five to 10 years? What are the biggest bottlenecks? Will that look very different from the rest of the world? Or do you think the evolution of technology will be organic and go in the same direction, no matter your geographical location or your domestic strengths or weaknesses?TP (51:30)Yeah, so I will first talk about where I think they can possibly see the most improvement, and that will be the semiconductor part of it. I do think they will have a fully domestic semiconductor supply chain pretty soon. And that, along with government support in terms of putting money into these high-capex, maybe lower-rate-of-return investments, will allow them to more aggressively build out the domestic semiconductor infrastructure.Once you have that infrastructure, then you can produce all the AI chips, all the phone chips, and all the analog chips that you need for your various embodied AI products and EVs and all these other leading sectors. And once you have that, that means you’re no longer constrained. You’re no longer constrained by compute. You’re no longer constrained by possible Western tech export controls on you.So then the AI players in China are equal in terms of AI infrastructure. And that allows them to compete a little bit better with their American counterparts. Now, they do have some obvious advantages over their American counterparts. We’ll have to see how this plays out, because China does have this entire grid build-out that is just unrivaled. And as we move to a more electrified global economy, being able to build not only data centers but the entire grid is actually a huge competitive advantage over the rest of the world.I don’t really want to say who wins the AI race, because I feel like you can only lose the AI race by not participating and investing in it. But if you invest and put a lot of money into it, like both the US and China have, both of these countries will have a huge share of the global economy going forward.TP (54:15)I just don’t see how you can put this much effort into AI in America and not get something out of it.Grace Shao (54:24)I just feel like it’s not a zero-sum game.TP (54:28)It’s only bad if you don’t try to build your own AI industry, right? If you don’t invest, that’s a problem. But if you invest, something good will happen, I think.Grace Shao (54:42)What about the smaller countries where they don’t have that capital, and maybe they don’t have that much capital to deploy into this, or even frankly the talent to build their whole AI stack? Where do they fit into all this?TP (54:53)Yeah, so I think that’s one of the factors that might help the Chinese ecosystem over time, because a lot of the open-source stuff is coming out of China right now. So if you’re from one of the smaller countries, let’s say Singapore, and you want to develop your AI sector, you are more likely to use an existing open-source model and do reinforcement learning training on top of that, and then develop your AI product on top of that, than use something you don’t have any control over, like Claude, for example.Grace Shao (55:41)Interesting that you use Singapore, because I was just there last week and literally OpenAI just announced their satellite office. I think they said they would employ 200 people. Singapore is an interesting story because, if anything, they’re super gung-ho on AI, from top-level diplomats and ministers to companies. So it will be interesting to see how they play out this strategy.My question for Singapore is: they can attract a lot of talent globally to go over. They can attract a lot of new companies to go over, which is what they did with the internet era too. ByteDance, Tencent, Facebook, everyone’s there. But then what is the value they propose for the locals? Or how do they plan to diffuse AI into the economy? I don’t know how they make themselves that relevant globally beyond being a hub for these companies.TP (56:35)That’s a very hard thing to say because I don’t see Singapore, just on its own population, actually developing anything unique. The people who would work in Singapore’s AI industry could work in any other country also. So I think Singapore has always put itself out there by being a country that attracts talent from all over Asia, right? And they attract a lot of capital also from the rest of Asia.There have been a lot of issues in recent years where they say all this money coming in hasn’t really helped the local-born population in Singapore. So that is something interesting to watch out for.Grace Shao (57:25)Yeah. I don’t want to go on a tangent on Singapore too much. So, last two questions. One is: what is one underappreciated hard-tech development you think people are missing?TP (57:38)Yeah. Last year, I wrote a thread about a list of what I call sanction-breaking tech that was happening in China. A lot of these are not things you see in the media as much, because they are the zero-to-one steps in the upstream supply chain that need to be achieved in order for an end product to be built three or four years later.So things like high-speed analog-to-digital converters and digital-to-analog converters, advanced diamond substrate for heat sinks and other purposes, high-end gallium chip designs, and a lot of the lower-level material science-related stuff that people don’t really see.But once China develops these things, that’s when you see this really fast iteration afterward. Because everything in China is kind of built upon the idea of having the upstream supply chain and the process knowledge. And then it can iterate through the end product a lot faster.So as fast as China has moved in the past 20 years, I don’t think the West is really prepared for what is to come out of China in the next 10 years. I really don’t.Grace Shao (59:36)Interesting. Okay. Well, I think that’s a topic that no one really has an answer to. No one really knows the future, right? But I appreciate your thoughtful answer.My last question for you is a question I ask everyone who comes on the show. What is one differentiated view you hold that you think is non-consensus?TP (1:00:00)Interesting. Well, one thing that I’ve talked a lot about with people recently is that if you listen to mainstream media, when they talk about China, they always talk about the economy not doing well and that China has this housing bubble that’s apparently a real problem, right? And that China has this demographic problem going forward, and that’s why China might have problems going forward.I’ve actually always held the opposite belief, in that I’m always under the impression that China grew overly rapidly for many years because it built up this real estate bubble, and all that money went to real estate instead of the tech sectors. And at a certain point, it decided that it could no longer blow up this real estate bubble because young people weren’t getting married and having kids because they couldn’t afford homes. So it deliberately deflated the real estate bubble in order to solve this problem.And then it still claims to have grown at around 5% a year for the past few years. If you can deflate a bubble and grow at 5% a year, that is quite the accomplishment, actually. So I would say the Chinese economy is quite healthy.You would rather have an economy that can grow strongly in the middle of an asset bubble deflation versus an economy that is growing just a little bit in the middle of a historically large asset bubble, like you have in the equity market in the US.Grace Shao (1:02:05)That’s a very interesting take, actually. I’ve never heard someone say that. But yeah, I kind of see where you’re coming from.TP (1:02:14)Yeah, that is my take.Grace Shao (1:02:17)I love it. TP, look, I’ve taken up an hour of your time. I really appreciate your insights. And you entertained my brain going in all directions as well. We’ve really talked about a lot of different topics today.Is there anything else you think we didn’t cover that you would like to share with everyone? Or do you think we can always pick this up again another time?TP (1:02:40)The only thing I would say to everyone out there is, if you enjoy AI, try one of the cheap Chinese models and see how it works for you. I’ve tried it myself. It’s great for my work purposes. And I highly recommend everyone use Kimi.Grace Shao (1:02:58)There’s a plug. No, I’m kidding. They are good, actually. I think I use different models for different things, but ultimately I find that if you’re really using them for more basic writing and everything, the Western ones are better. But if you’re really hosting your own models and running your own agents, then a lot of the Chinese ones are a lot more cost-efficient.So thanks again, TP. Thank you so much for your time.TP (1:03:28)I’m glad to be here. I’m glad to be on your show. And you can all follow me on X at TP Huang. I’m really glad to be on this show.Grace Shao (1:03:38)Definitely. And TP is on Substack too.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• The reasons to open-source and the future of AI bootstrapping with Tiezhen Wang 25.05.2026 1h 3min

  Joining me today is Tiezhen Wang (Tom), formerly of Hugging Face, where he worked with researchers in China, Australia, South Korea, Japan and across APAC, to help make open-source models more discoverable, usable, and visible to the global developer community. In this conversation, Tiezhen explains why Hugging Face became the GitHub for models and why open source is not just a distribution mechanism but a different way of coordinating research. We discuss why Chinese AI labs have leaned so aggressively into open models, how DeepSeek changed the commercial logic of open source, and why Qwen, Kimi, GLM, MiniMax, and others are using openness as a way to win attention, recruit talent, and accelerate the whole ecosystem.His core argument is that China’s open-source AI push has three layers. At the researcher level, open source preserves attribution and career mobility. At the company level, open models can become benchmark-led marketing, developer distribution, and a recruiting advantage. At the ecosystem level, government and university incentives are beginning to cultivate open-source culture among younger engineers.We also discuss why US frontier labs have pulled back from openness as research and business have become more tightly coupled, why distillation is much murkier than the public debate suggests, and how DeepSeek’s releases increasingly function as shared R&D for the broader AI ecosystem. The conversation then turns to monetization: why open-weight labs can still make money through API tokens, base-model access, post-training services, and inference optimization.Finally, he lays out his current thinking on AI bootstrapping: the idea that agents may eventually help improve their own harnesses, generate training data, and even improve the models they rely on. We close on a more philosophical question: if a handful of closed labs control access to frontier capability, open source becomes more than a technical preference. It becomes a check on the concentration of power.Tiezhen/ Tom is based in Sydney, Australia. Feel free to reach out to him on X to chat.To find the previous episodes of Differentiated Understanding, see here.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, researchers, founders, and product managers. For more information on the podcast series, see here.Chapters04:07 The Philosophy of Open Source at Hugging Face12:51 Challenges and Opportunities in Open Source17:12 The Role of Collaboration in Research21:50 The Future of Open Source and AI33:58 What Constitutes Distillation in AI37:18 Navigating Copyright and AI Distillation37:43 The APAC AI Landscape: Insights Beyond China43:08 Understanding the Ecosystem: Labs vs. Hyperscalers46:21 Monetizing Open Source AI Models52:02 The Future of AI: Bootstrapping and Self-EvolutionTranscript (AI- generated for reference only)Grace Shao (00:00)Tie Zhen thank you so much for joining us today. I’m really excited to have you on. We’ve been trying to make this happen for a while and just so glad the timing’s finally worked out. To start, can you tell us a bit about yourself, your journey, and where you’re at right now in your career and how you see the whole ecosystem? And also, just help us understand Hugging Face a little bit as well.Tiezhen Wang (00:19)Yeah, thanks, Grace, for inviting me. I know, sorry for the long delay. It has been a while, but I’m recently in transition because I just left Hugging Face. So to give you a quick information about very high-level overview, you can think of Hugging Face as the GitHub for AI. If you are not familiar with GitHub, you can think of Hugging Face as Amazon, where you can find all kinds of models in one store.And we are helping, so my job is to help researchers to get their models, which is the open source models on Hugging Face. And they can use the best, like all the tools, all the services on Hugging Face to make their models more discoverable and available to everyone. We also offer all kinds of technologies. For example, we allow them to create demos so that developers do not need to download the whole models.and they were able to try it out and see how it goes. And we also offer services so you can create your own agent using open source models. We do all kinds of scaffolding on top of open source models. another part of work that we do is to help them get more traction. We use LinkedIn.I use Twitter mostly to help them getting well known by the public. And we write analysis on their models and letting people know what are the new inventions from the model, et cetera. we work with researchers across the world. Like myself, it’s focused on APAC, especially Chinese researchers. Yeah, that’s pretty much the goal.quick overview of what I do. If you have any questions, just let me know.Grace Shao (02:03)And how did you get to this role? Because I understand you were with Google for quite a while as well.Tiezhen Wang (02:07)Yes, I was with Google as an engineer. work on ML frameworks. But then we had a bunch of reorg. And I was assigned to a project which is not open-source. But I really like talking to people in the open source world. It’s kind of very different. So when you are paid to work something versus you want to work on something yourself,Like you have very different mentality and very different feelings. So when I was working on the open source machine learning framework, I talked to people outside Google. And I can see the stars in their eyes. They do want to work on something they want. And even though they may not get paid, et cetera, I really like this feeling. So after I was assigned to the non-open-source project, I want to try something likenew but also in open source and I was like talking to people in Hugging Face and I really liked them. At that time, like Hugging Face was not like part of the mainstream. It was like a niche product for researchers where researchers can upload models. But I do see there’s a huge potential for Hugging Face to grow up because first I believe in open source and the second like Hugging Face is going to be the entry point where like all people will come in and search for open source models. But the most important of all is that I feel that Hugging Face is a company who understands how open source works. Open source is a huge leverage. If you use it well, it’s going to be very powerful. And Hugging Face is like 200 people, like very small companies compared to other companies growing up from the same area. But they are able to use open source as a leverage.and called for collaborations across the world and do very impactful things. a lot of people, a lot of big companies are doing open source, but they just don’t understand this age. That’s the essence of open source. And I do feel that Hugging Face is doing really well there. That’s one of the reasons why I want to join Hugging Face.Grace Shao (04:06)Yeah, I think that’s amazing. I think that’s something we definitely will double click on later, especially when we talk about why China’s labs seem to have been embracing open source. Just kind of one last question on just the whole ecosystem and how hugging face fit into it. What was the philosophy really held by the whole company? Because I actually listened to one of the founders interviews, Clem’s interview recently. And during the interview, he talked about how Chinese scientists have always been long term contributors to open source technology. And then he said it was really like kind of a pivotal moment around 2022 where American open source contributors kind of took a step back and then there was a sentimental shift in the ecosystem. Why is that and how does Hugging Face kind of view the whole ecosystem?Tiezhen Wang (04:47)Yeah, there are several questions. Let me try to address them one by one. The first one is the philosophy behind Hugging Face. I think it’s really the mindset. so anything that we see where we can have a collaboration, like Hugging Face will just reach out and see if we can collaborate. So if you go to see a lot of work released by researchers, they will have paper on arXiv.and also their project on GitHub. And you’ll see me on all of these issue number one, which is the first issue after the repository has been released. And we just write something saying, offer blah, blah, blah. Do you want to collaborate on something? So for anything that we can collaborate on, we will just call for collaboration. And some we’ll go through, some we’ll not. But this collaborative mindset is very, very different from.like a business point of view. From a business point of view, you will first think, what is my edge and how I win the market, how I compete with others, and what are the end areas. After the competition, what’s the end game, how it will go. So that’s the way of how you can justify the investment and everything. In open source world, it’s totally different. It’s like, I want to do something.I just say it and I do it and there are developers who want to join in and we do it together and we grow the pie gradually. we do not have like, let me put it the other way. So if you see an open source model coming from one of the Chinese lab, for example, GLM 5.1 is released and you may think like Kimi or Minimax like other open source model provider.in China would compete with them. But actually not. Like you will see they are commenting on the Twitter saying, congratulations, et cetera. This is a collaborative mindset where everyone is stepping up on each other. we can do a lot of, as a group, can continue to push the frontier forward. So I think this is very, very different.Yeah, and talking about your second question, the Chinese, well, I wouldn’t say labs. Chinese researchers, labs, companies, et cetera, they all want open source. I think there are three different folds. The first one is on the researcher side. A researcher would always prefer if their work is open source. That’s coming from their academia background, because when youLike on the CS world, when you write a paper, you have to show that it’s actually working. You have to show that all the numbers are real. Other people should be able to verify that. And you can only do that by releasing your code, releasing your models to the community so that other people can evaluate. So a researcher, after they graduate and they go to a company, they will bring this mindset forward. And by default, they are open source people.And another perspective is for their self, for the career development of themselves. So as an engineer in big companies, it’s very often that you are working on some project and nobody knows that you are working on that project until you say that out on the game or on your resume. But open source is very different. We know precisely who has contributed to DeepSeek before.And that’s very attractive for for researchers, because if I have done great work, I want the whole world to know that I’m doing excellent work. This will help me have better branding, help me to do more collaboration, help me in the future step in the career. So a researcher would always love open source, by default. So that’s the first part from a researcher’s level. The second one is from business level.So well for individual is quite easy to embrace open source from manager level from the executive, they need to justify the investment on open source. I have to spend tens of millions in training a model and you want me to give it for free. That’s crazy, right? That’s how people think before DeepSeek. Although we have lot of open source models before DeepSeek, but the trend is completely changed.Before DeepSeek, people were thinking, oh, maybe the model is not that good. Maybe I’ll just open source it. But if the model is good enough, maybe I’ll keep it for private. And that’s one of the reasons why you see a lot of people were saying open source is not that good, especially from Robyn. And lot of people do not understand how the open source works.works. But then people do realize that if they do not open source, they do not even have a chance to stand on the market. Because their model first is not really good. If they just compete on the marketing level, on the business level, they do not stand a chance, not even a chance. So you spend tens of millions and you get nothing. But if you open source, at least you have some sharing and people will remember. And also you can have the market from.for the researchers. I think Qwen team was one of the first team who understand it from a business level and start like open sourcing work. And as the result, it’s very, very good. Like they almost taken the ecosystem from Llama and now they are becoming the default for researchers to do research, which is like a huge branding for Alibaba. And like, I guess like if Alibaba wants to do any kind of business, like it’s quite easy for them.to approach to researchers saying, we are not nobody, right? We are the provider of Qwen and everyone wants to talk with them. And another side for the business is that they find it really hard to attract top talent if they do not do open source, because all these talents want their name on papers, et cetera. if they can pay a lot of money.but they still do not have the best talent. But on the other side, if they do open source and the researchers know that they come to this group and they can have their name marked on history, it’s going to be very attractive. So like this company, even not releasing the best models, they try to release something to make researchers happy. It’s kind of like their...company perk. So that’s another route. But after DeepSeek, everything changed. People know that if I do open source, I can have huge branding for my company. DeepSeek is not doing any kind of commercial stuff, like alteration to cusTiezhen Wangers. Yet they still have a huge evaluation of, I think the most recent number is [unclear: “14 million HKD” in transcript; confirm figure].That’s a lot of money. So by doing open source alone, they can make money. And that changed the mindset for lot of people. so after DeepSeek, Kimi, GLM, Minimax, and StepFun, they all come into this open source world. actually, they have made a lot of success stories, like GLM and Kimi, by doing open source, lot more people understand them. And they kind of open up.the global market, not just the market in China. for them, I feel that it’s not like losing a lot of money because they doing advertisement in a different way. Kimi was spending tens of millions RMB per year on advertisement. And the result is very short retention. People know them, come to their side, and they do not feel any different. And they just move away. Now, the researcher team, the manager, the executive means, knows that the best score on open source benchmark is the best advertisement. So they can concentrate all their power, not wasting them on advertisement, but concentrating all their money and resources on training the best model. But this best self, it’s the best marketing, and they can create great models and start earning money.So I feel that on the business level, everything starts to make sense. But now there is a new challenge, which is how you can stop people from taking the free ride. It’s a longstanding problem for open source. I did something, for example, I made a database. I spent a ton of engineering hours. I open sourced it. But I’m not making any money, because the cloud provider is taking that for free and start making money and monetizing it.it’s happening for open-source world as well. I open-source the model and all these inference providers and chipmakers and BDA-AMD are making money, but not the researcher who created the initial model. That’s why you see some licensing change and discussion on that. Kimi did the first non-commercial license, and then MiniMax made a more restrictive version. Tiezhen Wang (13:40)made a more restrictive version. But I don’t think that’s the final version. People are still trying different things. And I believe maybe in one or two years, we will have a more standard way of balancing open source and commercialization, et cetera. So that’s the second level. The third level is the third level. So the Chinese government is really encouraging people to do open source.If you do open source, you have extra credits on your bachelor education, et cetera. And Shenzhen recently announced a very interesting policy. So you can have housing points if you do open source on GitHub. basically, they are categorizing.Grace Shao (14:21)So the incentive, yeah, go straight to the students, like even in academia, while they’re still in university.Tiezhen Wang (14:27)Yeah, so it’s kind of cultivating this open source culture when other researchers and developers are still in universities, which is really good. So I do feel that the culture of open source is, if they are winning the young students, we are going to see more open source projects. And to be honest, I do feel that that’s the right approach.Because if you’re not thinking about open source, you are thinking like traditional way of collaborating with people, which is company or corporation. And I feel that the essence of why we had cooperation or company is not keeping peace with how we evolve now. I think about, you set up a company in Hong Kong 200 years ago. Why? Because you have a group of people. You want this group of people.That’s why it’s called company. You have a group of people and you want them to work together. And how you can make sure that everyone had their benefits. Everyone is doing a lot of work. Obviously, they want to have a return. And you do that by setting up the shares and also the voting system. that’s how a group of people is working together. But now the word company has changed. It’s more like amulti-international company where the worker in the company has no work in deciding how the company runs. Whereas open source work is more likely the original version of a company. You have GitHub, you know who has contributed what. Everyone knows your contribution, and you can have your name listed. the group of people coming from all around the world, can.collaborate on something. They do not need to be part of a big company going through all the interview process. They can just collaborate. So I think that’s very, very interesting. And now with Zoom, Tencent meetings, and all the Google Docs, it’s much easier to collaborate internationally. I don’t need to know who is contributing to the PR, but I know someone is interested in my project, and we can work together. And I feel that.That’s probably the future way of how people can collaborate. that’s to end the last point on society level. I think the society is advocating for open source. also open source is probably the way how the society will evolve.Grace Shao (16:48)Thank you. is like so insightful pack that I have to digest that. But you you mentioned quite a few different topics, which I can definitely take this straight, conversing different directions to start. have two questions and they’re actually unrelated. So one at a time. Number one is you really make a point about China being really, you know, strong advocate on open sourcing the LLMs. However, I thinkCould you tell us the history of open source in China in general? Was there a tradition to want open source technology even pre-LMDs? That’s number one, first half of that question. Second half of that is you say there’s a lot of incentive for researchers to actually want to open source everything, right? And then therefore they can claim their contribution. Well, in the recent interview between Zhang Xiaojun and...deep minds, Yao Shui Yu, I think maybe you’ve also listened to it. You know, one thing that really stood out to me was how he was saying people need to be like responsible. And like for someone who’s not technical, I actually really struggled to understand what he meant at first until like actually Jiang Xiaoxuan actually asked him to clarify as well. His whole point is that in academia, people are so used to only claiming a certain section of what they contribute. So for example, for a big piece of paper or research,that you would take credit for what you contributed, right? And you want to make sure that it’s best optimized, known, heard, seen, whatever, right? Recognized. However, in terms of how LLM can work properly in terms of the long run, whether it’s like, you know, further in post-training and further, you know, know, usage, whatnot, it’s important that people don’t claim so much credit to their own part of the work. It’s more important that people work collaboratively. But kind of to your point on open source that, you know, they can work collaboratively and make sure that each piece works together better instead of each piece working best on their own. So it kind of contradicts your comment on why people want open source, because in that sense, wouldn’t it make sense for people to not want open source? I don’t know. That’s another question. And the third part of this is really if open source makes so much sense for tech companies and makes so much sense for academics.then why are the American labs so anti open source right now? Like what is driving that? Is it purely because commercial reasons or philosophical reasons? This is very big, but you did throw a lot at me. So I’m going to throw these questions back at you.Tiezhen Wang (19:07)Yes, sorry for my very long answer. I think it’s probably by itself worth writing a blog post with enough content, and I can elaborate more. But great questions for the story. Can you remind me? I guess we can go through them one by one. Can you do mine? Yeah.Grace Shao (19:25)Just like in general, source China, China open source. What’s the sense on that? Beyond LLM, right? Like why did Chinese companies always contribute to open source technology? Clem talked about this in his interview, but he didn’t go into that about it, right? So number two was just about, yeah, number two was just about like, why do these academics want to claim their names, right? Is it better for the company in the end or is it just best for them, like the selfish reasons?Tiezhen Wang (19:37)Yeah, okay. Let’s try it. Yes. Mm-hmm. Yep.Grace Shao (19:52)And number three is why are American labs kind of anti open source right now?Tiezhen Wang (19:56)Yeah, so let’s try to address the first one. I think it’s a great question. And I do see the shift. So I feel that AI is probably one of the very few areas where Chinese open source contributors dominate. If you look back to, for example, I would say the initial days of modern open source comes from like anLinux or Apache or database and everything. And where you do see a lot of individual contributors from China, but you are not seeing enough Chinese company creating a project. And then the project gets adopted globally. You are seeing that gradually when we move to the area of cloud-native, like when the Kubernetes comes out.And a lot of Chinese cloud providers are trying to really pay attention to this whole open source world. And you will see that this grows. But now it’s like this. So it grows exponentially. So I think it comes from two folds. The first one is the Chinese participation in the global market. It needs time to warm up.Like for example, lot of Chinese contributors, they can only contribute two projects in Chinese because of the language barrier. So that kind of limits how much they can actually do. And now with larger language models, with better education in the new generation of developers, the language barrier is not that strong. That’s why.That’s how the Chinese open source contributors can make a better impact. And another one is, so in the traditional way of a company’s, like how a company’s structure itself, if you do open source project, it’s kind of hard to justify your credits because the open source by itself is not the core business of a company. There are very, very few companies whohad their core business made on open source. Like PingCAP could be one of them. PingCAP start with open source and then find monetization plan. But that’s so small. So few of them. And in the new areas, a lot of companies, their core business is open source and plus monetization. Even for IPO companies, for public list of companies, Minimax is basically one such example. They have their best models, open source.and then trying to make money. So this is very different. If your core business is open source, of course you will put more resource on open source. And it’s more likely for your project to gain a lot of developers. And I feel that the third one is the international collaboration has never been easier before, apart from language barrier. So after the pandemic, I feel that all of a sudden everyone is used tolike Zoom and Hangout and collaborating with someone who you don’t see face to face. And this is a great chance for open source project to ramp up. Because before that, you have to meet face to face, and the bandwidth and the people you can meet is kind of limited. And now you have a huge, like, as long as your project is great, like you have a huge pool of potential developers.And the last one is probably AI by itself, like coding agent itself. Although it does make code review much harder because there are probably a of AI scope. But it really lowers barrier of who can contribute to an open source project. Before, you want to contribute to a project. They are probably developing language you do not know. And also, the code base is pretty strong. A developer might not be able to.contribute to the project until he has a very thorough understanding. And that’s probably like months of work. Now you can just ask AI how this part works. And I only need this feature. And what are the code I need to modify? And I can just give it a test locally, and it works. I contributed to some Rust project without being a Rust expert. So that’s how AI makes everything better.So I think it has all these reasons. There are probably more, but I think someone at the end of the day, in two or three years, maybe starting to write some history about how every single aspect of technology, moment and everything, people’s mindset shift, how to cultivate the open source spirit. But I think, yeah, that’s the...Top ones coming out of my mind.Grace Shao (24:26)And then the second question was just that would researchers focused on their own name and frankly ego in this sense actually be the best way to help cultivate the best LLM or whatever whatever product that’s the end product that’s to be shipped. Does that make sense? Because it kind of contradicts what Yao Shunyi was saying on Zhang Xiaoxuan’s podcast, right? He was saying in that sense a lot of researcherswill try so hard to only own what they are working on, but they have less of a sense of responsibility for the bigger project.Tiezhen Wang (25:00)Yeah, I haven’t really read the broadcast entirely. But talking to your point, feel that open source or not, or having researchers name these data on papers or not, it’s actually a game changer. If you are a researcher and you might want to stay in academia, why? Because all the papers you publish is very important to your career. Everyone sees.Like you have published which paper with like who and the paper well like also mentioned that you contributed which part of work are you Look like my corresponding author. Are you the main person or you are just like contributing a part of it? And you there’s h-index to measure the impact of a researcher So like you have all that infrastructure working for you if you stay in academia But if all of a sudden you want to start workingfor company, you lose off that because the company might be able to say, we have a policy where all your open source and all your paper publications, even writing a blog is controlled by the PR team and they have to decide if you can do certain level of things. So your exposure is reduced. You might be very happy because you earn much more, like 10x the salary compared to be an assistant professor.Like after five years, if you ever want to go back to academia, that’s impossible because you lose all of your track record. And with open sourcing, things are very different. The top nailing app wants to hire the best talent from academia. And the people from academia wants to work for the app. But at the end of the day, the two systemIt’s the same system because all the record is public. So it’s very easy for people to come in and come out, come in and come out. It’s kind of different from people coming from academia and then lose track in and get lost in the company world. So I do feel that having your name listed on the work you publish is very important. It is kind of the concept of [Chinese phrase unclear]. So you.your branding grow with whatever you have done. So your reputation is based on whatever you contribute. So you need to pay extra attention on that.Grace Shao (27:16)I see what you mean. And the last bit of just now what we’re talking about was just why is that if you believe open source makes so much sense to these researchers, that so many researchers in the US or at least some of the companies, the entities right now are not willing to go open source.Tiezhen Wang (27:31)Yeah, there are lot of companies who change their position. Google used to be the company which impressed open source the most, like Google open sourced and like TensorFlow Kubernetes and bunch of other important open source projects. The open sourced transformer, which is the cornerstone of our modern AI system. So Google was really impressing open source. But I feel thatAt some point in time, Google stopped doing all the open source work because they kind of lighting OpenAI and other companies taking the free ride. Google did a lot of fundamental work and then it’s kind of taken by other companies for free. And also, OpenAI and Anthropic, I feel that they are still contributing to open source, but that’s not their main project.Their main project are hidden secrets that they do not want to share so that other people can catch up. And the research and business is getting so coupled. So for example, a researcher in OpenAI found a way to improve the intelligence level by 10%. Let’s take o1, for example. They managed to find out how to make model think. And by this chain-of-thought and thinking process, the model islike way much smaller, way much smarter. But they do not want to share the gist.As long as they do, other people will catch them up. So it’s kind of a very restricted environment. Although researchers may still want to open source some of their work and have their name listed there and sharing very detailed observation. But the business doesn’t just allow them to because you are trying to. Yeah. also, researchers can also talk to like in some conference. I know how open source, sorry, I know how o1 was roughly made by reading bunch of YouTube videos from the internet made by OpenAI researchers. But after that, you see less and less very detailed research sharing, even the videos or recordings by them. I guess they kind of learned the lesson.But like, yeah, yeah, that could be part of the rhythm. On the open source world, like, it’s kind of different. Like, so before DeepSecR1 was released, lot of people were speculating how OpenAI was doing o1, and they’re trying things in different way. And after DeepSecR1 is released with all the recipes and all the data they shared, like, the open source world seems to be converged on the path. Although that path mightGrace Shao (29:46)There can’t be compliance reasons.Tiezhen Wang (30:12)be different from o1 because we never know how o1 was made. But then because of DeepSeek’s contribution and sharing, everyone knows how to make thinking chain. And the whole ecosystem is evolving really, really fast. That’s one of the real value of open source because everyone can just collaborate. No one is holding secrets. Well, there are still a lot of secrets on how you can run as efficient as DeepSeek, but that’sLike too technical, like that’s not too much on the research side. yeah, like I...Grace Shao (30:44)So on that note, yeah, I want to follow up on that. I think I recently wrote about something which is, speaking to our researchers, it got me a sense that DeepSeek in a way is now becoming essentially like a foundation for everyone because, you know, a lot of the labs in China are looking to DeepSeek to see if there’s any like, you know, engineering breakthrough, like your point, and they build on top of each other. Help us understand like each of the labs, because you said, they’re cost constraint, they’re compute constraint.Tiezhen Wang (31:06)Thank you.Grace Shao (31:12)They’re teleconstrained, right? Their resources are constrained and every single asset you can think of compared to the American peers. Now, why does it make sense that they all open source and how are they all optimizing for their own goals at thisTiezhen Wang (31:24)Yeah. So open source by itself, as we just talked about, is an accelerator of the whole ecosystem. So DeepSeek shared all their like, knowings and discoveries and what things work, what things doesn’t work. This by itself is accelerating the whole industry, not just Chinese open source, but also like US open source and US like closed source. Like they just don’t say how much they learn from DeepSeek, but I believe everyone is learning from DeepSeek.Not just that, DeepSeek also contributed to GRPO, which has become the most used algorithm, reinforcement learning algorithm in the industry. So they did a lot of contributions. if you check recent model architecture evolution, what’s proposed by DeepSeek is becoming the standard and getting adopted by many people. For example, Kimi 2.5 was using a model architecture very similar to DeepSeqs. And GLM 5.1 was adopting a lot of components from DeepSeek architecture as well. So it’s kind of sharing and learning and co-evolvement is one of the, I would say, secret of how China is able to. catch up with the US in certain area, although having restricted compute and restricted capital, I would say. If US open source is working again, like the whole ecosystem, like everyone was trying to open source, I would say the human race would be evolving much faster than what we are doing now.Grace Shao (33:01)So on that, how do we understand the accusations of what is being distilled? What is technically shared? What is, how do I understand the gray area of that? Like the accusations from a lot of American labs, Chinese labs right now, like you just said, a lot of American labs are learning from Chinese labs. Frankly, within the researcher community, it’s not even Chinese versus US, it’s really just labs with each other and against each other if they have to, right? Intellectually competing. So then how do we understand what the, industry agreement is on a distillation, why is it so contentious rightTiezhen Wang (33:32)On distillation, yeah, that’s a great question. I can only give you my perspective. first, distillation is a very broad word. We are distilling from each other as well. I learn from you, you’re learning from me, and we are all learning from books and papers and all this public information. So I would say,distillation is a very common practice, like basically how you learn from others. Like you might have a model which summarizes the books and like doing bunch of explorations. And the way for the model itself to move forward and evolve is to distill from its historical data and historical experiments. And like that works for like another model trying to learn like your model as well.And on the research field, distillation is very common. DeepSeek R1 was released with MIT license. Specifically, so I actually asked the team about it. They choose MIT license because they want their model to be distilled by others. Because that was the only model that works really well with the thinking chain. And they want all the open source model to be able to have that.Like they have shared all the recipes, but others do not have data. So DeepSeek design like they’re like small models so that and also the recipes so that other people can easily distill DeepSeek, getting the thinking chain and use that on their own models. like this distillation is happening like everywhere. And I think like US companies are distilling from each other as well. Like I’ve seen like the recent discussion on Twitter in public.where Elon Musk and Sam Altman were kind of battle on that. yeah. And if you think about it the other way, so if you do not allow a model to distill, I mean, the output of a model to be able to train a model which is from a competitor, it’s kind of a very interesting point. Like if we say, I’m reading a book.I’m telling you the story. So you, after reading the output from me, which I think of me as a model, you’re reading my summary and you are not allowed to share the summary to others. You have to read the book, the initial book, not using my summary because of the license, et cetera. That’s kind of ridiculous. That’s not how human transfer knowledge in the past a few thousand years.Like I have a very bold argument. I think that anything like generated by AI should not be copyrightable. So like it should be in public domain, like anything generated by AI, because like anything generated by AI is a distillation of like human entire history and everything that human has created. And if you just take that for free and asking other people do not use that.Like it’s kind of a waste and it’s kind of like blocking people from evolving forward. Because like human content do not have this restriction and why you are putting this restriction on something not copyrightable and generated by machine. So that’s something I do not really understand. So I do see there are terms and conditions saying that my model output cannot be used to improve other models. But I don’t think that’s kind of valid.I’m not sure if someone eventually will file something on the court and we can have a case on that. currently, think there are a lot of things to discuss, but it’s not about if we can distill a model or not, but about something bigger. Should the model creator even have this right to restrict others from distilling from their models?Grace Shao (37:18)That’s really interesting. I think that a lot of the discussions in the public space is really about whether you can use copyright work of human output. And then the argument is always like, just you cannot distill because the company said there’s no distillation allowed. But like to your point, there is no actual clear black and white rule of regulation around this right now. And in fact, it’s it’s bit murky. Yeah. Yeah, yeah, that’s interesting.Tiezhen Wang (37:37)I’m not a lawyer, but I can find a clear answer on that.Grace Shao (37:43)Okay, I want to kind of go to China. Like we’ve kind of talked a bit about the big picture. Well, a lot about the big picture. But let’s look at just the China labs. mean, I know that you represent APAC back then with Hugging Face and you worked around APAC, you lived in Australia. But for the sake of this, know, Chinese labs right now probably are the most relevant out of APAC. Do you think I’m missing anything actually on the APAC conversation? Like, do you think anyone else in the region is relevant in this space that we can talk about?Tiezhen Wang (38:08)Korea is doing really, well. Yeah, Korea is really well. Well, the most, one of the best model is probably Upstage. they, initially they create a Korean model leaderboard, like open source version of like model leaderboard. Well, no, no, the leaderboard was not funded by government. The, the, the,Grace Shao (38:10)Yeah, yeah, give us some picture on that. That’s funded by their government, right? That’s their government funded.Tiezhen Wang (38:26)So Korean, it’s actually a very impactful country, but as the other days, there aren’t enough Korean data. Even for ChatGPT, I think until ChatGPT 4, the model doesn’t speak good Korean. So the model was able to speak very good Chinese from day one, like from ChatGPT 3.5, but because of the data volume, et cetera, speaking Korean was always a challenge until ChatGPT 4.At the time, like now, the open source model is able to speak like Korean. So Upstage create a leaderboard. So the way they solve problem is very interesting. They’re not solving problem by solving problem. They’re solving problem by helping others to solve the problem. So instead of creating a model right away, they create.Grace Shao (39:09)I heard about Upstage from VC in Korea as well, but I don’t know the detail about it. Tell us more about who they are, what they’re doing.Tiezhen Wang (39:15)Well, I don’t know too much about who they are, but I only see their open source contribution. I think the founder is a professor in Guangzhou, but he’s Korean and moved to US. Correct me if I’m wrong. I’m sorry. I’m not really up to date with that information. But I just want to call out because I think that’s a very interesting paradigm. For example, if you are a company, you have your own problem you want to solve.Tiezhen Wang (39:42)Like, how do you want to solve it? Like, you are going to hire some people and define a problem and try to use your own people to solve it, right? So that’s the old way. What’s the open source way? Is you publicly define the problem. You have a leaderboard. Like, you might do a private eval or public eval. It all depends on you. the problem is you have to list your problem.publicly and you have to tell everyone that you can contribute to this problem by submitting a model to a URL and we will do evaluation and see how each model is evolving on this area. So they basically have a leaderboard. And you will like a lot of researchers would be very interested because now they have a problem to solve before they do not even know Korean what’s the problem. So now they have a problem to solve and you will see that the curve goes like.it goes up because there are more and more researchers coming in and all their work are open sourced. So a new researcher wants to jump in the field. They will first have a look on the leaderboard to see how far away from a really usable benchmark. And then he can investigate all the previous attempts and find his own way of kind of just changing something a tiny bit.and apply that to the past people’s work and submit to the leaderboard. And now we are seeing people making progress on the leaderboard. So that’s a very, very clever way because it’s not one company solving the problem. It’s like we are opening the door for everyone to come into this playground and try to solve the problem together. I think within a few months, they were able to get thousands of submissions.which is really massive because just imagine you hire 10 % people, you won’t get that. And now it’s by this new way of doing things like building public, evolving public, you’re having a lot more submissions and you are educating people, et cetera. So they have this very impactful and inspiring leaderboard and then they release a model called Upstage for something. I can’t remember it has been a while.And the Korean dataset and the Korean models are accelerating very fast on high-netics. I think it is now the fourth largest models, speaking Korean. Yeah.Grace Shao (42:04)Very interesting. Yeah, I’m going to shamelessly self-plug in. People can listen to the episode I recorded with one of the leading Korean VCs as well that was published last week. He gave a good AI ecosystem breakdown of stuff.Tiezhen Wang (42:12)okay. Yeah, could you help me like do some like DD first and like just make sure that are correct. Yeah, you can.Grace Shao (42:22)Yeah. No, no, no, he did talk about Upstage as well. It’s very interesting. Yeah, I want to... sorry, go on.Tiezhen Wang (42:29)Yeah. And also, so you asked for APAC. So in Singapore, there are a lot of great researchers, like lot of Chinese researchers will go to Singapore as well, like Cancun too. Yeah.Grace Shao (42:43)Yeah, I think the ecosystem is a bit overlooked by I think Western markets, but definitely there’s a lot happening in around Asia. Like APAC has been including Australia as well as Southeast Asia, East Asia, and Northeast Asia. Okay, I want to bring it back to China. We’ve been kind of talking about China kind of more on the high level sense. Now looking at the companies themselves or the labs, we want to break it down. Just give us a sense like, how do we understand moonshot?Mini, Max, Deep Seek, Zhipu, if you have to put it in one bracket, versus the hyperscalers, Tencent, Alibaba, and ByteDance, in terms of their strategy, in terms of the capabilities. Like how should we understand this ecosystem right now? Are there other relevant players that you think I’ve missed, maybe like Xiaomi or anyone else?Tiezhen Wang (43:25)You mean like how the model creator, model lab, are collaborating with hyperscaler? Is that your question?Grace Shao (43:32)No, no, I just think it’s like the people, the people who are creating LLMs, like are researching on how to deploy LLMs. These are the main players, right? Now, how do they defer? How are they similar? What are we seeing like on the ground? Are some of them becoming more irrelevant? Are some of them becoming maybe say, we just talked about DeepSeek becoming almost infrastructure provider for the whole ecosystem.Tiezhen Wang (43:39)Yep.Grace Shao (43:58)You know, are that mini-max is very, focused on multimodality. Zhipu is very focused on coding capabilities. know, Alibaba really trying to push out commercialization by their existing applications. How successful that is, that’s a different question. Just like an overview of these players.Tiezhen Wang (44:14)Yeah, I do think they’re kind of converging. Yeah, because everyone knows that coding is, the whole market for coding is booming. And if you have a good coding model, you can sell it for profit, for large profit. And I do feel that everyone is rushing for coding. There are people exploring different things, like,For example, Tencent is putting a lot of efforts on Hunyuan and doing OCR stuff. And lot of other companies are doing video generation. But at the end of the day, think from a strategy level, I don’t feel that there are a lot of difference. It’s more likely a case where, you have data? For example, it makes a lot of sense for ByteDance and Kuaishou to work on video generation models because they have a ton of data. And also, do you have?like a large enough scale. Like for example, Kimi is not very active in making all the apps. Like Tencent is making models. They are making like great apps. Like for example, Yuanbao, like a QA app, like based on all the Tencent data, it’s very popular. Like they make QClaw. Like Tencent is able to do that because Tencent has a huge talent pool. Like Tencent is a huge company, Whereas like if you look at the Kimi, Kimi is very conservative in...doing all that because Kimi is still a very small company. So I think from a very high level, everyone was on the same page about the strategy. It’s just more, how much resource do you have? What are the advantage of you? Do you have data? Do you have distribution channel? Do you have product design, success story, et cetera? So yeah, I’m not sure if I answer your questions.Grace Shao (45:57)No, no, that’s good. So we kind of talked about why researchers want to open source. We talked about these companies are somewhat doing the same thing. So then this leads me to the question. We know that open source, open weight does not actually mean they don’t make money. However, obviously means that it’s harder to commercialize as we like alluded to with the US labs, why they make those decisions. Then how do these companies find ways to monetize and sustain their businesses then?Tiezhen Wang (46:21)Well, in the US, are also labs dedicated in making open source models and still making money from other donations or from other parts, like selling apps, et cetera. It’s basically the same way in China, too. For example, DeepSeek is run by, I would say, donations from the people who play the stock market.like there are labs run by VCs and lot of labs are already profitable by like selling tokens like GLM has recently raised the token price because like they see a huge number of demand and they’re like running short on compute. yeah, like open source can make money. Like there are a ton of ways for open source model provider to make money. I have a lot of ideas. if in case you are interested in like making yournot profitable, can contact me. But honestly, there are lot of ways. The simplest way is to sell token. If you have the best model, you can sell a token for profit and people will actually buy your token. so it’s very interesting because when we combine science and technology, always consider it’s the same thing.Grace Shao (47:15)Yes, everybody find Tiezhen Wang.Tiezhen Wang (47:37)For model, it’s the same. When we think about models, we just think of a model that generates tokens, et cetera. But actually, there are two different parts. The first one is training, where you have the model. And after you get training, open source the weight you trained. Another part is the inference. So you need to run a lot of optimized CUDA kernels in order to make your token cheap and fast.Either bracket can make a lot of money. For example, you can open source the fine-tuned model, not the base model. So if a company want to use open source model for fine-tuning on their own data, they cannot be building on a fine-tuned model. cannot build. They have to find the base model. And if the base model is not open sourced, you can sell that for profit.And also different clients might have different requirements on the model. The NeoLab can collaborate with the client directly and provide some kind of training and post-training support. So that’s a way of making a lot of money, actually, because training is very expensive. It involves very expensive researchers and data and compute. On the inference side, too.Grace Shao (48:45)Yeah.Tiezhen Wang (48:52)Because the inference is tightly coupled with the data center you own. So your optimization strategy does not, there’s no guarantee that your optimization will work on a different cluster. So a lot of people just do not open source the inference recipe because it’s not that useful. And also it’s kind of a moat. So the model provider who creates the model, they know how to optimize the model best.when the model is released because they have seen the model for four months and they have done a lot of optimization on the model inference. And when the model is out, like everyone else, it’s just starting to know the model and doing some optimizations. So of course, the model provider will sell token in a much efficient way compared to all other competitors. Three months later, when the outside inference provider gets toknow all the secrets and do very optimized kernels, there’s a new model coming up. So the model maker, the people who know the model from day zero, always have an advantage on selling the tokens. So that’s one of the very important ways how they can make money.Grace Shao (49:59)I see what you mean. Mm-hmm. Yeah. And does DeepSeek v4 coming out have an impact on how the GLMs of the world or Kimi make money? Like essentially their strategy with the fact that you just said they just raise prices on their tokens.Tiezhen Wang (50:16)Yeah, so GLM and Kimi doesn’t sell DeepSeek or Qwen. So they are not competing with each other directly. I would say the capabilities are on par with each other. So it’s more like a user test. Which one is better? There is no clearly winning between all three models. So we’ll see like Zhipu’s stock price was getting down because people were so worried about DeepSeek. But then they realized that like theZhipu token selling is not quite impacted, so the stock price bounced back. But at the end of the day, I would say it’s actually a good thing for them. So GLM 5.1 is adopting a lot of core design in DeepSeek with 3.2, I think, model architecture. And they were able to cut down the cost.by adopting all these exploration from DeepSeek. And now V4 Pro is out. I don’t know the details, but a very simple guess is that Zhipu is able to cut down the cost because they can adopt new things from DeepSeek architecture. So Zhipu on one side, because of the demand is so high, so they can increase the token price.and they can learn from DeepSeek and cut down the cost. So Zhipu is going, yeah, exactly, exactly.Grace Shao (51:34)you have a higher immersion. Yeah, this is something I think they’ve talked about as well, like really being able to learn from the engineering breakthroughs that DeepSeek puts out every time. Okay, I have mindful time. I just kind of want to have a few questions on the future outlook. You posted on X recently saying that you’ve been thinking a lot about how do we make AI bootstrap itself? And you you’re going through this transition yourself, you’re thinking about the future of AI. What does it mean for the open source future as well?Tell us a bit about where you stand right now and how you think of this bigger picture.Tiezhen Wang (52:06)Yeah, I’m still doing some exploration on my side. I think this whole AI bootstrapping logic has already been implemented by a lot of big lab internally. The idea is very simple. In compiler world, you can design a programming language and write a compiler probably in well-known languages like C. And then you will first implement this language using the C code.In the next iteration or after a few iterations, you are able to implement this language using your own language. So it’s called bootstrapping. You are basically evolving on your own. You are not relying on something which is not from your language. So it’s like putting it another way. If you see how normal living creature, how they replicate itself and how they evolve.I don’t need to have a screwdriver somewhere to engineer my kid, right? My kid’s just born. All by itself. But how far are we from AI to do similar things? Now we have a coding agent very powerful. We have our AI training pipeline recipe kind of stabilized, at least for small sized models. So are we really far fromlike AI able to get one of my idea, like I give him the direction, and he’s able to like first bootstrap a very simple version and gradually evolve towards that goal. Like I think it’s like highly possible. So at the end of the day, we might be able to like just tell him what I’m going to do without like giving him all the harness and all the like detailed guidance and.I’m not talking to him 100 times, and he’s able to first lay out what he needs to do and have a plan, and then probably design a DSL or agent all by himself. And probably he will create a model ways to help him to get adapted to this goal. And then he can just keep evolving. All I need to do is to give him more fuel.which is compute, and he’s able to do some evolution and all by himself. It’s kind of like if you have recently read Andrej Karpathy’s Twitter, there’s a concept called auto-research. But auto-research is just evolving on the model weight. It’s not evolving on the agent and harness. I think on the agent level and harness level, there are also a lot of things to do too.So I’m quite new on this journey. What I was able to do is to bootstrap a very simple agent and I can use that agent to optimize the agent. But I think eventually we will get the weights involved too. When the model realized, okay, I’m not just needing an agent, I can create a bunch of data and improve my weights. He’s able to evolve from that too.Grace Shao (55:05)So in the future, how important is the capability of the models versus the harness and then the industry expertise then? Because right now, so much of conversation is still about, you know, the models are very strong. We are seeing what you’re saying already, the agent’s starting to build out things auTiezhen Wangatically. But we still need the taste. We still need the industry expertise to guide them. I find it hard to imagine that, you know, you can plug in something just say, want this to be done. And the agent just starts doing it exactly to your taste and your...imagination? Do you really think that’s happening?Tiezhen Wang (55:34)Yeah, I do feel that it’s happening. Like we are using agent to like especially coding agent to code something that we are completely unfamiliar with. And I’m quite confident that it will actually work. The reason is that like I have defined a set of goals and as long as I see that is moving towards that direction, like I’m good. I do not need to understand the code line by line. Like it’s just the box. But like the difference is I’m using the coding agent.Grace Shao (55:58)Mm-hmm.Tiezhen Wang (56:02)to do something else. What I can do is to use the coding agent to improve coding agent itself. And using the coding agent to generate the data and train the model that coding agent is using. And I would call that a bootstrap, not like just using, like, I think I’m already quite happy with coding agent to do something else. But just like, yeah, yeah.Grace Shao (56:23)Interesting. I want to end on a more philosophical note. So do you view the argument that AI is going to replace humans then? Or do you think AI is going to be in the role to support humans if we could keep on going down this path?Tiezhen Wang (56:35)Well, it’s actually a very, interesting question. And I feel that people do have different feelings. But from a pure technology point of view, I do feel that it’s one condition of like, so technology is not the only thing that will decide everything. You mentioned that if it’s going to help humans, well, it’s not really technology by itself to decide.Like it can be used in different ways, in different social structure, in different like tradition and such. It’s like giving you a gun and you can do things in good. Yeah, it’s not. Yeah, yeah. But like just imagine that you’re going back to history with all your knowledge of modern society. Are you going to help theGrace Shao (57:12)It’s not a good analogy. But yeah.Tiezhen Wang (57:26)like the society, like the history you go back to? Or are you able to help? Like I think it’s basically the same. If you have AI that knows everything, you can just think of it as human being in like 2000 years in the future. And now you have it. And what it is going to help on the society. Like it really...Grace Shao (57:31)Yeah. It’s like that saying, your own capability of using it is the cap of itself. Also, I think there’s a lot of argument and discussion around the fact that even the society as we know it today, the knowledge work that we all have, that we normalize, are not even created until the recent 100 years. And if AI is to disrupt that and replace human in that sense.Why is it so bad? Because it alleviates us to do other things that human multifaceted beings that we are can do. Is that kind of part of the argument as well where like, even if it replace us or helps us, it’s only helping us actually alleviate some of the things, if you take a step back, the things that we don’t want to do, right? Where we can maybe go touch grass. I don’t know, maybe this is very optimistic view of it, but there has been people saying like,The cap on AI capability is a cap of your own intellectual, your own cap of your own ability to navigate or use AI. So the more you can use AI, the more it can help you. The less you can use it actually, the more it will replace you.Tiezhen Wang (58:45)Well, I think it’s very interesting to define what is you. Are you defining you as everyone, or are you defining you as people who have compute? Well, no, it’s not. It’s actually a very, very, very interesting question happening right now. You know, Anthropic coding agent is able to do lot of things. But people are of imagining that we areGrace Shao (58:53)We’re getting really philosophical now.Tiezhen Wang (59:09)Everyone is getting a lot more powerful with models. But what if one day Anthropic just say, you cannot use your coding agent to do certain things? It already happened. Anthropic said, you cannot use your agent to do auTiezhen Wangated tasks. The other thing, there could be other limitations. I have a very bold argument is that the reason why we are able to use AIso cheap that even us, like we do not own a data center, right? Even us can use that. It’s because our data is still valuable. You know, if you use subscriptions, your data is going to be distilled by Anthropic to further improve the model. And like they are able to give us a discount because they still need our data.Grace Shao (59:52)So your point is that one day when they capture enough data, they will not even give us this kind of access for free or for cheap price.Tiezhen Wang (59:59)It depends on how they define you. You ask them, like, you or something. How they define their user. How they define who could be part of the game. Like, if one day, like...Grace Shao (1:00:09)So then the question is, no, but then my question is, then there’s another argument where they’re saying too much power is in the hands of a few companies right now, right? Or a few founders, what not. We need open source, that’s your point, right? No, that’s really interesting. And that was actually gonna be the last question I was gonna ask you. What is one differentiated we hold? And I think you’ve already answered that in that sense, right? Yeah, I think it’s for us to really think about it. But then as the average user, my question is, how do you actually boycott?Tiezhen Wang (1:00:18)That’s why we need open source.Grace Shao (1:00:36)these companies or if not boycotting, how do you actually make an impact? Because if I’m not the developer creating an open source model for the average person to use, me as an average user, what do I do?Tiezhen Wang (1:00:47)Well, just use the model to do the thing you want to do. Try to embrace the model and be more patient for open source models because obviously the open source model is not as good as top tier closed source models. you kind of like, well, I mean, with open source models, you keep all your secret to yourself. So you can have.like better security and you have better control. Open source model will never betray you if you just write on your local laptop. So although the model is not performing as well because he’s not distilling you, right? So still you can trust on your open source models and give it a more task to do.Grace Shao (1:01:20)You host yourself.Tiezhen Wang (1:01:34)I do feel that a lot of open source model is actually capable of doing things. But the expectation might be, think of it as six months, like cloud version of, sorry. Let me put it another way. So think of it as old closed source models and be patient with that. And you can grow up with the open source model together.Grace Shao (1:01:54)That’s very interesting. Thank you so much for your time, Tiezhen Wang.Tiezhen Wang (1:01:56)And thank you, Grace.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Nathan Lambert Reflects on China’s AI Labs: DeepSeek, Open Models, and the 'Race' with the U.S. 19.05.2026 1h 3min

  Joining me today is Nathan Lambert, author of Interconnects AI and a post-training lead at the Allen Institute for AI. Nathan recently returned from a major tour of China’s leading AI labs, where he met with researchers and teams building some of the most impressive open models in the world.In this conversation, we discuss what Nathan saw on the ground: how Chinese AI labs differ from their U.S. counterparts, why open models have become such an important part of China’s AI strategy, and how labs like DeepSeek, Alibaba, ByteDance, Kimi, Z.ai, MiniMax, and others are navigating compute constraints, data access, and commercialization.We also dig into some of the most debated questions in AI today: Are Chinese labs really 6-9 months behind U.S. frontier labs? How meaningful are distillation accusations? Can domestic chips like Huawei’s make up for restricted access to Nvidia GPUs? And is China’s AI ecosystem actually government-directed, or is the reality more fragmented and commercially driven?Ultimately, this episode is a more nuanced look at China’s AI ecosystem that looks beyond simplistic narratives about subsidies, copying, or geopolitics, and instead examines the technical, cultural, and economic forces shaping the future of open models.Check out his two recent articles here:* Notes from inside China’s AI labs* How open model ecosystems compoundTo find the previous episodes of Differentiated Understanding, see here.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, researchers, founders, and product managers. For more information on the podcast series, see here.Chapters00:00 Insights from the China Trip11:51 Cultural Differences in AI Research18:15 The Role of DeepSeek in China’s AI Ecosystem25:26 Overview of Major Chinese AI Labs30:56 The Future of Open Source in AI37:50 Market Dynamics and Consolidation in AI42:28 Distillation and Model Convergence Controversies51:58 The Gap in AI Performance: US vs China61:09 Monetization Strategies in AI: A Comparative Analysis62:32 Government Influence and Misconceptions in AITranscript (AI-generated for reference only)Grace Shao (00:00)Nathan, thank you so much for joining us today. Yeah, really, really excited to finally hear your thoughts on your big China trip, on what’s happening between the Chinese AI labs and the U.S. AI labs, what you think the potential compute constraints might mean for these labs and their performance in the future, and obviously the open-source ecosystem. So before we get into all of that, could you...Nathan Lambert (00:02)Yeah, thanks for having me.Grace Shao (00:23)Briefly tell us about how you ended up actually working on post-training and open language models. Just a bit about yourself.Nathan Lambert (00:29)Yeah. So I actually started my PhD at Berkeley in 2017, not working on AI things. I was an electrical engineer by training in undergrad, which is funny looking back, because that’s the same year that the Transformer paper came out. And I was like, I think I should do this AI thing, and tried to get the famous advisors to mentor me. And they’re like, we can’t take you. So I had my PhD as this wandering path to become an AI researcher. And then I ended up at Hugging Face after that, which was, realistically, the only industry research job that I had, but also a very hot startup and very fun to learn kind of at the intersection of these tools that people use a lot for AI and research, which is what I was doing.And then when ChatGPT hit, the kind of RLHF thing blew up as the hot word on the technical side of things. My PhD had ended up being in reinforcement learning, which is just the first half of reinforcement learning from human feedback. So it was kind of a natural pivot to be like, well, I might just do that. And Hugging Face was a good place for doing that, because the whole company is kind of all for that, which is like: figure out how to support the community on the hot thing and build platforms there. So they were very happy about that. And I helped build a team at Hugging Face.And then I was kind of burnt out on the remote-work time-zone thing and found out that the Allen Institute was doing such similar stuff. And I was like, wow, I have people that could be in-person friends and do similar things. I was like, quality of life — I need to do this. And a few years later, I ended up building a bunch of models. And I think being at a nonprofit opened me to this ecosystem vacuum of information, where there aren’t many people who can talk about what they’re doing. So then, with some luck and committing to write every week, I just feel like my influence filled the vacuum of nobody saying reasonable things.And it is this nice synergy between what I write about and what I work on in my day job, and it just kind of got bigger and bigger in a very fun way. I think that, generally, at the highest level, I’m motivated by wanting AI to go well on this trajectory. And I worry about a lot of near-term things, whether it’s social unrest in the U.S. and just kind of the massive hatred for AI — I think is a very big near-term problem — and then, medium term, concentration of power, because I think AI will be super powerful in ways that people don’t expect. So generally, open models are a nice way to curb both of them by being a bit more transparent to people, and it naturally is a hedge against concentration of power. There have been different reasons throughout that, but that’s kind of a recurring theme in my life in the last few years.Grace Shao (02:50)Definitely. I love your work because I think you help non-technical people like myself really understand what’s behind what’s happening in these labs a lot better. And then I actually just spoke to your former colleague, Tiejin Wang, and he was with APAC Hugging Face just last week. He was saying the same thing. Open source, in many ways, is kind of the best way to go forward as we know that this technology will not stop evolving, but it’s the best way to kind of put up guardrails and checks and balances for the monopolies.Okay, I don’t want to take up too much time on that side of things today because our focus really is about your China trip. Before we get into the weeds of all that, I want to hear about the trip itself. Most people who are writing about Chinese AI are getting their information secondhand. You really went there, you spent time with the researchers, you met with people who are building the models. Tell us about what you meant when you said you came back with great humility, right? Your eyes are a bit more open, whether it’s the good or the bad. Tell us about your trip.Nathan Lambert (03:50)I feel like I kind of went in — I mean, I had this horrible English phrase in my writing, which was like, “I knew I knew nothing about China,” which kind of tried to indicate that I knew going into the trip that I knew nothing. And it was still the fact in my current writing. This is a horribly written sentence that I had in there. And I only talk about it because somebody called me out on it. It’s like, what is this? And it’s like, leaving, which is knowing that it’s such a big country, there are just such vast amounts of talent working on these problems, and how unpredictable it is as a human to model people with very different worldviews and upbringings and training systems. Realistically, the way that people are trained in China is very different.And I just think that even being there, you can’t fully grasp: what are the pockets of three to six researchers doing that is actually a bit different than in the West, even if they’re working on the same goal? I think you could get down to that level of granularity and a sociological study and actually see differences in what they’re working on, and that’ll always change the output. I didn’t get to that level of granularity, but it’s just to start having real experiences and understanding how people explain how they work on these problems.And for me, realistically, a lot of it is coalition building, which is just like: I want there to not be vitriol at the level of the technical companies doing things in international bodies. So just meeting all the labs on both sides is really nice, because you need to do that for them to talk to you about more sensitive issues in the future. I got some criticism on the piece, which is like, this is how you shouldn’t visit China. And it’s like, well, what are you going to do if you’re going on an official visit to a bunch of companies? How do you expect to get in the door without being nice? You have to start somewhere, and I think it’s important to be respectful.Grace Shao (05:31)I think the piece was, frankly — I don’t think the criticism was fair, to be honest, because I think you were really transparent with the fact that you’re not a China person, right? It’s not like you’re going there and exoticizing everything. And if anything, a lot of people, even with China backgrounds, like to use certain dragons and tigers to describe things. I feel like you actually were really humble going and being like, I’m just a technical dude meeting with these labs, talking about their technical research, right? And then because you were physically there, you had observations of the culture and the people. So yeah, I actually thought your piece was quite good. And yeah, sorry.Nathan Lambert (06:05)I agree. I was willing to let that sail past, but I think it’s important for people who listen to realize how actively these companies are trying to court Western audiences, which is why we could get in the door. I mean, we had some prominent people on this trip, but that’s why we got all of them in the days that we wanted them, except for DeepSeek. So essentially some, like Catherine Rintel, who works with me at Interconnects, and some other creators...Grace Shao (06:23)How did you get everyone? Yeah, how did you get everyone?Nathan Lambert (06:29)He used to live in China and has connections in China. So he kind of orchestrated the mix of his connections and leveraging my connections to labs. We had some bigger names on the trip as well. Just stringing all of these together to get all the various labs in place is a few months of networking to make sure the trip lines up with people with established networks and contacts with the various labs. But these people want to look good to Western audiences, so they’re only going to say yes to the right researchers.And the researchers know that there are two to four comms/ops people in the room, hanging out, making sure that it goes well. Especially the bigger the company, the more comms people. You go to Alibaba and there are three to five various people, from the head of comms to some special offices. You’re not going to get these people in the office, or at all, without accepting the cost of these types of handlers. It’s the same thing in the U.S. You’re not going to just plop a senior executive into a chair.So it’s also good because now I have the WeChats of a bunch of researchers from China that I could just text about things. It’s like, hey, congrats on the new model release. It’s like Lay Lee works at Xiaomi, Xiaomi MiMo. It’s like, talk to this guy for an hour at a mall — I don’t remember the name of the tea store — but it’s like...Grace Shao (07:27)No, of course. No, of course.Nathan Lambert (07:49)Now we have these relationships, which is very useful, and that helps information spread across the ecosystem to these trusted parties, which doesn’t really exist. There are not that many, I think. And the opposite direction of the trip is very hard because Chinese researchers can’t really enter the U.S.; the visa purgatory is too complicated. A lot of us on the trip were either Canadian or entered on a transit-without-visa entry, which makes it very easy for American technical talent to go to China right now, which is why I think there are so many trips. I think there’ll be more of them.We’ve got a lot of inbound from VCs and open-source labs in the U.S. that want to establish collaborations with these various labs because they’re the best open-weight models, and they want to build a stack for companies in the U.S. building open-weight models. So I think there are going to be more prominent, but not gigantic, U.S. startups going to try to build these relationships, which I think is a really interesting technological development because we’ve never seen this type of professional work trip in China from U.S. tech companies. Most tech companies have a “bring a device to China, it auto-bricks itself, and you have to hand it into IT.” So to actually proactively send people in a professional capacity is a really big change. There are a lot of angles you could take this, and I think it’s cool to see how it unfolds. This isn’t even really about the trip. This is the follow-on that we’re hearing from people that are like, hey, how’d you do this? We want to do this trip.Grace Shao (09:06)Yeah, definitely. Actually, from my end, I hear about VCs or investors always being quite active going to China because previously American funds were very, very active during the internet era. People were kind of always trying to find a way to either get into these good deals or potentially keep their pulse on it. But I think it’s really, really positive for the whole AI ecosystem to have this kind of fair, transparent exchange in some capacity. But to your point, there’s no way that star researchers can come out and talk to you off the record without any compliance, because that doesn’t happen in the U.S. either. That’s just companies protecting themselves.I just think your trip was quite meaningful, and I want to bring it back to your observations. You talked a lot about the cultural aspects of it. You talked about how you felt like in China there was less of this star-researcher celebrity status around people. People were more humble, or there was more humility. It was very focused on execution. You argue that Chinese labs are particularly well suited to the current LM-building game because they’re very focused on meticulous stack-level work. And there’s less ego sometimes to work on the dirty work, or the non-sexy work. So kind of unpack that for us. Why do you think that is? You kind of touched on it — you said they were brought up differently, they were taught differently — but what’s so different?Nathan Lambert (10:27)So essentially, an interesting part that synergizes on this trip is that we stopped by some academic institutions. I think it was like AIR and Tsinghua and stuff. And you hear all of these academic leaders talk about how they’re pushing hard to try to change it. So yes, they know China is producing more papers than anyone else, but they still think that it’s not as transformative of research. And they think that they’re trying to cultivate the academic domestic ecosystem to change just the type of work it works on, and the distribution, and take more risk.And then you would talk to some industry leaders off the record behind closed doors, and you would hear things like, it’s never going to change because the education system is so structured. There are so many layers of the funnel that reward things like memorization and stuff that they’re just like, this research culture is not going to emerge. And then the follow-on with the AI labs is that these labs are doing fast-following. They kind of have a proof of concept, and they know what it needs to look like. Therefore, in that domain, you’re not trying to invent the new paradigm. You’re not trying to make the model that is o1 or o3, or the first model to work in Claude Code. You’re like, I see it, and I’m going to try to do that and make it the best thing. And I’m going to try to make it cheaper and just maximize that goal.A lot of companies don’t need to invent the new paradigm. OpenAI has done this so many times. That’s their bread and butter: never doubt OpenAI’s ability to release a blog post and a plot that changes how people think about AI. I still think it’s going to happen a few times in this massive boom over the next four years. OpenAI just kind of has that sense of what is the thing that you can push on a bit earlier and just transform things. But I don’t expect — and other people wouldn’t expect — the Chinese companies to do that as much, because it’s just such a culture of, I guess, building. I don’t know how to describe the positive version of this. Maybe it’s slightly more practical-minded, in terms of: it’s your job to build this thing.A lot of the researchers, maybe because they knew their managers — some of them had managers in the room — see their role in the company as being to make the models excellent. And especially for students, I work with students and that’s what they say. I work at the Allen Institute and we have students that will co-lead our language models. It’s not that surprising, because if you do an industry research job in the U.S., a lot of mentors will tell you that you’re kind of free of the burden of bureaucracy and politics. So the naivety of students, and the simplifying, is actually so good at just getting a lot of technical work done.There’s also the life-stage side. If you’re younger, you don’t have as much family, and you normally haven’t built up as many habits and other things you do with your life. Language models are so complex, and the amount of context that you need to absorb to understand what the bottleneck is — there’s so much information, and you have to be able to pick what the bottleneck is and break it. If you just don’t have the mental space to absorb all the context, you kind of end up doing things that are cute but don’t make breakthroughs on the model.So that’s kind of a difference that I’ve seen in people who were both very successful academically before language models. Some of them are able to pivot to this practical mind, which is: what is the state of the system? How do I improve it? And then some try to make kind of these abstract frames of what’s happening and approach it like an academic, and it normally doesn’t improve the model as much. So I just kind of see, if the academic system is a bit more practical-minded, a bit more structured, and the work you’re doing is structured in the language model — make this kernel implementation faster, make this idea work — then maybe it can be...I think it’s an oversimplification. I push on that a bit in the piece just to really contrast what you could think a U.S. lab would look like. And I have a few anecdotes. I’ve heard a U.S. lab paying off a researcher to be quiet about their thing not being in the model. All of these one-off things are more storytelling devices than anything, because most one-off things don’t matter at all. But also Llama 4 imploded, and that was because it was described as a Game-of-Thrones political-style environment, with all the VPs vying for influence and showing that their thing made the benchmarks go up. It kind of fell. Many, many people will tell you that. And we’ve had the Qwen turnover, but it doesn’t seem like it was quite the same type of thing as Llama 4 or xAI. xAI barely exists now. There have been some dramatic things in the U.S. with how these companies have kind of come and gone out of the fold.Grace Shao (14:55)Yeah, I kind of agree with you, but also I would push back on that. I think there’s obviously a more rigid and competitive academic system, which by default in East Asia results in a culture of students following the bureaucracy and authority a bit more. So I agree with you in the sense that they’re very pragmatic. They focus on the task that is given to them. However, I wonder if things will change with how AI will disrupt education. That’s number one. But also, a lot of the young researchers that you’re working with today seem quite different. At least a lot of the entrepreneurs I meet today are born in the ‘80s and ‘90s, some even younger and born in the 2000s. And I think there’s a kind of aura or confidence coming from them. If anything, you want to say they’re a bit more individualistic-minded. You went to Shanghai, right? They are dressed very, very uniquely. They have these outrageous outfits on the streets. People are seeking individual ways to showcase their personality. So I wonder if that will shift.But for sure, for the academic institutions like the Tsinghua and the Beida of the world, they are still very old-school. But I would say that is the same maybe in some academic institutions in the West still. Okay, I think on this topic we can go off on a tangent on academics, but let’s go back to China’s ecosystem.When DeepSeek V4 came out, we talked about it offline, the two of us, quickly about a piece I wrote saying how DeepSeek is starting to look a bit more like a base layer for China. And if anything, some of the labs kind of admitted to that. They’re like, we have very limited resources. And to your point earlier...Nathan Lambert (16:11)Yeah, you could take that in so many tangents.Grace Shao (16:34)Limited people — these labs are tiny. They’re run by 100 to 200 people max. Limited capital, obviously limited compute. They have constraints all around. And in that sense, in a way, the ecosystem’s looking less like a zero-sum game and more like different players optimizing their own strengths. So correct me if I’m wrong, but DeepSeek is providing a base layer where a lot of labs will quickly follow and basically adopt a lot of their engineering breakthroughs. And then Zhipu, Z.ai, will focus on the coding; MiniMax focusing on the multimodality, et cetera. There are a lot of these different players. ByteDance, obviously, very, very focused on their video models. And Qwen, like you mentioned, had the whole open-source saga break apart with Lin Junyang leaving. But in general, they’re still kind of the leader in hyperscalers on that front. So everyone’s doing their own thing almost, instead of really...Nathan Lambert (17:27)I agree with the people specializing, which I think is normal business evolution. You figure out a bit where you’re good at. And there’s so much opportunity that they are like, okay, I’ll follow this because they see that they’re good at it. I just am more skeptical of DeepSeek as a base because I have no idea what DeepSeek is doing. And some of the labs when we were there, because DeepSeek V4 had just come out, were like, yeah, we look at the things they’re doing, but they seem more intricate than needed. And if you read the paper, there’s just so much going on in this model. As a researcher, I’m like, some of it seems a little fake or a little dependent on their setup and not necessarily going to work in every model.Grace Shao (18:04)What does that mean? Break it down for me.Nathan Lambert (18:18)Essentially, I will say that building an LLM is dependent on where you have your GPUs, your pre-training dataset, your intended deployment setup, and stuff like this. So you make decisions based on your constraints, and you build the model. DeepSeek has these constraints and they end up with their model, but Moonshot and Zhipu have different constraints, maybe more flexibility, and they ended up building a different model. They will test the DeepSeek innovations. So they’ll say things like, X innovation doesn’t improve our model. These two organizations are on different development paths that have core similarities, like these large mixture-of-experts models and the general methods are similar, but a lot of the parts end up being a bit different.That’s why I’m like, I don’t know exactly. If DeepSeek was a base, you would see the Chinese labs just do post-training. We just take the base model that’s out there and we adapt it to our domain of specialty. And we have users that do that, which is something that I think about a lot. I’m thinking about starting a post-training lab and how to format post-training research better. So I think about this a lot. I think about what a shared base actually would be. They go through — some of these labs put an extreme cost on creating their base model. And if they didn’t need to do that, they wouldn’t.One of the labs told us how long their pre-training run was, and my jaw dropped. I was like, that’s way too long. Any U.S. advisor would be like, you’re taking way too much risk on this pre-training run. If they didn’t land that pre-training run from one of these past big MoEs at a Chinese lab, I don’t know if the company’s dead, but that’s a huge amount of time. Most U.S. companies now know that you don’t want your big pre-training run to be more than a few months because it’s just so much risk and time to put all your eggs in that basket.That’s a sign that, in that case, they don’t have as big of a peak-size cluster. Essentially, pre-training time can come down a lot when you have a bigger overall cluster; you can just get more throughput on it. But if your biggest cluster is smaller, it’s harder to get a certain amount of throughput, so you use that one for longer. That’s a compute constraint. To loop it back, I think the specialization is real, but I’m more like, I have no idea what DeepSeek is doing. I know they’re raising money now. I don’t know what the plan is there. They seem the most without a specialty in the Chinese ecosystem.Grace Shao (19:59)Dependency on. Yeah.Mm-hmm.No one knows, though. No one knows. They’re secretive.But that’s my point, right? I feel like they’ve been kind of nationalized, whether willingly or not, because they’re taking the Chinese government’s money. They’ve kind of gone secretive. And it’s not like there’s a secret that they prefer Chinese-educated researchers. They’re keeping a very domestic stack, from talent to capital to the whole stack. So to me, it seems like they’re being Huawei’d, in some ways, because they did well and they got their name globally, and then by default they’re becoming the next Huawei, willingly or not.Nathan Lambert (21:01)I don’t think nationalization makes you a base for the other companies, at least not at this stage. There could be something, but it’s hard to force.Grace Shao (21:06)But then you have some incentive, right? But then it is some incentive. You’re like, well, if you can propel one of the teams and propel the whole industry as a whole, it could be in your KPI or some kind of unspoken expectation.Nathan Lambert (21:17)The coordination problem is so hard. Essentially, both in the U.S. and China, even the open labs, what they do is they fork open-source code and match it to their internals, and every company does this. Therefore, all the improvements that could potentially be going to the open code and forming this base that is far more efficient — they’re not completing the feedback loop. I think China could be closer to it. If people really lean into DeepSeek as a standard architecture and DeepSeek shared their training code and all the specifics and how to do this, from a Chinese economic perspective, that would be a huge win because you’re just saving compute. But I think it’s too decentralized and too competitive to have that happen. It wouldn’t happen in the U.S. either.Grace Shao (22:04)It’s so cutthroat. Yeah.Nathan Lambert (22:08)Even though I think for open models to be closer to the frontier, it would be better. I talk about open models in the U.S. needing a consortium. But there’s definitely enough money to make a consortium in the U.S.; then you fail because the model won’t be good because you’re feeding too many asks into the model. That’s the only way to create a shared base.Grace Shao (22:25)Interesting. So it’s not really just commercial. Yeah. It’s not the commercial reason.Okay. So if you had to give a high-level commentary on each of the major labs, what would it be? If you look at ByteDance, Alibaba, Tencent Hunyuan, if they’re relevant, DeepSeek, Moonshot, Zhipu, MiniMax, Meituan, Xiaomi now being part of the ecosystem too.Nathan Lambert (22:46)You might have to prompt it or say more, but I could just kind of ramble through them, which is kind of fun. Alibaba: cloud-focused, understands that open models can enable more usage of platform. So I would say Alibaba is very, very cloud-focused. ByteDance: mostly characterized by everybody else being intimidated by them, and very user-focused, including multimodal. Kimi: vibes of the office were great. It would be one of the best startup vibes that you would visit among U.S. or China. Zhipu: very AGI-pilled, surprisingly cautiously excited about being entity-listed, even though they have no idea why they are, because they’re like, it stamps them as a big deal. And then there’s some...Grace Shao (23:27)I think they previously worked with SOEs. That’s the main reason. Or they still do, but that was one of their main sources of income. And unfortunately, because a lot of these labs spun out of Tsinghua, and Tsinghua is, for people’s context, in Beijing. It’s really close to the government, obviously. But the thing is, when it’s close to the government, it could mean there are three layers of agency underneath the actual government apparatus. But then people like to link it to the fact that it’s taking government money, so therefore they are suspicious. It’s very unfortunate, I think. A lot of companies get thrown into that category. Even companies like Lenovo and a few other Chinese companies have previously been called out by U.S. senators saying, they’re taking Chinese government money, but really it’s that their scientists or their research labs spun out of a certain government-affiliated or government-funded academic institution. That’s what it is. Anyway, yes, go on.Nathan Lambert (24:23)Yeah. Some more would be: Xiaomi — surprisingly great research vibes for a new team at a random company. They seem to be crushing it.Grace Shao (24:31)What do you think of Luo Fuli? The star researcher.Nathan Lambert (24:31)I didn’t get to meet her. I think she’s as close as they have to a star researcher right now. There’s the tier of star CEO, which there are obviously others — Dario and Sam, the analogies are there — but the star researchers, like the Sholtos of the world in the U.S., obviously you can come up with many more. She’s the closest you have to this. I need to watch more interviews. We’ll see. But she wasn’t in our meeting.But they just seem to be doing the right thing. They’re making general models. They don’t really have specialization yet. Florian, the person who helps me write about open models on Interconnects, and I took a detour to go see Meituan because we’re like, why is Meituan building these models? And they’re very practical about it. It was a less glamorous visit at a normal tech office. It wasn’t an official visit for them. They were like, yeah, we’re a major online platform. We obviously are going to use LLMs everywhere once we need to build our own LLM and specialize it to our products, which, surprise, is very practical-minded. I’m guessing there are many more companies in China like this.Grace Shao (25:39)That’s what Tencent’s saying too. It’s because they want to serve their existing consumers and optimize their LLMs for their own distribution and their own basic interface or activity loop.Nathan Lambert (25:52)Yeah. After I left, some people in the group went to Xiaohongshu, like RedNote, and they’re there. They’ve released some language models that are multimodal. They’re like multimodal data-processing things. So a lot of them are not that surprising. The startups just have different cultures. I have met some MiniMax people before, so I left the trip early before MiniMax on this one. But MiniMax was quirky. They have a ton of women in their company, which was very fun. And they have products. They’re maybe slightly more product-focused, but I feel like the quirkiness of the company kind of matches maybe Western confusion over what their products are doing and what they’re trying to do. But it kind of matches their language models that are a bit more efficient.Grace Shao (26:35)Well, they came out with a lot of very consumer-focused applications, right? They had Hailuo and Talkie, all these character companion-bot products before.Nathan Lambert (26:45)Yeah. And then the last one I went to was Ant Ling, which is also very corporate, but in a less intense way, because I think they see it as serving their own products, whereas Alibaba Cloud is like, this is the gold mine we have to win. It’s a much bigger deal for them than Ant Group. But a lot of these things, when you list them — I don’t know, eight to 10 companies — they’re all pretty reasonable with respect to the age of the company and what the company does best. There’s not as much confusion.Grace Shao (27:14)Yeah. And Ant is low-key best at medical chatbots right now, which I guess makes sense because everyone has access to Alipay. And then for seniors, apart from WeChat, it might be the only application they’re using on a regular basis. So it became the default medical consultation app, which is really random, but it’s their niche now. Yeah, I think you’re pretty spot-on. It’s pretty cool that you got those takeaways, even just meeting with them for a couple hours.Nathan Lambert (27:41)I have been reading about them for so long, so a lot of these priors are easy to confirm when they kind of fit with things you have seen. The Chinese showroom culture is so interesting, and also one of the most surprising things to have at software companies. It’s so funny. They’re definitely appealing to Western audiences. Z.ai had poorly translated merch. What was it? Something so — it would be borderline inappropriate translation in the U.S. It was like “ship big, go hard,” or something. Just some really weird translations. And they have live API statistics in their showroom. So Z.ai was like, we’re serving 5.5 trillion tokens a day. All the U.S. companies are so closely watched for when they announce token statistics.I know at least one of these numbers is wrong. It’s something like Fireworks does either 30 or 300 trillion tokens a day — or I meant Together for that one — and then one of Fireworks or Together, and the other one, are like 100 trillion tokens a day. Don’t take these as sourced; go look them up. There were some public announcements recently, but those were the first updates that anyone has on major infra companies in the U.S. Inference is a huge market. You don’t hear anything from Fireworks because they’re just struggling to demand and they’re making bank, because inference is a much better thing to sell than bare metal.Essentially, inference is selling the software implementation to serve tokens more efficiently, and you can just get more margin when you improve the stack for a fixed model. So a model comes out and you host it, and then you can make your stack more and more efficient on that model. You just get more margin and hopefully growing usage. That’s way different than GPUs, where the best case is that you lock in a huge commitment for a long term.Just being able to walk into an office and learn about their API is interesting because they also had geographic distribution, which was like: China was, I don’t know, two-thirds; U.S.A., 20%; and then the last percent was Singapore, Korea, Japan on the Z.ai API. So that’s cool. This is s**t that I always want to know about the companies, and I have no idea. One of the things I always want to know is: how are open models being used outside of the U.S. and China, and has this decades-long process of technological diffusion started to kick in in a way that any company can measure? I don’t think anyone has good data on it yet, but I think it’s obvious that at some point, open models that are cheap to run are going to have some interesting playbook across the globe for the long tail of countries. Maybe I’ll just walk into the front door of a Chinese open-weight company and get my answer.Grace Shao (30:30)But actually, I think the culture of these labs — a lot of them, because they’re run by really young, passionate people — you would feel like they’re a lot less commercialized or less corporate, or at least less sleek. They’re not sophisticated with, you can say, the capital-market side of things, but you can also say that they’re just really naive and open-minded and passionate about the product they’re working on, with less of a corporate guardrail built around them.Nathan Lambert (30:56)Yeah.Grace Shao (30:57)Okay, I want to talk about...Nathan Lambert (30:57)Yeah, go ahead. It’s like one of the people at Z.ai who’s known on X — I don’t know, 9,000 followers — it’s like Lu. She came up and was like, hi, I’m a student, I’m 20. I’m Lu from X. And I was like, that’s hilarious. There was a lot of s**t like that. It was like, oh, okay. I don’t want to call her a kid, but it’s like...Grace Shao (31:06)Yeah, yeah, yeah. And I think the one that runs Moonshot’s developer ecosystem or something is literally a girl fresh out of school, right? And she just posts hilarious memes all day long. There’s no filter on her social media. It’s funny.Okay, we go on these tangents, Nathan. We need to come back on track. Open source, open weight. Why? Why do you think Chinese labs are adopting it or embracing it, however you want to put it, especially after visiting them? Is it because they simply have to, because of what we talked about — they are leaning on each other because of all the constraints they have? Or do you think the philosophical drive is actually bigger in that ecosystem? Or is this a bigger strategic thinking for diffusion in the long run?Nathan Lambert (31:54)I actually don’t feel like it’s that special ideologically. I think it’s easy to say the ideological line when you are doing it. Now you can look at Zuckerberg: he said the ideological line when he was doing it, and then he stopped. I think it’s mostly just that, for one, distributing within the U.S. ecosystem, especially to enterprises, is the highest-value market, and they can’t sign many enterprise deals. And the closest best thing is things like Cursor adopting Kimi’s model. Even if Kimi doesn’t get paid for that, they’re happy. That’s the biggest sign of credibility for them, and they can figure it out in selling tokens or whatever in the future.Practically speaking, one, the only way to influence the U.S. market is by releasing these models. And two, it seems like they don’t feel like they’re losing as much if they release and share things. If the model was closed, they just think they would get less influence, they would be seen less, fewer people would use the model, their actual paid offerings would be adopted less. It just seems almost overwhelmingly obvious, because there are all these benefits and not as obvious of a drawback. There will always be better models, and just keep going. But I think every scientist loves...Grace Shao (33:08)Then why are so many U.S. labs against it, or not willing to?Nathan Lambert (33:12)Because they can make as much money without it. Anthropic and OpenAI make more money by not releasing them. They can just make so much money, so why bother thinking about an open model that doesn’t make money? There are different scales of influence. Same with Google. Google’s making so much money. I think Meta will make a lot of money by having good AI models in their products, if they get their act together. Even Google could release more models. They have so many surfaces other than Gemini that need AI to be commoditized and used, like the cloud and all of this. Meta could release the models. It’s just not worth the effort for some of them. They’re like, we need to do this high revenue target; it’s too much of a pain to go through legal and make it ready to release. Why bother?I don’t know, maybe it’s a little bit of a cynical take, but I think Microsoft and Meta could release their best models openly because they benefit if it’s a commodity layer. But I don’t expect them to, because it’s just kind of like the benefits of focus are so high, and they just kind of see it as something they don’t have to do.Grace Shao (33:56)And it’ll be good for them. But then eventually, we will see some consolidation in the market as well, assuming — because you can’t really have 10 labs in each dominant country right now all exist.Nathan Lambert (34:26)I do expect consolidation. I think this is potentially a subtle cultural point, which is that the U.S. labs are more likely to buy into “we’re special, we need to go fast, keep it closed,” and the Chinese labs are not. There could be something there. That’s also who the decisions funnel up to. I don’t know. I talked to the Alibaba people that make these decisions. I can’t say all the things that they say about them. Some of these were two-on-one and off the record, so I can’t say all these things. But at all the other labs, there is a person that makes the call, I’m guessing. I think those are senior leadership that we’re not talking to. So it’s kind of hard to know exactly what they really think.I definitely expect consolidation. My thing is that I expected it in China faster because the capital markets aren’t as strong as in the U.S., but I don’t have a model for that. I think you can model it, which is: what do you think the revenue growth would be? What do they need to do to raise to keep training bigger models? What is the compute cost? Then you look at the potential raises and think about which country would not be able to do that race first. But also, it’s this wild thing with OpenAI raising $120 billion. Are you kidding me? What is that?Grace Shao (35:47)Yeah, the valuations in the U.S. are not really understandable by anyone else right now. I think in China — so on your point on that, I’ve been writing about this and I think it would make sense for Tencent just to buy out one of the labs. They have the money, they need the capabilities, and frankly, they’ve really been struggling to compete with their LLMs, with all the labs talked about just now. So my...Nathan Lambert (36:05)Their licenses are so bad. They release all these models that have horrible licenses. They’re not that good, and the licenses are just horrible.Grace Shao (36:13)So I feel like it financially makes sense for a company like that to optimize and just buy out a lab. Then the labs can also lean on their distribution, because at the end of the day, how are they going to win consumer mindshare or distribution in China right now when it’s really just dominated by Alibaba, ByteDance, and Tencent? That’s my spiel. But when I spoke to some of the researchers...Nathan Lambert (36:33)I think big companies have a lot of inertia, and the senior leadership has the call, and they can have inertia. I still think Apple just ends up buying some lab for $25 to $50 billion. It’s not the worst thing. Just golden-handcuff the researchers. Some will still quit.Grace Shao (36:43)Yeah. But I think right now they don’t want to. The labs still have a dream. Some of the researchers still have a dream. So when I spoke to a lot of them, they’re like, no, we don’t want to do that. We want to commit to our own frontier research. If I wanted to join one of the big tech companies, I could have. So why would I want to sell? That’s what the researchers think. But to your point, we don’t know what actually the one person or two people at the very top think, especially if they continue to have hurdles with compute access and capital access, which brings me to the question.Nathan Lambert (37:14)It also depends on your view of inference. You can ask your next question. I don’t need to cut you off. It depends on your view of inference. If these agents are just so much inference, I do think it’s going to be an oligopoly-style market, not a monopoly-style market. And what’s the difference financially between two and four or five big companies with great models? Is that actually not sustainable if there’s so much demand? There are a lot of cases where we have two or three, like the cloud, but what’s stopping that from being four?Grace Shao (37:40)I think they will be the infrastructure providers. Yeah, yeah. And they would kind of lean into each of their existing ecosystems or distribution, whatever you want to call it, and serve certain specific models for specific uses. So enterprises can choose what matches their needs the best as well.I do want to bring the conversation to a more contentious topic, which is on distillation and model convergence. You raise the question of whether Chinese models are structurally different. Often we are hearing claims saying a lot of these labs are about three to six months or six to nine months behind U.S. labs. There’s obviously a lot of noise or allegations and accusations from certain U.S. labs saying Chinese labs are distilling them. How do you actually see that accusation or that kind of dynamic?Nathan Lambert (38:36)The biggest unknown that I don’t have an answer to, which actually has a lot of sway, is how much of the Chinese companies are actively trying to hack APIs versus just showing up as a customer and paying. If you’re trying to hack the APIs, normally you get reasoning traces out so that you can create a reasoning foundation that would be similar to the model that you’re trying to do this from. That’s very different than the API standard form, which is just the output of the model, which is a less direct process for learning from.I don’t know the magnitudes. If it’s more just like, I walk up to an Anthropic API and I use it as intended, but I’m making a competitive model, I’m not very sympathetic to Anthropic. They could ban it if they want to. And I think the impacts are kind of a standard practice. You can do it with many different models and so on. The evidence Anthropic provided is not large enough scale where I’m like, this is industry IP theft at mass scale going on 24/7/365. So there’s definitely some gray area to what is actually happening in distillation.That’s why, on the policy side, I try to push people to not call all of it the same thing. Essentially, using any API endpoint to make synthetic data to train your model is some form of distillation, but it’s very different if you’re trying to break this model so that it gives us a different behavior that is hyper-useful for training and not get caught. Those are pretty different actions, and they’re all looped into this common phrase of “distillation” right now. That’s my biggest problem, which is that academic researchers and small companies use distillation extensively as the core of their business and the core of research methods. So if the U.S. government nukes that as a thing that could be done in the AI ecosystem, it’s mostly bad for small players, bad for U.S.-China tensions, and bad for academics. That’s my primary concern.And then trying to get the labs to actually say more. There’s a distillation side and then performance is the other side, which on benchmarks, it does seem like the Chinese labs tend to be six to nine months behind. When it comes to general use, I’ve always found the closed models to be better in ways that are hard to measure. So I go very back and forth on whether the closed models are better. I think we will especially see Anthropic and OpenAI pull ahead on knowledge-work tasks like legal, healthcare, financial services, because I just don’t see the Chinese labs paying for that data. All that data is going to be people that charge hundreds of dollars an hour to annotate and create these environments. So it’s a whole new capital build-out that goes on there right now. It’s going to be billions of dollars if you’re going to buy a billion dollars of data and a billion dollars of compute and a billion dollars of talent to train your model.Grace Shao (41:30)They don’t have the money.Nathan Lambert (41:30)I don’t think they have that. Mercor has some of these evals, and I think there is a bigger gap there. So it’s very interesting. Florian, the guy that helps me, and I disagree on it. It’s this fine line between, yes, the evals — coding and lots of these things, and even random evals that surely the Chinese labs aren’t training on — the open models really are genuinely crazy impressive scores. So I think there’s also a tester’s bias, where I don’t use the open models as much. Maybe it’s hard to ground in my head what I was doing with AI six to nine months ago. I wasn’t even using Claude Code as extensively.I guess the question is, at the end of this year, can I use an open model in something like Claude Code and feel like it works at all? That’s the test on the performance gap, starting in June, June to August, and whether or not that hits. I don’t think the open models have hit that yet. I think it would be way more of a narrative if all the companies spending billions of dollars on Claude are like, oh, we can spend 1% and just use DeepSeek. These CIOs and all the big companies — some companies spend more on tokens for their employees than on headcount. These are normally startups. But they would happily reduce that token cost to 1% expenditure if it really was that similar, because then you could just use 10x the tokens. I don’t expect that to happen. And I expect things like the latest Claude and GPT-5.5. I expect more of these things through the year, and we’ll see if I end up being right. Both are right at the middle of us, as a world, getting more clarity on them. They’re like 18-month-long stories unfolding, and I feel like we’re just in the middle of performance gap and distillation and learning more.Grace Shao (43:25)Yeah, it’s interesting. You mentioned — it helped me recall a conversation I had with other people as well. The point on distillation is that I just had a conversation with your former colleague at Hugging Face, who leads APAC, called Tiejin Wang. He was just saying, look, the distillation accusations don’t really make sense because we’re all distilling off of each other as we speak. I’m learning from you; you learn from me. We’re distilling. It’s so vague of a terminology to just use that to accuse all these various behaviors.So to your point, I think people in the technical world who understand what’s happening actually want more clarity on what is the gray area, what is actually black and white, and what is not appropriate or unethical. That needs, I think, the industry to come together to really put guardrails and rules around.Now, number two on the compute side and the data side. Something anecdotally will be interesting to you is that when I spoke to one of the lab researchers in Beijing, I think in February around Chinese New Year, they were saying, look, they want to get better data, but they can’t because usually a lot of American labs would pay tens of millions, if not even more, like a hundred million dollars, for a set of very obscure or niche datasets, but they would have an exclusivity contract. What the Chinese labs will do is that they will literally wait out the exclusivity contract and then, say two or three months later, pay for it at one-tenth or one-twentieth of the price for that same dataset. So then once they start post-training on that dataset, that’s where the three to six months or six to nine months come in as well. Yeah. On that note, I want to...Nathan Lambert (45:00)Yeah. I think the data industry in the U.S. has two things. One, the lab asks the data vendor, we need this specific type of data. And the data vendor is a network that connects the people to the lab. The other thing is the data vendors know evals that are important, so they try to create good data for hill-climbing on specific evals. That data could be sold to multiple people, but is less expensive because they make it once and expect to eat margin or take margin on it. There could be a pipeline where once OpenAI is at the cutting edge, creates this new thing, they create deep research, then the data industry is like, let’s make things that are a little bit cheaper to sell. So there is time lag in these various things.But I heard the same thing on the ground, where they have a negative view of the data industry. It’s like, quality is bad, we don’t really have access, we do some in-house. That’s a very big difference from today, which is that you have the data companies in the U.S., which is insane.Grace Shao (45:53)Yeah, the American data companies are so mature. It’s its own sophisticated ecosystem.Before we get into data, I actually want to ask you this question. I think recently a lot of the narrative is now saying, look, Anthropic and OpenAI have kind of proven that pre-training scaling laws continue to hold, especially with the recent models. There’s an obvious compute constraint on the China side that we talked about. And then it will likely be even more amplified with the absence of Blackwells in the coming months.So as we move forward in this race, per se, if you have to put it in China versus U.S. in that sense, will we see a wider gap between the performance and benchmarks between the Chinese labs and U.S. labs? As in, will we see the gap going to 12 months, 24 months, as Chinese labs are very, very constrained on compute for pre-training breakthroughs?Nathan Lambert (46:40)I think it’s more of pre-training as a thing that you could actually finish. How big can you pre-train a model that you can finish and serve? The Chinese labs could train models that look like GPT-4.5, which is this giant model, but you can’t serve it. They end up training a model that is 2.5 trillion parameters and they release it, and no one can use it. They could barely serve it on their API because they don’t have Blackwell NVL72 racks or something — these racks that are definitely what are serving these large MoE models. They just don’t have the quantity of these.So there’s a difference between models that you can build and models that are actually useful. I think some of the Chinese labs are definitely like, we don’t need to release the gigantic models because nobody is going to use them in open weight. The biggest models end up getting served via API. So there might be some segmentation in that market. But I do think the inference and amount of economic resources that you have to serve your customers is becoming a thing that dictates what models are built. That’s why I think the gap will continue to rise. All signs point to GPT-5.5 being a bigger model, and I don’t expect that to stop.And then the economics of it is just the basics of: you need a certain volume to have the margin to support the research, because you can’t keep raising these ridiculous rounds forever. I think OpenAI, Anthropic, and Google are the only people with that AI usage volume to keep marching down the scaling laws to another 10x of training compute, which is mind-boggling amounts of investment in a model. That’s why, when the economic markets slow for fundraising, the model gap between these big three will just show a lot more. That’s the distilled way to say my prediction of when things will look different. It’s like these labs can’t fundraise, they go public, they can’t generate revenue more on their paid services, and then it’s just: look at how much training compute can be allocated or can’t be allocated.Grace Shao (48:41)Yeah. Basically, we’ll see a bigger gap, I think, in the coming months. Then what can make up for that? Domestic chips, or, like you said, better data. And why is it that sometimes people assume China has a very strong data ecosystem or data products, but actually the data vendor ecosystem is very weak in China?Nathan Lambert (48:41)So generally, I think I agree with what you said. I don’t know on the data side, but the way domestic chips could help is that if Huawei chips are fine for inference, and if they have sufficient volume to support the inference economics, which then trickles back into revenue, my read is that they just don’t have the volume of the chips, especially spread out across the amount of companies that they have. Essentially, the total FLOPs of Huawei, all the things produced, and it’s going to all these different places — it’s just not big enough.It could be something like ByteDance and Alibaba, with offshore data centers, can keep up a lot longer because they have access to Nvidia compute and have for a long time through this kind of offshoring. Maybe that stabilizes the ecosystem, and we’ll see what the AI startup, the younger startups like Kimi and Z.ai, end up doing. No one wants to do this, but if they pool resources, they last an extra year. You get another order of magnitude if they all pool together, but I don’t see them doing it.Grace Shao (50:00)But that’s the thing we were just talking about, right? MiniMax and Zhipu, how can they possibly compete with the hyperscalers at this point if you need offshore data centers? And the fact that Zhipu is on the Entity List doesn’t help, right? It’s not going to be easy for them to access these data centers either.Nathan Lambert (50:12)Yeah, I think they can’t. I think they won’t. Human nature will make it so they won’t collaborate. They’ll just do something smaller. They’ll just have successful businesses that are different.Grace Shao (50:22)They just have smaller ambitions, want a smaller piece of the pie. Yeah.Okay, so you wrote something like, nothing’s a secret, but everyone wants Nvidia chips. They want it, they don’t know how to get it, they’re fighting over it.Nathan Lambert (50:34)Yeah. They’re the only thing that works for training. All the models are trained on Nvidia. I don’t believe the DeepSeek propaganda that it’s trained on Huawei. The only models that are trained on Huawei are tiny. Inference on Huawei works. Every lab is like, inference on Huawei works. The labs that don’t have meaningful inference are like, we are told to get Huawei, so we buy them, but we don’t use them. Earlier research labs are like, we don’t have any inference and we don’t have a need for Huawei. Any company that has meaningful use of their models has figured out how to run them on Huawei for inference, which, to Jensen’s credit, is like — it’s happening when he said it was going to happen, but it’s not that surprising.Grace Shao (51:11)Yeah, the Dwarkesh interview. I don’t actually understand why he got so much hate for it because even without your political stance, what he said actually made sense logically by saying, if you don’t sell them the crappier versions of what we have, they will have an equally quite crappy version to serve themselves, or they would just want...Nathan Lambert (51:28)I think they would buy both. Buying both is actually true. The amount of Nvidia chips that you would have to sell to China for them to stop buying Huawei — because Huawei is almost surely way cheaper because Nvidia margins are insane — when would they actually stop buying both?Grace Shao (51:43)But then you have to go on CANN. You have to reroute everything back on CANN. The developer ecosystem is not there. That’s Jensen’s point, right? Or the habits are not there. So I think that’s what, when I talked to a research lab...Nathan Lambert (51:50)Yeah. But I’m saying they would also use Huawei. I think they are so supply-limited, they would use both. Anthropic uses everything. A lot of companies in the U.S. will use multi-platform. Meta is a huge buyer of AMD. Demand is so high that any chip that is potentially viable on the models within a few generations is very valuable. And the fact that you can run some reasonably large model on any Huawei chip is a big line crossed for Huawei.I don’t know if they can produce the volume of chips and scale that quickly, especially as they try to move to lower nodes. That’s the standard semi debate. But the question is: can Huawei scale production? That’s the only question. And if Huawei can manage to scale production, Jensen will just look really right. If Huawei can’t scale production, Jensen will look a little bit like a lunatic, but it will be outside of his hands.Grace Shao (52:43)And we don’t really know what happened during this trip. It seemed like nothing really substantial happened after this big Trump delegation. It was more like a high-profile tourism trip versus an actual deal trip.Okay, I want to ask you something you wrote about that’s a bit niche, not something you usually write about. It’s on the SaaS side of things. You said that there’s a common argument that China struggled to monetize AI because they’re unwilling to pay for enterprise software. We looked at how China tries to monetize on consumer AI, but clearly that’s not really been proven yet. In your piece, you push back on the claim and say that there’s a distinction between SaaS spend and cloud or inference spend.Tell us about what you think about that ecosystem and how Chinese AI labs are trying to make money maybe a bit differently from American AI labs.Nathan Lambert (53:32)I don’t know if it’s necessarily different, but I ask a lot of researchers about this. They say that everybody is trying the new AI tools when they come out. If they don’t like them, they stop using them. If they like them, they keep using them on the consumer side. So something like Claude Code would be an example: tons of people tried it. I’m guessing lots of them churn in China, just like in the U.S., but consumers are very quick to adopt and try new things, but won’t stick if it’s not actually serving them.And then the enterprise is like: there’s definitely cloud that exists. Digital services are gigantic. They essentially think that there’s more runway for making money on AI models that falls into that. And they all use coding agents; they all use Claude. It’s a hilarious thing. They’re all very Claude-pilled. There’s almost no mention of Codex, where in the Western media, Claude versus Codex is this whole thing. They all use Claude. And that is obviously a paid service. So I think there are cracks in the argument, and I expect AI models to be seen as a bit of cloud, but potentially it is the thing that changes some of the expectations, where it’s just so transformative because they’re so competitive, and it could be seen as a bit of a phase shift.Grace Shao (54:41)Yeah, and I think it’s a generational shift, a phase shift. Also, actually, recently Doubao raised their prices on Seedance usage and whatnot, and it’s a shift into trying to capture the prosumer market. You can say the average uncle and auntie on the streets still don’t want to pay for a consumer app, but I think there’s more prosumer market share that could be captured in China, maybe not fully enterprise either.I want to ask you about government roles and geopolitics. I know there is a common narrative that usually people assume Chinese AI labs are heavily subsidized. Actually, when I was in San Fran in March, I was at a dinner with a couple of investors, mostly public investors, and one guy asked me, “Hey, are all labs just basically subsidized by the government?” I was like, definitely not. The majority of them are not. If not, they frankly don’t want to take money from the government.It was really hard for him to understand that, because I think the misconception is all Chinese labs or Chinese tech are just funded by the government. Kind of to our point earlier, where any affiliation to any government agency, just by default, is assumed to be therefore backed. First of all, the government, I don’t even know if they have that much money to give out. Number two, I don’t think that’s how competition works, right? So what’s your thought on all of this?Nathan Lambert (55:55)It seemed more like a provincial government trying to help the companies do stuff, which is like get offices, get talent. I don’t know what the provincial government can do. In Beijing, there’s Beijing Academy for AI or whatever, which is a real research institute that’s just funded by a certain neighborhood in Beijing. It was like, okay, the U.S. could do that. But much less of the Ant Group-style thing, which is government takes major ownership stake in an investment round and goes on. Maybe Kimi’s latest round, there were mentions of government-backed VCs, and I don’t know how that kind of intermediary works. So I still think it’s very indirect. And because the government system is so competitive across the different layers, each of those layers are competing to help the companies, but they don’t have piles of cash sitting around to buy GPUs.Grace Shao (56:44)No, they don’t. And they frankly don’t know what they’re doing half the time. This is an argument like what you said: Haidian District or Chaoyang District of Beijing will be funding an academy, and the academy will be in the effort to help AI go toward AGI. But the reality is they’re trying to follow this high-level KPI of being like, let’s make AI happen. All they want to do is write in their report and say, we funded something about AI, so we’ve hit our quota. I don’t think it’s as hands-on as people assume.Nathan Lambert (57:09)Yeah. If you read Breakneck — most U.S. tech people haven’t read Apple in China and Breakneck — and all you need to do is read these books and learn a little bit about the interface between tech and China and understand that they are also hyped about AI, and then you’ll understand that it’s a messy trickle-down process in the Chinese government. It would be very obvious if they were nationalizing a lab. It would be as obvious as if it was in the U.S. It has not happened.Grace Shao (57:18)Both great books.Yeah. So to close, what is the biggest disconnect between how the U.S. AI ecosystem right now thinks of China, Chinese AI, and what you saw on the ground? And what is something you think we didn’t touch on today that you want to share?Nathan Lambert (57:52)This is the thing that everybody asked me. They normally asked me the first thing when I got off the plane: what’s the big thing? And it’s like, I don’t think there’s anything that shocking. I think that many people just haven’t read basic books about how tech is interfaced with the government, and know these things, or hear narratives that are very geopolitical, which is targeting the top end of the government system and how that in the U.S. engages. And there’s a lot of shrapnel from that. Anthropic pushes very aggressive China narratives, and Anthropic is a very followed company in tech.Most people don’t spend the time on this in the U.S. ecosystem and just don’t go deep on it. I don’t have anything shocking. It’s good to encourage people to do some of that because these dynamics impact things like: the Chinese open models are really influential, and now Silicon Valley is building AI. So it matters to a lot of people, but they don’t study the causes of why they might do this. They just are like, it’s here, I don’t need to think about China.Grace Shao (59:00)Yeah, and I don’t know what it is. You and Bill Gurley were saying this in a couple of his public appearances. It seems like Chinese researchers, tech people, CEOs, whoever, are a lot more aware of or following more closely U.S. leaders and thought leaders, tech leaders, business leaders, than vice versa. There’s something about that. I don’t know if it’s just easier to dismiss it or easier to not have to learn something new. But the goal of AI...Nathan Lambert (1:00:27)I think it’s American culture. American culture is very obsessed with its own weird world. Yeah, it’s hilarious. American culture is ridiculous. It’s so ridiculous.Grace Shao (1:00:27)As a Canadian, I can’t say things like that. You said it. I’ve lived in the States, but I can’t say this. But I think, look, shameless self-plug here: as a Chinese Canadian, my life goal here with AI Pro is really just to bridge that gap. I think to your point, there’s going to be geopolitical narratives and rhetoric at the very top, but for the average person, or even for builders, tech people, whatnot, it’s probably in everyone’s benefit to understand what’s happening on the other side and stop alienating it or stop making it as if it’s so different.I think throughout this conversation, it’s really just to say, look, so much of it is so similar, but so much of it is slightly different. The difference is not really a government mandate versus maybe a cultural difference or resource-constraint difference, especially in building technology. But that’s kind of my view.One last question for you, which is a question I ask everyone on the show. What is one differentiated view you hold? Throw me something crazy.Nathan Lambert (1:01:28)I know, I’ve always kind of been open-models doomer, even though I build on them. It’s just that it’s so unsustainable, and there’s so much money to be made with building closed software, that I’m constantly doomy about the prospects of open models. I’m always a skeptic.Grace Shao (1:01:40)It is a bit sad, isn’t it? How does a company like Hugging Face actually make money?Nathan Lambert (1:01:45)I don’t know. You can look into how much money they actually make. It’s not very much, unfortunately.Grace Shao (1:01:48)Yeah, I think that’s the unfortunate reality of the capitalist world we live in. As much as it incentivizes the competition and breakthroughs, that doesn’t help with what we just talked about earlier. Yeah.All right, Nathan. Thank you so much for your time. Really appreciate your insights and your sharing.Nathan Lambert (1:02:04)Yeah, thanks for having me. Good to see you.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• AI x education, a contentious but unavoidable future. Designing tech for children with Dex's Reni Cao 18.05.2026 1h

  I spoke with Reni Cao, the CEO and co-founder of Dex. Dex Camera is a language-learning camera for kids. Reni is a dad, a former product lead at YouTube, and on a mission to build technology that does good for kids and gives digital autonomy back to parents. We dive into his personal story from his high school days that drives his passion for AI, and why he believes the current education system is a “cookie-cutter” that fails curious kids.We get really into the nitty-gritty of what makes “good” tech versus “bad” tech for kids and why the category of ‘children-first tech’ is very overlooked. Reni explains why most children’s apps are built on an “attention economy” model that forces them to compete with addictive content, and why his team needed to build physical hardware to break that cycle.We tackle the hard questions, including the pushback from parents who believe in “no tech” childhoods. And he shared his most non-consensus view: that the era of standardized, industrial education is over. He believes we are entering a golden age of “scaled homeschooling” where AI meets kids where they are. Whether you’re a tech investor or an anxious parent, this conversation about nature versus nurture, “nei juan” (involution), and raising resilient humans in an AI world is a must-listen.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, founders, and product managers.For more information on the podcast series, see here.To find the previous episodes of Differentiated Understanding, see here.Chapters00:00 Reni’s Journey to Dex Camera03:48 Designing for Children: Principles and Insights08:05 Technology’s Impact on Child Development12:09 Bridging the Gap: Business and Product Design15:36 The Role of Parents in Tech Development25:20 Leveraging AI and Language Models29:48 Value-Driven Pricing Strategy32:05 Defining the Product Category34:33 Subscription Models and Content Delivery37:58 AI and Parenting: Balancing Technology and Safety43:29 Unexpected Use Cases and Impact47:29 Personalized Education and Parenting PhilosophyAI-generated TranscriptGrace Shao (00:00)Reni let’s start with your personal story. Who are you and who are your team members? Because when I met you in SF, I was so enamored by the product and I thought your story was so interesting. So please share that.Reni Cao (00:11)Hi everyone, my name is Renny, CEO and co-founder of Dex. We’re a technology company in San Francisco, almost all parent company, which is pretty special in a startup setting. We’re a bunch of parents that having trouble with the same kind of like a reality where like our education system is a sort of like cookie cutter and our entertainment is also cookie cutter for children.So we’re like, can we harness technology, especially the latest development of the AI, in different way for families that really gives children a chance to become the best version of themselves and ⁓ give the digital autonomy back to parents themselves rather than accepting the fact that they have to struggle between technology versus no technology. So yeah, we’re the parents, of like a bunch of missionaries in this journey together to explore how can we make the best use.of the AI and our first product is called Dexta Language Learning Camera where kids can take pictures and turn the whole world into language immersions. And it’s a product targeting young children three to eight. And we’ve sold 10,000 pieces so far and ⁓ ratings has been high and we’re pretty excited about this. But yeah, this is pretty much about us.Grace Shao (01:24)But Reni, tell us a bit about what you did before Dex actually. What kind of led you to this path? I know becoming a parent really did inspire you. You have a young daughter, I think similar age to mine, around three years old. But before that, what really led you to this path? Were you always passionate about children’s tech or education?Reni Cao (01:41)I actually have been a product management guy for the last decade in Silicon Valley, some big companies like YouTube and LinkedIn, some smaller s***, ZFS, Wish. But I have been a builder since the beginning. I would actually say that my passion for decks actually originated much earlier than I started my career. It actually started right when I was at school, but happy to say more if you’re interested.Grace Shao (02:07)Yeah, no, do tell us a personal story there.Reni Cao (02:09)So I was always this random kid with tons of questions back in high school. And very unfortunately, I think the education system, especially in East Asian countries, is not designed for meet kids where they are. So every time when I come up with a random question, my teachers are usually a little bit impatient and will be like, can you just go back and finish your quiz, et cetera, et cetera.So the moment I saw when GPT-4 comes out, I was thrilled and I posted a long like blurb on LinkedIn. Basically saying like, you know, if I had, have this as a kid, I would have grown into a more complete human. So this kind of like, I feel like this like generative AI’s capability to meet kids where they are, especially meets your needs for curiosity. It’s game changing. So.I feel like I’m building this product first and foremost for a younger me that could have benefited so much from this. That’s pretty much the story about me. yeah, I know we see it and of course our parents right now we see there is a tectonic shift in terms of the skill landscape and what the future of workforce is going to be and even the existential challenge of what does human mean in a future society.So we do want to build something that’s centered around children, centered around the family to help them find what they love and build agencies around it at the end of the day. So yeah, that’s the two main driving force of me coming to Dex. But I would be honest about it. It’s like very random. When I want to start a company, a lot of my colleagues are very surprised, being like, oh my god, Renny, you’re getting into this field. But yeah, I guess I finally find the work of my life.Grace Shao (03:48)I love it. think you need to understand the passion and the personal reason behind the businesses to really understand why the design was frankly so intuitive and why you’re so passionate about building this and leaving such a comfy, know, like cushy corporate role. I think that’s the one thing that stuck out to me. The product itself is actually so natural to how children behave to your point, like my three year old.from morning to night, know, morning she wakes up, it’s like, mommy, what’s this? What’s this? What’s this? What’s this? How do say this? Why do you know that? Sometimes she gets angry at me. If I don’t know something, she’d be like, but you’re an adult, you should know everything. But the reality, especially with languages, it’s really difficult. So for example, yesterday she was coming back from her Mandarin class and she said, liu shu, she was pointing at random tree. And I was like, that’s not liu shu. All I know is not liu shu, but I actually don’t know what liu shu is in English because I think it’s only really common in mainland.I’ve never seen that kind of tree. Well, I guess it’s a willow tree. You don’t see it very commonly elsewhere. And then she kept on pointing at trees, but in Hong Kong, you clearly don’t have liu shu because Hong Kong is like tropical. And then she got really, really mad at me. And that moment I was like, wow, if we had a Dex camera, that would have been perfect. But I was literally trying to take a picture of it while we’re moving car and try to upload it to GBTB, like what tree is this? What’s the name of it? So anyway, I think it’s really great product design. And I want to kind of get into that a little bit.When you were designing it, what was the thinking? Like, what does it mean to be children first?Reni Cao (05:10)I think there are three layers of children first as a principle. The first layer we already touched upon that. So young children, their hand anxiety is very different from adults.they tend to use one hand to operate a device and another hand they want to use for sensory explorations, like they want to touch. Sometimes they want to just move things around. So this requires a different form factor that one handed use, very tactile, very intuitive for young children such that they can explore a world while harnessing the power of AI in this case. So this is kind of like the user, the special things about the user.And it’s a different design. think that’s layer number one. I think the layer number two is also that the device itself is a metaphor for the market as well. And in the market, we want to build something that’s drastically from the so-called adult-centric smart devices, namely the phones and tablets, to send the market a message that there could be a different option. There could be a good technology. There could be a family-centric technology. And we’ve picked this form factorutilizing the metaphor of magnifying glass. It is something you use to see some hidden wonders, otherwise you cannot see. I do think that’s the ⁓ second layer of the things, which is like metaphor and category creation. And at the end of the day, I do think we intentionally make the device kind of worth finding this fine balance between engagement andlearning or kind of like a healthy aspect of the technology, meaning like we add a assistive screen, but we make it really kind of like limited and not the center of the whole kind of like a user journey. And we want to kind of like find a new way to put all the components in our consumer electronics world in a way that it strikes a more delicate balance and ⁓ let the device itself to be kind of like, you know, retentive.for children without getting them to be addicted. So it’s kind of like we intentionally make it a little less stimulating, actually much less stimulating than a lot of a thought-centric ⁓ product. So that’s the main three kind of like principles around the product design. There’s a lot of conflicting constraints here, as you can see, but we do our best trying to find what is the answer. And here we go. Like what you see right now is our first, you know,⁓ answer we have thought through and I think the market validated the answer quite well so far.Grace Shao (07:39)Yeah,definitely. think exactly to your point, know, like a lot of times, I think when we as young parents looking at introducing technology to children, really worried about the big screens, addictive nature, or even the parental, even though a lot of them allow parental control, it’s the unlimited access to a wild, wild internet out there. Like all of these things are basically concerns and or reasons why we hold back technology from our kids.So actually on that note, do think you kind of mentioned it, right? Like technology over the years, especially big tech frankly, has garnered a bit of a bad reputation. And I think that was really tied to the rise of social media and all of this mental illness that came with it. And obviously like you mentioned the addictive nature. So what do you think is actually harmful to the children’s development when we are looking at tech? What are areas actually we can really embrace technology?I think you kind of touched on it lightly, maybe explain it to us in an even deeper, more technical way.Reni Cao (08:37)Yeah, our thesis is that why a lot of parents think technology is negative for a good reason. And the reason is that all the main status quo technology for children are built on top of the attention economy, as we call it. Everything revolves around time spent and how much attention, how much engagement in terms of like a minute, seconds, sessions you can get.That, is the reality because, think about it, you build an app on an iPad, immediately you’re entering a competition with Roblox, with YouTube Kids, with all the videos, all sorts of things out there. You could do well. You can try to do good for the society, for the families, but you’re effectively competing against more like...addictive kind of like a form factor of information and it’s a losing battle and as we call it is a rat race. So no matter what type of like educational apps or content you’re trying to deliver at the end of day you have to deliver them in more and more engaging way more and more gamified and more and more animation used etc etc. That’s I think that’s why it’s another reason why we need hardware at the end of day. I think the first stephow we can create alternative reality is that we need to create a new world, a new kingdom where the business is built upon outcome rather than attention. Meaning like it’s not the time spent logic anymore. It’s like, can you use this device? For example, for Dex, you can use the device and you see the child speaks better after two months or your kid starts to have a like a love to speak Mandarin and not preserve the rest of their childhood.I do think there is a business model there like that, but I believe that business model warrant a totally kind of like a different design of the experience from ground up, from the device layer to the software, to the content, all the way to like user interaction. So I do think like that’s why the current technology is considered bad because it raised towards attention. And I think ultimately, inside Dex,I believe the final answer to create that alternates like a reality is can we deliver something that’s purpose built for children before we build a general sort of like, you know, like time spent logic, like a product in, in, in the case of the decks, is something that, you know, purpose built around the languages. cannot do a lot of things. It cannot, it’s not a chatbot. It cannot, it cannot play videos.But I do think even do one thing super well with the Frontier technology already delivers so much value to the families such that you can build a viable business model on top of that while creating values for families. I think being courageous enough to limit our scope to something to begin with, like really hold onto our principle, deliver a promise, create values there.is another internally operating principle to get there in terms of how to harness the technology. And I want to say that it’s very interesting. What we noticed that is a lot of people are trying to use the AI in quite an all-in-one way. So you can see a little device with tons of features in there. can generate pictures. You can talk to celebrities at chatbot. You can talk to Elon Musk on that device. And we think, actually, that would be a very slippery slope.⁓ in terms of harnessing the technology at the end of the day. yeah, purpose-built is another very critical principle we’re holding on to, to create a good technology.Grace Shao (12:09)No, I love that. But I mean, from a business perspective, sometimes people might not have purpose built businesses, right? Unfortunately, some are not. Then thus, how do we basically help the industry align the business incentive to the product design incentive? Because, know, like what you’re saying right now, it makes a lot of sense. And I think once I saw Dex Camera, I was like, wow, why is there not something like this on the market?but it does feel like there’s a huge gap where, like you said, there is a big devices and the big tech. There’s this tiny niche little products, whether software product or hardware for children’s ⁓ use, but it doesn’t feel like people are taking it seriously, even though we all know parents are willing to spend on children if it’s for their good. It’s not like the economics doesn’t make sense. So why is there’s that gap right now?Reni Cao (12:56)I think you’re hitting on one of our most recent realization that the parenting needs and the children’s needs are quite long tail or as we call it, very like a versatile, right? Different parents have different parenting needs. Even when you look at the language as example, there are tons of different languages and even more dialects you wanna learn. Like let’s say you wanna learn Mandarin, you still got so many like a dialects there. There hasn’t been a real...kind of like technology that can enable a venture scale business that attracts talent, that attracts a good backing in terms of like a capital to build something that’s like a generational. But I do think this is the moment AI is strong. We finally have to make sure we can build one system.that can consolidate all those long tailed needs. Even for Dex, very specifically, you can learn a lot of languages and even more dialects with just like a nine person team building the hardware plus software. I think it’s the catalyst that’s much bigger than Dex itself. And I’m really excited about that. But I think another very interesting angle is like, despite the technologies there, you have another question. It’s like why there is notmore company like Dex. I have a personal opinion here. When new technology comes out, people will tend to use it in the most sloppiest way possible. They were trying to just like, OK, you can chat with the AI, so why don’t we just shovel AI into a little box and put it into a Talking Fluffy and call it an AI toy. And that’s it. That’s my business. I do think it is like a gravity that’s pulling people away.from deeply think how to harness technology and pulling them towards something that’s so trivial and it’s just almost like a shortcut. I think that’s kind of like also, I would call that a trap on the entrepreneur side, that the technology is changing so fast and everyone’s a full mowing, everyone just wanna use it in some way. But I think in this sense, we as Dex, the company, we believe in that.we need to think very deep about how should we use this technology to meet users where they are and deploy like AI in certain ways so Shell can deliver the value. So that’s why we start small, but we’re going to expand from there.Grace Shao (15:09)Yeah.No, it makes a lot of sense, but I think I wonder if you guys all being parents like you just said have made a huge difference. I hate to overgeneralize, but like, you I’ve been in the tech space for 10 years, but usually either I meet men who are like 20 years older than me or they’re very young men who have not, you know, settled into a family yet. And I’m just saying when I tell people my mom, it scares people. They’re like, I don’t know what to say. I’m like, OK, like.I’m not trying to scare you off by telling my mother, but the reality is most of us one day will all have families. And when we do, we start thinking about the things around us very differently, our perspectives shift. And I think to your point when you guys had a lot of purpose designing this product, I wonder if it has a lot of, you know, reason because you guys are parents. Whereas if someone is an entrepreneur for the sake of being a business person, they might not have the nuanced understanding of what a kid needs and what they even think is good for a kid.So to your point, they create little stuffed animals with an L-I unplugged into it, which is horrendously scary. I would never introduce that to my kid, right? I’m getting very, very agitated about this. But you know, another one that we talked about kind of offline was like, I should be ambassador and be paid by Tony Box at this point, because I probably gifted at least like 20 of them out to friends with kids. I think they’re just like, on the surface, you think about it, they’re like, ⁓ a little box that plays music. You’re like, this is so easy. I can just use my iPhone.Reni Cao (16:13)Me neither.Grace Shao (16:32)to exactly to your point. It gives the kids agency, allows the kids to start navigating the world themselves and have preferences. For context for people who don’t have Tony boxes or kids at this point is you put these little miniature IPs, essentially they’re Disney or whatnot, and you can put them on the little box as a magnet. And then the box starts singing and has like seven or eight pre-programmed music or ⁓ stories. And then you can control with your little hands. And basically like you press theReni Cao (16:54)stories.Grace Shao (16:58)big ear, the ear just like the volume goes up, small ear, the volume goes down. It’s like really, really great. So basically introduce technology to kids where they’re like, oh mom, I can control what I want to listen to today. But I don’t need to nag you about it to control the iPhone. I don’t get exposed to a screen. And I can sit there and be entertained for like half an hour myself. So I think Dext really falls into that category for me. Like, you know, we kind of skip the part where we explain how your technology work really and in a very day to day way.It’s basically like you hold a camera, you point at things, you click the button, you can say, what is this? And you default choose languages, right? You actually explain better than me, please.Reni Cao (17:35)So there are actually four questions here. So I want to actually react to all of them one by one. I think this is a lot of good insights here. I think Tony Box and Dex share one thing in common, which is they are children-led, or they are child-led in this case. Think in the POV of a child. The world is kind of like a scary place that you’re told to do this or that.you are brought to here or there, there’s not much quote unquote autonomy you could have. But now there’s a device that your parents actually are willing to let you operate and you can decide what type of content media or interactions you can get. That is just a huge reward to children’s like unlimited curiosity and their like a strong needs to be considered sort of like, you know, a big kid or aeven grown up in a way. I think that’s the intricate magic that if you were not a parent, you haven’t interacted with children a lot, you will miss. So instead of saying like a parent’s made us a better product builder, I think at the end of the day, it goes back to the product 101 that you really need to know your user. You really need to know who are using your product. We spent such a long time with our kids every day.And early days, which is very funny, like ⁓ the first group of users using DAX is just our own children. And that gives us a huge edge there. Right. And I do think you mentioned that a lot of like startup founders in this category, sometimes they’re doing something with raised eyebrows of the parents. I do think they’re a little bit distant from the kids is one reason. And another reason is I do think there is a misconception that children are less.at the end of the day, lot of founders think, you know, those are toys or some gimmicky stuff. Kids, you know, you just give them something that can flash, they can make some sound, and children would love to use them. But I reject that answer. I think that assumption is completely wrong. Children are actually smarter than adults in certain ways. They just cannot verbalize it. But as I said, they already got their little taste.as the famous word, popular words, they got their taste and they sometimes can tell what’s a soulful piece of story versus it’s a very sloppy kind of story. So children actually knows that and they want quality experience, they want quality product, they can actually absorb something that’s really built well for them. I think that just gives us kind of like this endless.sort of motivation to polish our product as if we’re building this for the most critical sets of adult users because we think actually children are more and they deserve more. Now, coming back to how Dex works at the end of the day, I think the core loop of Dex is quite simple. You just take the little camera. I’m happy to actually send a video to be the bureau here. You just take a picture.⁓ And they would just literally just tell you, let me actually take a selfie here. Hi. Let’s see what I can learn about this. Look at that big smile. It’s like spreading happiness everywhere. Can you say a smile?Just smile.smile.Yeah. This is like you get unlimited, like smile comes with some laughing too. It’s when you make happy sounds like, ha. Can you say laughing? Laughing.like the ones we use to listen to music. Do you like music too? Can you say headphones?This is actually English immersive mode. So you can, you can improve your vocabulary there.Grace Shao (21:06)how many languages you have now.Reni Cao (21:08)We have 16 languages and more than 30 dialects and it’s still expanding. And interesting observation here is like the smaller, the more niche the languages is, the stronger the demand is there, which we find is super interesting.Grace Shao (21:21)probably just harder to find offline solutions otherwise, right? Or like harder with the communities, assuming you’re an SF, finding a Mandarin community is not that difficult. You know, if you’re in England, finding a French community, probably not as difficult. if you go, you were saying like maybe like Arabic languages like that are not as mainstream, maybe in San Fran, you have people in San Fran wanting to do that, right? Or like people in Dallas last time you said, trying to learn Mandarin, which again, you don’t have a huge community. Very interesting.I’m sorry, I got very passionate about the topic. So I want to of swerve back to our conversation here about raising children with technology. I’m sure you get pushback. think people right now, there’s the other side of argument where everything should be organic. Everything should be very simple.Reni Cao (21:53)Yeah, of course.Grace Shao (22:08)And I myself, I’m a big fan of a lot of the Montessori toys. You know, they’re not buttons or not even power charged. They’re just little wooden blocks, but they’re designed very well for them to, you know, develop motor skills. So how do you kind of explain to parents today who are saying technology should be rejected in the childhood. Kids should just be reading physical books. should learn the way that we learned or even like previous generation learned. We should go back to touching grass only. SoLike, yeah, what’s your argument there?Reni Cao (22:37)First of all, you are completely right. Every once in a while, we got a comment on our social media that, why don’t you talk to your own daughter to teach that language? Why do you need a device to do that? So your assumption is completely right. And my response to that is, first of all, actually, I respect that parent a lot. I believe in the most ideal world, organic human-to-human interaction and free play in the real world is great. There’s a lot of tech, like researchers actuallyProve that right, right? However, I do think the parent miss out constraints here. Number one, you may want to talk to your daughter, but you don’t know Cantonese, for example. So there’s no way for you to teach some subjects or some skills that you want them to learn or you want to immerse them with. And second, all of us know that the contemporary society is more and more fast paced. Not all the parents enjoy this privilege.of saying, let’s slow down, set up a dedicated time for children to go out to places. All sorts of this ideal family style back in the 80s and 90s changed a lot, I would say. So we are, believe, rather than just blaming the parents, not spending enough organic time with their children, I do believe that technology should be introduced more as an option, as kind of like a gap stop.as one of the extra tools on the table. That’s why when we design decks, we don’t introduce chatbots, but we spend so much time on sharing the insights that what your children are interested in. What did they take a picture of? What do they want to geek on? What did they learn today towards the parent app? And just give them this little window to see the world through their children’s eyes. Give them good downtime topic.giving them a way to reconnect even as asynchronous. So I do think the concern is real and the overall kind of like, you know, judgment is very well reasoned. But I think what that’s the approach here is much more nuanced than saying like, let’s use technology to replace human. It’s not, it’s actually using technology to connect the humans, connect the parents and kids better. That’s the nuance I have to take a bit.Grace Shao (24:42)I see what youYeah. No, no, I love it because actually I’ve seen some parents even give kids like little Kodak cameras these days and these little toddlers go around the world, take pictures of how they see the world and they’re so cute. My own daughter sometimes takes my phone and takes pictures around the home and I come back with a lot of selfies and pictures of her sister’s foot or it’s just very cute because you see the world through their eyes, right? And it gives like, it’s like technology doesn’t take all connection away.on technology. wanted to ask you about the technology. How do we understand that? Like how are you actually leveraging LLMs? How do you route through different LLMs or different languages? Is this something we talked about briefly? But I wanted to understand that bit more.Reni Cao (25:20)to share details. Where should we start?Grace Shao (25:22)Like how does it work? right now? So basically for the little Dex camera, can’t ask it, like he’s to your point, you didn’t build a chatbot. So I can’t ask a question. I can’t have a conversation. It’s not a companion, but I can ask it what’s this? How does all that work in terms of the back end technology and the guardrails you built up?Reni Cao (25:38)Yeah, I thinkin a 30K feed view, Dex are utilizing basically all the multimodal LM capabilities to understand what the children are looking at. And on top of that, we build sort of like a profile, interest profile for the children and the parenting need profile for the parents to help contextualize, you what responses should we give in that case? To give an example, if you’re a three year old,just starting to learn Cantonese and you are sort of like interested in a bunch of like a museum topics or you love like dinosaur skeletons and stuff like that, we will render you more challenges around kind of like hey let’s bring Dex to a museum and learn about different terms there and it will be English the primary languages teaching entry-level Cantonese things there. So basically like the visual understanding you certainly use like a multimodal LLMThe response definitely use kind of a conversation API of a lot of like an ALM. And I think building out this context layer or this memory layer of like a children’s interest and parenting needs, that actually is more complex. That takes kind of like a full agent system to try to understand what matters, like condensing or distill insights into a profile and gradually kind of injecting that into our responses. I think that’s on a very high level. That’s it.We do use a wide range of LLM, mostly with Gemini and OpenAI. yeah, that’s kind of like the high levels.Grace Shao (27:08)I’m going ask a question you might not like, but I’m going to put you on the spot. When we talked last time, said specifically on Cantonese and Mandarin, you do use different LMS, but the accents can be quite funny. Like they’re a bit off. They’re not native sounding. Why is that? And how do you overcome something like that? Or other maybe non-English languages. Yeah.Reni Cao (27:12)No, ask me.First of all, you need to try again because we have a solution already. But definitely, hit. We are already squeezing. I’m so hard that we’re hitting the boundary of a lot of like, in this case, it’s a TTS of the leading providers. Because I think about it, I’m pretty sure you’re using English plus Cantonese. It’s basically using English to learn Cantonese. Is that the case?Grace Shao (27:50)Yes.Reni Cao (27:51)is a mixture of languages cases. The challenge there is that without fine tuning, there is very limited sample of someone that speaks very good English and very good Cantonese, and they mix them in like one sentences. So the data, the training data to begin with is a little flawed. Either you have accent English or Cantonese as the more common cases. That’s the fundamental root causes of this. And we’re having kind of like heavy lifting tasks to kind of like solve that.And with the foundational model getting better and better, think one day we’ll get there. And we can see that to be fully fleshed out in the next six months. You definitely hold us accountable. And I think this is right observation for mixed languages. It’s really hard. Yeah.Grace Shao (28:32)Yeah, I bet. how does it actually work right now? Like in terms of economics, like people pay you about $249, right? That’s the price of the product pre-tax. That’s not cheap. Like it’s much more expensive than a toy, but obviously bit cheaper than an iPad. How do I understand the pricing decision there and price? And then how does that relate to, I guess, how you pay for your token usage right now? Does that cover it?Reni Cao (28:57)Yeah.Yeah. Oh, big time. We actually have a pretty healthy margin and the tokens are getting incredibly cheap. Much cheaper than where we started. I’m talking about like in 96, 97. It were a fraction of the token cost of where compared to when we just getting started, which is back in 2024 February. At that time we don’t even have GBD4, we have GBD3.5. that’s the kind of like, that’s the kind of like, actually that time we have GBD4 butis we don’t have GPT-4.0. So it’s very expensive at that time. So now pricing. Actually, I have a let’s talk about the user-centric view and a business-centric view. On the user side, we’re actually adopting this value-based pricing model, which is like any enough day, language is a high value skill to acquire. I sent my daughter to a language immersion in the US. I’m very embarrassed to mention how much I spent on that school.Grace Shao (29:33)Okay.Reni Cao (29:49)And if DAX can offer 1 % lift or enhancement on top of that school, the price is fully adjusted and much more than that. So this is what I mean by like, and very funny that you mentioned toy, right? Toy is something that you get it, you play it for a couple of days, then you don’t see it, you don’t worry about it. And this is not what we’re trying to do. What we’re trying to do is we want to use a relatively high price to keep ourself honest aboutthe value we’re delivering to the parent. Do we really teach a language or do we really get the kids to fall in love speaking that language? If we do so, that price is well-justed. If not, we’re going to give you 90 days of free return period. No question asked, just return it to us. I do want to use this pricing model to push us to deliver more value for the user. So that’s one aspect of it. And on the business side, very funny, you mentioned, I hate when people box us.into toy category. I don’t blame them. Natural reaction, but I want to send a signal to the market that if a team of talented people, hardworking parents, put their heart and soul in building a purpose-built device that harnesses AI and delivers concrete results, we could get out from the typical, stereotypical, like a toy average order value band and go much higher. Above that, it’s less about, I to keep myis more kind of like, want to send a signal to prove that the market, we have enough parents waiting anxiously for something similar to this and want to pay a perceptually higher price for it, a premium for it. But yeah, that’s kind of like we landed on that price. And it’s so funny that so many people in the early days tell us, you’re going to do $1.99, because anything that started with a oneGrace Shao (31:22)Premium, yes.Reni Cao (31:36)is night and day different than like two, that it started with two. But I actually, I’m like launching a suicidal mission. was like, let’s actually make it start with two, but let’s deliver more value there because it’s never like, it’s not a retail business at the end of the day. We’re trying to create a new paradigm of digital parenthood and childhood. We need to hold a high bar for ourselves. And the price is very telling, like in that case.Grace Shao (31:59)No, I actually agreeand I think would you categorize yourself in the same box as Tony box vertical? Would you?Reni Cao (32:06)Not really. ⁓ Tony Box is a, I would say they are a content business. they are, same thing with Yoto. Actually, their founders have deep backgrounds in labels, music labels specifically, and IPs. So they are effectively a distribution business that they are creating a new channel to distributing those IPs from Disney, from Spin Master, and et cetera, et cetera. And the other side, you can see that at Dex, we’re notGrace Shao (32:19)I see.Reni Cao (32:31)I think like IP partnership or putting characters on our device. And we actually optimize for value and outcomes, like I promised to you in one of our principle. So I would put ourselves in, I don’t know, the de facto smart device for families. Just very honestly, the family device, the family technology, maybe like this is where we’re trying to go to, but it’s a completely like non-existent category before we’re still exploring.Grace Shao (32:33)Yeah.Yeah, yeah.Okay, like family tech device.Reni Cao (32:59)and it may change how I call it.Grace Shao (33:00)think there’s some moresimilar things maybe in East Asia because the audio learning like you know even when I was very young like I remember my grandma had a 步步高步伏机 I don’t know if you know what that is it’s like those like tiny little yeah yeah basically what it is it’s like people learn English with it and I think it’s very very like mainstream in China for a while but like you know these things been around I think in East Asia because everyone is using it to literally learn EnglishReni Cao (33:11)The steps are fine.Grace Shao (33:24)But it’s very one dimensional. It’s like one language to one language. They basically embed a dictionary, make the dictionary into a digital one. And you can ask search questions. You can ask what this word is. might, more advanced one might be even like with images, but I think, I don’t know, in the 90s, I didn’t see any images. But yeah, it does remind me of that technology and that vertical. haven’t seen something like that too mainstream in the West growing up, you know?I think if I was when I was learning French and German growing up, that would have been so helpful to your point. But yeah, so I want to bring it back to sorry, I just want to bring it back to the the business. On the Tony box comment, I do believe their business actually could be really high margin because their product is only say like 199 or something like that, right? Like they’re the box. But each character is not a 20 bucks or 30 bucks.⁓ My daughter is drying me up here because every two months she asks for a new figure. But my point is, it’s a great business, right? Like that thing just keeps selling. It’s like Spotify and a physical thing. So would you guys have add-on any services, software, hardware, anything?Reni Cao (34:32)We do.That’s a lot of investor has been pushing us regarding this razor razor blade business model. I think for us though, what we are ultimately delivering is a business more like an app store.It’s like where you can get personalized content and software for your parenting needs and for your children’s growth needs at the end of the day.We’re launching, not we’re launching, we launched two tiers of subscription so far to validate that. One tier, $10 per month, you got unlimited LTE, plus you actually got a curriculum packed in like a content library. Every day we give you one topic and in the topic you can explore a lot of new vocabulary, expression, know, new languages and it’s good kind of like content to consume. And I think what’s most interesting is our future vision is actually a $20 per month tier.In that tier, you can actually create activities for your children, personalize. Grace, can be like, I run this podcast. I’m a podcast host. How do I explain that to my kid and make it a little bit fun, exciting, and even adventurous as if the recording a podcast is a little journey? And by the way,my kid likes this way of storytelling. You could give a lot of like a prompt there. They’re actually based on the profile, the context layer, we’re gonna build sort of like interactive, like a content that involves taking pictures, speaking, and just like looking at the device for explaining what does podcasting mean. And this tier actually got really good like attraction. And when we look at their subscription retention,is above like 90 % in three months that shows early signs of product market fit. But this is what I mean by like our business setting of day. We are a channel to deliver like harnessed intelligence to parents such that they can build whatever content and software that adapt to their needs rather than just a purely search, then filter or control kind of like a timer. I really want the digital world to revolve around them, running out of way around. So in this case,Put it in a simple way, we give them a tool to build whatever they want, and we charge on the usage of the tool, pretty much.Grace Shao (36:41)No, I actually really see that. I love it. Because I think my husband was trying to use chat GPT for a while to create stories with my daughter. Like, add a pig, add a dog, add a whatever in this. And obviously, it’s not made naturally for this. So the stories don’t come out as, I guess, natively understandable for children. So I see where this can go. And the funny thing, you use my profession as an example.Reni Cao (36:49)Exactly.Grace Shao (37:05)example, like my daughter just thinks I talk all day, that’s my job, and she thinks that her dad sits at a computer and press buttons all day. So between the two of us, none of us are doing it much, just talking and pressing buttons. So it’d be really great if, you know, I can, I guess, lean on technology to find a better way to explain to children modern day careers, you know, that may be not as easy to explain as, know, mommy’s a doctor, and doctors go help people and save lives, which is like what my family has.you know, explained to us when we were growing up, you it was very clear. I want to kind of go on a little bit more about AI and parenting. I think there’s a huge discourse right now in the US, especially, I think from my point of view, where I sit in Hong Kong, in Asia, even yesterday, I was speaking to someone from South Korea, venture capitalist, they’re saying that parents and society seems to be a lot more open to bring technology into their day to day lives.They’re much more open to the idea of leaning into technology for personal use and less worried about privacy and you know these kind of issues I guess. So at a high level, what do you think, should we be concerned when we introduce technology to children? will they, you know, for example, taking pictures themselves that automatically goes into one of the LLMs. Is that something that...he should be mindful of or are there guardrails that can be built in?Reni Cao (38:26)We should be definitely mindful. That’s why we enforce ZDR, zero data retention across our stack for images. So even let’s say your kid take a picture of themselves, you cannot retrieve that picture even you want. You can ping me through my personal email. You cannot find that picture anymore. And OpenAI and Google signed a contract with us to burn a picture immediately, like zero data retention on all the usages. But overall, I do thinkGrace Shao (38:48)See.Reni Cao (38:51)It’s the company’s responsibility to introduce technologies to family and the family should hold a high bar there for sure. Because like the AI is so early and it’s way too powerful in certain way. And it’s like a kind of like a black box in certain way in a lot of different ways. that I definitely, I’m not a, I’m not that one of the technologies that wanted to like, you know, glorify AI and it is the future and stuff like that. comes with a lot of risk, especially like unproven.aspect how it impacts the children’s cognitive development and something like that. That’s also a reason why we work with researchers and professors ⁓ closely like in Mount Eucalon from UCSF and Harvard professors doing education and doing research using text. I do think there is a substantial risk here such that theAnd we as the entrepreneurs and we as the parents, we need to hold a high bar for ourselves and roll out things one by one. So I guess that’s why you will hear more about like, oh, that’s like, you could have done this. You could have made it more engaging. You will hear this much more often than you’d be like, oh, there is like an incident because, you know, we always prioritize, you know, safety first. We’d rather the device to be boring in certain way rather than introducing consequences that we don’t understand.So I think there’s a very interesting dynamic between the Western and Eastern in terms of their views about technology. And I don’t think it’s a family parenting only. It’s also even a whole society, general perceptions. Happy to chat about that, but maybe it’s a little bit off topic here. Yeah.Grace Shao (40:12)comes from the mindful design as well.⁓ No,we can definitely talk about that a little bit, but I kind of just follow up on what you just said. So how should a parent evaluate an AI device or tech device when they are purchasing for children, right? ⁓ I’m sure there are different devices out there, maybe not exactly doing the same thing as what you’re doing, but other devices are tech native ⁓ or AI enabled for children. How should parents kind of go about this?Reni Cao (40:52)I’m not a parenting coach. I will share my views. Number one, do think we should bias, we should start from our needs first. Maybe let me put it this way. Don’t get carried away with all the possibilities of the AI. Ask yourself, what is the unresolved parenting needs you have and find solution there. Rather than, this AIX, then that’s just to buy that AI device and give it a try. That’s number one, I would adopt that.Number two, I do think it’s important to see what a company’s method is. They definitely put their methodology somewhere, their belief somewhere, their principles somewhere, they’re kind of like, like how you ask me about how we use ALM. I believe that the parents should definitely hold the company accountable to explain those details and ask, verify, and that’s crucial step. That’s the due diligence on them, right? And I do think thatFinally, for any sort of AI product, I actually even think the parents doesn’t have to be getting into this searching and validating mental model. They could literally build their own in some sense. Given all the agent codings rising up and reducing the piece cost of software so low, I do think for lot of stuff, they should try.to accommodate their own parenting needs in certain ways. Like I saw tons of the parents go into cloud code generating like a, know, almost like a story writer for their daughter. That’s actually my previous colleague at Wish. And it was awesome. It’s just different blanks to fill in. It’s kind of mad lips type of like a story. I do think the parents can change also their way that they are in the autonomy right now to build whatever they want to build.Having said, it’s still a little bit of kind of like a Silicon Valley bubble type of answer, because honestly, in the world, the adoption of a cloud code is probably less than 2 % or 1%, I’m pretty sure. But I do think I would encourage parents to use AI themselves and explore a boundary, it can do, what it can does well, what it doesn’t. So then, the kind of I make a decision from there.Grace Shao (42:45)Yeah, no, I I appreciate that. It’s a very like thoughtful answer because it’s not just like A or B. of the day, think it’s parenting itself is so personal. It’s on how your family dynamics work, how you prioritize your time, how you want to parent. So when you want to buy technology for your children or incorporate that into their lives, it’s also a personal decision. I wanted to ask, actually, do you have any good case studies to share with us just a little bit?Reni Cao (43:11)We have quite a lot. What aspect, what, what type of case does he want?Grace Shao (43:14)Just like, I don’t know, like things that unexpected people use. For me, I mean, by default, just assume, yeah, people use it in urban areas, right? But then I think when I met you, you said, actually a lot of people use them, you know, in unexpected places, like orders come through all over.Reni Cao (43:19)Alright, I’ll give you one.One of, I immediately think about one thing, one, almost like ⁓ close to 5 % of our users, they bought Dex to help with speech delay. That’s something we never anticipated, but those parents are very frustrated with all the, as we call it, sometimes autism tech or the speech therapy tech there. It’s not meeting their bar and they saw Dex, they’d be like, I would try everything right now for my kid. And surprisingly Dex helped them.and it makes them real happy. And you can find actually all those real reviews in our review sections. Quite a few family mentioned that their kids refuse to speak certain languages or just even English, but that’s kind of necessitate the language as a fun activities. And all of a sudden, the kids start to open up and speak more, and the parents are really happy about it. This same exact story happened with my co-founder, who is really worried about his, at that time, two-year-old young son having speech delay.But I want to disclose the name, but the song he actually first time spoken like coherent like ⁓ Chinese phrases using Dex and he caught it on a video. That was one of the most wholesome moment of our kind of like a user feedback in our channel. And right now we’re actually ⁓ volunteering to develop this special need mode. That’s kind of like, you know, customizing to special needs children.especially like April is the world kind of like autism awareness month. And yeah, we just want to do it. And we want to donate to Dextre researchers and speech therapists to help us do it. This is a totally kind of like a side quest, but it just like give us it gives us so much kind of like energy. You’re thinking about technology can be used in a way that’s like immensely helpful.Grace Shao (45:00)That’s amazing.Yeah, and something unexpected, right? Okay, I think I want to wrap up our conversation because I don’t take up too much of your time, but I do want to ask you one big macro question. With you working on whether you like to call it physical AI or not, essentially like a physical product hardware time software, how do we understand that trend going forward? Do you think AI will be essentially integrated, plugged in to more more hardware devices? What’s your view on that?Reni Cao (45:33)I do think there is a consensus that every wave of software technology revolution, there will be kind of like a device revolution following that. We are at the tipping point there. That’s like people starts to reimagine, where is this? this cloud? Is this the recording card? Maybe it should be a separate, like an ⁓ AI. Or this is a sort of like a little pendant that can kind of like ultimately listen to your life, help you organizing. I do think we’re at the ⁓the dawn of a next wave of hardware. But it’s less about we’re doing the hardware because of the, I do think this is a software or technology driven type of hardware revolution out there. I do anticipate that. I do at least what I’m 100 % sure is like smartphones are not designed for children. Tablets are not designed for children. Families deserve something built with their interest.their needs in the center of the spotlight. And I see that happening. And that’s why we started this company. And I bet there’s going to be tons more use cases there.Grace Shao (46:33)No, amazing. Thank you. I think ⁓ one last thing. Is there anything I missed or anything you would like to share with us?Reni Cao (46:39)By the way, I time. If you want to turn through all the questions, I’m happy to be here. I don’t have anything else after this meeting.Grace Shao (46:44)no, don’t worry. think it’s a lot of times like I use them as prompts. But you know, when we’re chatting, like we actually covered most of it, you know. ⁓ Yeah, is there anything you think we missed? But from my end, like I feel like I covered most of it. You know, we did technology, we talked about children, AI philosophy, talk a bit about your business model.Reni Cao (46:50)Yeah. Yeah.Yeah.I do think you would want to talk about. Yeah, go ahead. Go with one last one, and I have one for you. Yes, go ahead. Ask yours first.Grace Shao (47:04)I think one last one. You go.So I wantto ask you one last question, which is a question I ask every guest that comes on the show. What is one differentiated view you hold? I feel like your whole thesis around devices right now on the market are not made for children is already a differentiated view. But is there anything else you think that you hold that’s non-consensus?Reni Cao (47:29)Yes, with this view, I got beaten up so many times, but I still got to say it, right? I believe that education should not be cookie cutter. It should be highly personalized. So is entertainment. So is the parenting software. And we’re about to enter the golden age. Finally, this is becoming the reality. And let me say it this way. You look at a school in the US, how you tell the school is good or not, you look at one ratio. It’s called a teacher-student ratio.One teacher taking care of less, but why? Because then the teacher can accommodate, individualize the needs. I actually have a very radical view in terms of our education system is definitely lagging, significantly lagging against how our society evolves, how the technology evolves. It’s still a one size fit all and industrial way.to handle education, handle like, you know, testing, standard testing. It hasn’t really changed in the past couple of decades, but the world is a different place now. And I guess my view is like, it shouldn’t be that. The default shouldn’t be that. The default is like every kid should almost have their personalized tutor and the playmate that deeply understand them. Unfortunately, that’s impossible before, resources-wise. But I guess we need to strive to get there.as a race, as a humanity. Because each kids just come up, come with their own spark.that will miss out the window to make that spark their lifelong journey. But I’m not trying to attack on educators or school systems, something like that. I just feel like there needs to be more forces from the society, especially from the tech side, to help together build this alternative, enhanced of like a system that really delivers individualized education.Sometimes I use the word scaled homeschooling. And you cannot imagine how much people hate that. people are like, homeschooling, you’re taking away the social aspect of it. People are very constrained on the vocabulary of how they describe things. But I guess when I say homeschooling, it’s not about keeping the kids at school and hiring a teacher. And that specific process right now, I’m talking about really meet children where they are in terms of their growth, in terms of their needs, in terms of the skills they’re going to develop.I call that a differentiator, but maybe actually lot of people will share the same views. I’ll be happy to know who shared the same view and please join us in the journey. Follow us along.Grace Shao (49:49)I definitelythink that view is definitely, feel anecdotally a lot more prevalent in SF when I visit. I’ve met other people like yourself, other people in the tech space or, you know, investors who are embracing this idea of modern homeschooling. And they say the same thing. They’re like, we don’t like to use the word homeschooling because, it sounds like a bit more cultish, but it really isn’t right. Like it’s really focusing on individual ⁓ growth.I think it’s amazing because I also think it’s because Silicon Valley itself kind of harbors this kind of growth and mentality and that the fact that people can succeed without degrees, people can succeed by building different things, people can succeed in just being different and being themselves, but the best version of themselves have always been, I think, what drives a lot of people who want to go to Silicon Valley because it’s like in many ways, as a mayor, talk to see like the best version of my talk to see, right? I thinkin East Asia, even as I put my kid in school right now, I find people definitely a lot less like that minded. ⁓ I don’t know if it’s a cultural thing because like, know, for you, you know, I grew up in Canada for me, I always felt like, you know, having that freedom to learn, explore when you’re young, which is more Western kind of way of, I guess, education was good. But I think a lot of peers here actually believe that, you know,for the first like say eight to 10 years, that foundational education should be drilled in. know, ⁓ grit should be taught, discipline should be taught. But it’s very interesting because it does kind of, I guess, manufacture different kinds of stereotypes. And I think it’s fascinating. And I think one more comment on that, I know this conversation has been more personal than we thought it would be, but I love it, you know.I don’t really get to talk about motherhood that much in my podcast. It’s usually about tech and bros and tech bros and about, ⁓ and about finance. but I think even, you know, when you have kids, people talk so much about nature versus nurture. And what I realized is I was shocked to see the nature come through, as young as like six to eight months in a child.Reni Cao (51:36)YouGrace Shao (51:53)their personality starts coming through and by the time they’re one to one and a half, they start kind of babbling, start demanding things. I realize 80 % of it is all nature. It’s like their preferences for how they socialize, their preferences of even noise, even you can realize like your point, your taste. You’re going to find a six months old who just wants to sit in a corner in a play group who just wants to flip through books literally and just undisturbed. You’re going to find someone who’s screaming in the middle of the whole group.you’re gonna find my daughter who’s rolling over everyone and just like trying to knock everyone out. And I don’t know why. You know, you’re gonna realize all of it is nature. And even I believe agency, autonomy, grit, and desire to actually succeed, that itself is nature. And I don’t think you’d be taught. And I think this is a bit controversial. But definitely I think my husband and I have been thinking a lot about this. We’re like, we can just provide them what we can. But there is no...point of even pushing them when they don’t want certain things. the best is to push them in a direction that they want to be pushed and they will tell you. I think this is like kind of the difference in our generation of parents. yeah. Reni, thank you so much. ⁓ Yeah, go on.Reni Cao (52:49)Exactly.Yeah, but can I, I know thisis over time, but can I add one last comment towards what you say? But I think what you said, especially growing up in East Asian, like, you know, education system, it has been industrial for a good reason, right? At a time where stuff like AI doesn’t exist, the most effective way,Grace Shao (53:03)No, of course, of course.Reni Cao (53:19)to develop fundamental knowledge workers, plus finishing the job of dividing the children into different segments and give them different levels of education. That education system works perfectly. I entrance exam, as I’m talking about, taking standard tests and stuff like that. But all we know that is AI is sweeping through all the knowledge works.and specialized in knowledge works, honestly, Asian parents like favorite jobs, like being a doctor, especially radiologist, you know, and or being a lawyer, you’ve got to start somewhere as associate. Now it’s getting kind of like his hardest. The world has already changed. The tsunami already hits. But I don’t think people actually understand the level of the s**t. A lot of like everyday people in the world, they haven’t felt.this like a tsunami, right? So when you say you want to kind of like, you know, like find define your children’s nature and push them towards kind of like what they are intrinsically motivated about and give them resources to set them up for success, building grades are on the way. I do believe that I think I will 100 % agree with you that it will become the most fundamental aspect or element of education in the next like five years or even sooner to be fair.That’s why I don’t send my... to put it in a simple term. I don’t send my daughter to Kumon. I don’t want my daughter to do Russian math. I never benchmark her against like, oh, like the other kids can read at the age of like three and a half. Why don’t you? Actually, I don’t because I fully understand that kids have their own time zone. Kids have their own spark. All you need to do is think deeply to define that, to understand that, understand why my daughter sometimes is super sensitive, understand why sometimes she got frustrated and want to hit.Grace Shao (54:31)I feel very validated.Reni Cao (55:00)Don’t take that on a surface level with the other tools you have. Go deep, understand that, and build these programs that’s personalized to her and help her. And I think like this is why I if I, talk about the word of Nei Juan a lot. If I have to dream on anything, right? I have to like a rather like ruthless compete on anything. I complete the deaf understanding of my daughter rather than anything else.Grace Shao (55:03)100%.Reni Cao (55:26)Because I actually think that’s the thing people gloss over. People must be like, education is just checklist. You got to check, check, check, check. And there is a better checkbox. Like Ivy League school, there’s a OK checkbox. There’s a worse checkbox. Forget about a checklist. That checklist is obsolete already. So I respect. I think we vibe together in terms of our schools of parenting.Grace Shao (55:33)Yeah.Yeah, 100%. No, I agree with you.parenting style. Yeah, yeah,yeah.Reni Cao (55:51)But you’re so fully intuitive. don’t know whether I’m right or wrong, but this is what I firmly believe in. And I believe someone’s going to join this journey.Grace Shao (55:58)I think there’s more people who are aware, especially people who are more plugged in with the technology because they realize how fundamental society will change. I just thought about when we were young, I’m sure your parents also told you to go to university, go to this, go to that, right? For sure there was a hierarchy in their mind, what kind of school you should go to, what kind of degree you should get. Now I really don’t think that’s the case. Actually, a lot of my readers would even know.my dad really forced me, well, pushed me, encouraged me to go into finance. And at one point he was like, if you don’t study finance and don’t work in finance, you’re not like following my footsteps and blah, blah, blah, blah. Right. And it was a very, it became a personal reason to do it. It’s not because I wanted to, or I was good at it. And there was actually a battle between us being like, I want to go into journalism. And he’s like, no, I was like, no, I’m going to go to journalism. He’s like, I’m not going to pay for it. You figure it out. But the beauty of it is actually found a way resourceful enough to get a full time, full scholarship.And I still want to journalism. Again, I recognize how lucky I was. I I found the opportunity to do that. But most kids actually just end up then doing what their parents told them to do and they never, and they never actually live their best life or become the best versions of themselves because they’re doing something not actually fundamental.Reni Cao (57:08)you hit a very critical, I think it’s a background or context. There wasn’t an abundance before, right? Growing up, let’s say in the eighties, It’s a relatively kind of like a society. It’s relatively kind of like not that sort of like, you you wouldn’t call it abundance at the time. Let me just put it that way, right? You still need to compete for stability, compete for resources. That’s why there’s a rat race in education, which I totally understand. That’s kind of like.It’s like a whole economy there, right? But I think that changed. No matter what we’re talking about, like, I mean, in China, I’m talking about US right now, I think abundance really will hit at some point of time. At that time, the challenge shifted from how can I avoid getting into a property or like ⁓ job loss towards kind of like, how can I find the meaning of my life? And how do I deal with this kind of like a journey?It’s a generational theme there. It’s very funny that your dad wants you to go into finance rather than journalism. mean, for those who understand Chinese internet a little bit recently, there has been a famous influencer called Zhang Xuefeng. He almost helps everyone to pick their college major. And one of the college major he hates the most and advice everyone to not go to is actually journalism at the end of the day. Because it’s just not a...stereotypically stable job that can make a lot of money, that can give you social status, quote unquote stuff like that. But I think that’s a lack of view of things. Our dad doesn’t know how our skill landscape is to be 20 years later, just like we’re not going to know what our kid is going to deal with. So we need to give them a more sort of generalize the resources and the skills at grit to survive. And it’s funny enough.I made my, I named my daughter Simone because we are big fans of Simone de Beaufort, the kind of like, you know, foundational philosopher of feminism and a lot of like a sociology thoughts. So, you know, if I have, if I have to give a set of expectation for my daughter, I want her to actually do something that’s not commonly seen as a very, like a prosperous or stable kind of like career ladder. I want her to do something that’s kind of likehas a mission and some sort of like outlier type of journey. let’s see how that goes. She’s so young, so who knows.Grace Shao (59:20)I love it. All right. Thank you so much, Reni.Reni Cao (59:23)Thank you. AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• There's more to Korea than just chips. TheVentures CIO on the country's AI stack 11.05.2026 54min

  In this episode, I spoke to a leading South Korea VC, TheVentures’ CIO Ethan Cho. He argues that South Korea’s low fertility rate and aging population put pressure on Korea to be one of the world’s fastest adopters of AI technology, similar to its rapid embrace of high-speed internet in the early 2000s. While not a leader in foundational LLMs like the US or China, Korea’s strength lies in application and adaptation, particularly in B2C areas like personalized agents and commerce, where cultural familiarity with chatbots and digital transactions lowers resistance.The Korean startup capital funding landscape is shaped by three forces: Chaebols (Samsung, SK, Hyundai), the government, and VC firms. CVCs from Chaebols tend to reinforce existing semiconductor and hardware value chains rather than explore tangential innovation. To counter this, the Korean government has become a dominant LP through initiatives like “Everybody’s Entrepreneurship,” injecting capital to encourage novice founders. On sovereign AI, he believes the government’s push is less about global dominance and more about securing sensitive areas like finance and defense, though he warns that domestically-built software has historically struggled to scale beyond Korea.Ethan is shifting focus from purely domestic champions to founders with global ambition but local execution, often Koreans educated abroad who return dissatisfied with traditional jobs. He wants to back ventures that change the world, not just build another food delivery app. He also recognizes key opportunity areas, including defense tech, K-beauty, fashion, and mental health, as society adopts AI at scale.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, founders, and product managers.For more information on the podcast series, see here.To find the previous episodes of Differentiated Understanding, see here.Chapters00:00 Introduction to Ethan Cho and His Journey02:47 Korea’s Role in the Global AI Supply Chain05:24 Cultural Attitudes Towards AI in South Korea11:06 Government Initiatives and Sovereign AI16:37 The Future of Commerce and AI Integration28:03 Consumer Behavior and AI Adoption28:43 Enterprise AI Solutions in Banking and Manufacturing33:39 Investing in Founders: The New Generation of Entrepreneurs39:39 Korea’s Future Exports: AI and Beyond41:41 K-Beauty and K-Fashion: Cultural Exports45:15 The Future of Mental Health in the AI Era49:53 The Limitations of AI and Human ExperienceAI- generated Transcript Grace Shao (00:00)As mentioned, our guest today is Ethan Cho. He has been active in the Korean VC space for over a decade with experience in the venture investing arms at Qualcomm, Google, Samsung and more. Now as a partner at the ventures, he leads a team focused on finding and nurturing the next generation of AI native startups. Ethan, thank you so much for joining us. So good to have you.Ethan Cho (00:18)Thank you, Grace. I’m very excited to be on the show and I would love to discuss with you more in detail.Grace Shao (00:24)Yeah, to start with, tell us about yourself. Tell us about venture investing in South Korea and the firm’s background.Ethan Cho (00:31)Sure. So I was born in Korea. I kind of moved internationally quite a bit. I moved to England when I was kid, when I was four years old. That was where I first learned my English, lived in England about four years, came back to Korea, then moved to Hungary, lived in Budapest for a year, came back to Korea again, spent the next 20 years in Korea, moved to the States, lived in New York for ⁓ my business school years and worked there for another year, came back to Korea after then. So I’ve been in and out of the country quite a bit.I loved startup investment very early in my career, so I wanted to move towards startup investment. I actually started out as a hedge fund analyst right after business school, but I quickly found out that I’m more interested in finding good companies and good stocks. So then I moved towards the private side, started with Samsung, and moved to Qualcomm, et cetera, et cetera.What fascinates me about Korean startups and startups in general is that everybody’s trying to change the world. I’m just such an honor to be a part of that and talking to entrepreneurs on a daily basis really excites me.Grace Shao (01:34)Awesome, I think it’s really interesting because you’ll definitely bring a very international perspective and not only just the Korean perspective and also kind of understand, you know, where a lot of our listeners are coming from as well. You have, you know, you have exposure to UK exposure to Europe, exposure to US. I think I want to ask you quickly, because you’ve actually worked in the public sector as in public investing, it’s kind of interesting because right now, obviously, the frenzy and the, you know, the global interest right now.Ethan Cho (01:57)Mm.Grace Shao (01:59)is in a lot of the big semi providers in South Korea, you know, focused on infrastructure layer that are public listed. At a high level, how should we think about the Korea’s role in the global AI supply chain? And then of course, we’ll shift gears into talking about the startup scene that you’re passionate about.Ethan Cho (02:02)Yep. I think that’s a great question. think one of the very obvious factors of AI is memory because you can only use AI based on whatyou or the agent knows about you or any company. So because of that, the demand for memory is exponentially increasing. I think that’s definitely a blessing for the career semi-players and also the current industry as a whole. But at the same time, think nature has always found a way to become more efficient. So the capacity or the demand constraint will not beexisting forever. There’s going to be significant improvements such as Moore’s Law. There’s always an innovative solution that comes out every other year. So I think there’s definitely going to be a lot of exciting opportunities down the road, but it always evolves. So it’s going to be very different next year. It’s not going to be HBMs anymore. I think it’s going to be something else going down the road. We’ll have to see, but I think the trend is here, but the trend itself will also keep evolving down the road.Grace Shao (03:15)Awesome. So, you know, if we have a really candid assessment of what’s happening in Korea right now, what’s genuinely really strong do you think? Like the chips are strong. ⁓ What do you think that weaker maybe compared to, you know, China and the US, you know, from outsider lens, is it really the LLM labs kind of or is it diffusion? What’s happening?Ethan Cho (03:24)Hmm.Yeah, I mean, it’s a complex situation. I think one thing that a lot of people quote and one kind of fate that we cannot deny is that we have a very low fertility rate. So the birth rate is decreasing very fast. We have a very rapidly aging population. A lot of people think of this just as a curse. I think there is an aspect to it that it’s a blessing in disguise because we are one of the countries that most desperately needs AI and robotics.And because of that, I think we will be one of the more adapting or welcoming countries for AI and robotics. If we go back to the early 2000s, we were one of the countries that adapted most rapidly to high speed internet as well as mobile technology.because we had to, we talk a lot on the phone, obviously. So because of that kind of demographic instinct, we were one of the much faster adapters to that technology. I think that’s gonna repeat in AI and robotics. As you mentioned, I think that AI and robotics is definitely not, we’re not the strongest when it comes to AI robotics in the world.But in terms of adapting and using it for actual use cases, we may be one of the very strong countries. So I think there’s a lot of challenges and opportunities ahead of us.Grace Shao (04:50)That’s really interesting. think you hit something that like, you know, people are starting to pick up in the West, which is like in East Asia in general, the embrace of technology is a lot more optimistic. Some come from very realistic reasons. Like you mentioned, whether it’s in China or Japan or Korea, there is a potential labor shortage that’s coming to the next generation, right? But not only so, I think just in terms of culture and social sentiment also feels that way. So if you have to give like a high level kind of assessment onYou know, just the cultural attitude and political attitude towards AI, what does it feel like on the ground in South Korea?Ethan Cho (05:24)I think AI itself, I think people think of AI in different forms, obviously. I think as far as I know, China thinks of AI closer to robotics. The US thinks of it as, I don’t know, maybe a chatbot or something that they use for the industrial usage. In Korea, as far as I’ve experienced, I think it’s more about becoming a personalized kind of agent, not agent.per se in terms of doing purchasing or actual daily tasks. We always had kind of chat bots, especially for instance, for our financial system or the banking system, we’ve always used CS bots very frequently. So we’re very used to it. So I guess there’s less resistance when it comes to adopting or adapting bots or AI featured functionality, especially in the B2C area. So although...When we say AI just by the name, it could sound creepy. I think it’s already very well embedded in the Korean startup ecosystem and the overall society as well.Grace Shao (06:27)That’s interesting. So you did kind of touch on one thing, you know, the Chinese view this way, the Americans view that way, you know, definitely there’s a bit of a difference in terms of how maybe AI is being seen as whether it’s a political agenda or economic ⁓ aggregator or, you know, how it’s being diffused to the seaside. So in that sense, ⁓ my question is, is Korea building domestic AI champions or is it more right now kind of working around, you know, working on top of U.S. frontier models?or leveraging a lot of the Chinese open source models. Like how do we understand that in the ecosystem?Ethan Cho (07:01)Honestly, this is a personal kind of statement and my personal observation. I think it’s all of the above. I think we were going to talk about this anyways, but the sovereign models is something that the Korean government really wants. I kind of understand it in a way. The SOTA models, the state of the art models are obviously the ones that the enterprises want to use.But at the same time, think a lot more people and developers are looking into Chinese open source models, especially with the recent changes in cloud code and Gemini and everybody, who are basically hiking their token prices. It’s getting more and more expensive to actually do recurring work. And at the same time, think, including myself, a lot of developers are quickly finding out that the fine tuning of the models can only be achieved by very,⁓ almost redundant loop work, which can only be achieved through open sources if it is to meet economic sense. So right now, for instance, if we use a certain American LLM model to do these hundreds and thousands of ⁓ very repeated work, that costs a lot. So from an individual standpoint, that’s not really easy to achieve. So I think everybody’s trying to find that sweet spot of mixing those three models.Grace Shao (08:16)That makes a lot of sense. I think for stars, especially the ones that you work with, know, the economic driver is probably one of the biggest reasons why they choose what. So I do want to save the sovereign AI kind of piece for a bit later to help our listeners understand, you know, the create ecosystem a bit better. We all hear about Chibbles. We hear about, you know, obviously the big tech like the Samsuns and the whatnots in Eskihainix right now that are getting a lot of attention, right? Help us just even understand how these different companies and the startups, how they work together. Because for example, in China, a lot of that big tech are actually the incubators and initial investors of even the startups. So even the leading LLM labs, they actually have taken 10 cent Alibaba money. In the US, it does feel a bit different. There’s vast amount of venture capital money that are kind of funding the current growth rate of OpenAI and anthropics of the world. How does it?Ethan Cho (08:54)Mm-hmm.Grace Shao (09:09)ecosystem work in South Korea.Ethan Cho (09:11)So I think there was ⁓ quite a few phases of evolution. when the startups were really founded, that was, I don’t know, that was like late 90s. Those were very purely internet domains, internet online communities. Like that had not a lot to do with the Chebals. But then came mobile technology and everybody was starting to invent something on mobile. That quickly got the...Interest from the big Chebos, but actually as far as I know there were some Interests in very early on in neighbor and cacao by all these like really large companies in Korea But they never actually fully understood what that what that was and they kind of let them grow Which was a blessing for us at the end of the day. So neighbor and cacao was you know established and they grew like crazy after that thethe big companies quickly found out that, we need that DNA of innovation. They started to set up their own VC firms. They started to set up their own accelerators and everything. But as a typical CVC does, they inject a lot of money into their interest area, but not so much in let’s say, tangential areas. So because of that, there are definitely a strong value chain around semiconductors, for instance. But that kind of reinforces thealready existing ecosystem of the chaebols, which is not exactly what the startups are intended to do. So there’s kind of pros and cons there. And on top of that, after the capital was kind of concentrated into that value chain, the government kind of now is more active in kind of leading investments. large chunk of investments in Korea is led by the government.by the mother fund or fund of funds of Korea injecting money into the ecosystem and the other funds matching to that. So there’s a layer of chaebols, there’s the government and also the capital firms who are also ⁓ acting as LPs for the Korean startup ecosystem.Grace Shao (11:06)Yeah, that’s a perfect segue into understanding, you know, the government’s play. So I visited Korea just recently, I think last November, and, you know, it seems like there’s a huge policy push as well in incorporating AI into the everyday everything. And, you know, it’s top down driven. And like you said, there’s capital also injection. So how do we understand the government’s current priorities in terms of embracing AI? How do we understand sovereign AI andwhat kind of role it plays in the economy South Korea going forward.Ethan Cho (11:38)⁓ I think the government is definitely ⁓ making a very interesting and important bet. So there’s this huge initiative called Everybody’s Entrepreneurship, loosely translated into English. The government is actually injecting a lot of money into the ecosystem by giving money to...people who want to become entrepreneurs, who are novice entrepreneurs, first time entrepreneurs. I think it’s a good thing that a lot of people are trying out their ideas at the end of the day. The downside honestly is that entrepreneurship isn’t for everybody. So there’s gonna be people who learn their lessons the hard way, but still I think all in all, it’s gonna be a positive impact on the overall ecosystem. I think...⁓ The AI drive is definitely very serious for the government. As we mentioned earlier, there’s a labor shortage coming in. I think ⁓ East Asia most of the time has a little bit of issue with immigrants. We don’t shy away from immigration, but I think traditionally we don’t have the most welcoming immigration system compared to the States, for instance. So we’re trying to buy some time there, I think.And I think because the strongest point of the industry, as we also mentioned earlier, is semiconductor and hardware and technology, we want to build upon that. And because of that, I think AI seems to be a very interesting and promising area for the government and Korea as a country overall.Grace Shao (13:03)How do we understand sovereign AI though? Like what is the, I guess, reason for like, you know, maybe the non too large company, sorry, too large economies to really start pursuing this? Because we’re seeing this kind of rhetoric in the Middle East as well. You know, a lot of the local governments are really pushing sovereign AI. South Korea for sure has been openly talking about this. think, you know, a lot of European nations are also thinking about this. Is this just for, I guess, owning?Ethan Cho (13:06)Hmm.Grace Shao (13:30)the future infrastructure or how do we understand this?Ethan Cho (13:33)I think that’s one point. I think owning the future infrastructure is one. But I think if people are realistic, think we don’t want the world, we don’t hope the world is going to use our own sovereign AI. I don’t think that’s the case. What I’m expecting or I believe that the government people are wanting is that to use sovereign AI in very sensitive areas, such as our financial backbone, for instance, orYou know, because Korea is technically set or on the defense part, maybe we’ll use that for that purpose specifically. I think in the early days when everybody in Korea started to talk about sovereign AI, I was actually less persuaded. The problem is, or the status is, as we see all these leaks all over the place, like even for the top, you know, bleeding edge,builders like Anthropic or OpenAI, there’s always issues here and there. And it kind of shows. I’m not saying that sovereign AI is going to be perfect either. They’re going to have issues too. But if a foreigner comes, a foreign entity comes in and kind of screws up an operation, that kind of blame and whatwhat something domestic spills over. There’s going to be a different kind of anxiety in the society, I guess. So that’s maybe the angle that they’re kind of anticipating. But I mean, I am worried a little bit too, because I’ve seen multiple cases of software built in Korea domestically, which has never been successful to Korea. And it just has been kind of a wanted wonder just in Korea. I just hope that doesn’t repeat. But we’ll have to see.Grace Shao (15:06)Actually, this is not completely rated to AI, but just on that note, why do you think a lot of the times like Korean companies are huge? Like you just mentioned Naver and like Kakao or like, you know, Japan, and LINE and China that we chat when not like these internet companies never really go abroad. That’s just intellectual curious question just on the topic.Ethan Cho (15:16)Mm.I’m a kind of linguistics buff. So I actually think the reason is in the language. The language that we speak is just different from English. because of that, think the like Naver, Kakao, WeChat and Line are all basically rooted in language. And because of that, that’s just universally different from WhatsApp, for instance. Like it’s not language per se, but if you look at WhatsApp, how they control their UI UX, for me at least, is very boring.Grace Shao (15:27)Mm.Ethan Cho (15:52)⁓ I would prefer a cacao or lime or WeChat much over WhatsApp if I could choose without being specific in which geography. So I think there’s a cultural preference, a very strong cultural preference that really is hard to translate across territories.Grace Shao (16:09)Okay, that’s an interesting take. Yeah, because I think it’s interesting because like, it’s basically the West has this one or even like, you know, Africa based Southeast Asia, they all fall under the American big tech kind of umbrella. And then Korea, Japan and China’s have such strong domestic players. Actually, on the note on you know, the consumer side of things, what are some interesting trends you’re seeing out of South Korea in terms of consumer AI right now? What are some companies you’re investing in that are in the consumer AI space?Ethan Cho (16:16)Yeah.Consumer AI, think, is still at a very early stage of growing. I think right now the most used cases that I see with my bare eyes are actually foreign tourists coming to Korea, visiting like big K-Beauty.department stores like Olive Young, and they go and get their skin scanned and they analyze it with AI and give recommendations to basically ads. But still, I think that’s a very clever way to scientifically analyze the customer demand. I think a lot of players, and I see a lot of players trying to replicate that into basically recommendation engines. Personally, I think that’sclever, but it’s not good enough. It has to get better. The trade-off there is obviously privacy. So if you want a Uber personalized recommendation, you have to somehow yield on the privacy part. think we’re still not clear on where is the kind of safety line. So I think they’re still kind of struggling towards that. We have invested mostly these days in consumer brands because how⁓ I think of it at least, is regardless of which AI becomes the winner or winners, I think as long as we have the best product in our portfolio, if the AI is clever enough, it will choose that product. So before actually deciding which AI algorithm will actually win the war, think we’re trying to get hold of the monuments before the, know, who,before we decide who becomes the winner of the war altogether.Grace Shao (18:12)So you’re looking at brands as in like, like retail brands. What are you looking at? Like, okay.Ethan Cho (18:16)Yes, yes, for now,yes. But at the same time on the kind of the hardcore AI part, we also have invested in sovereign AI companies like Trillium Labs, which actually develops SLMs instead of LLMs. We’ve also invested in a drone company called Bone AI, which does physical AI using drones and they’re targeting the Korean big defense industry. So that’s kind of themore of the hardcore AI part that we’re looking into. We’re still waiting for that sweet spot where consumer meets AI. I think that’s still kind of in their very early stage in Korea.Grace Shao (18:46)I see. I just want to say it’s so funny you used the Olive Young example, it’s really topical. So this is really a bit of a rant, but my friends and I were saying we need to go to South Korea to do the color palette, right? And all, you know, all the girls are like raging about this right now. I literally asked Claude Coe today to do it for me and it gave me the whole like color palette assessment. I was like, wow, I just saved myself a flight and a trip to South Korea. So definitely can see like there are consumer uses in that end, but I guess what is the monetization from that bud, right?Grace Shao (19:22)So that I can see how that will be hard to invest in that space. On the consumer end, know, again, I read headlines about South Korea, right? And, you know, I hear about companion bots being really big. I know actually even in China, they were group bots. They were like so-called boyfriend, girlfriend bots. And to your point, you know, South Korea, China, like even like a lot of East Asian countries are in are all faced with this issue where there’s mass urbanization.Grace Shao (19:48)loneliness issue everyone is like, you know faced with evolution and competition so they don’t have human companion and Do you see this as a trend and do you see this as something that potentially would not be actually within the cacau’s and the lines of the world that could be a spun-off on its own and to fall off of that I was just even just kind of thinking because Korea has so much IP right now in obviously k-pop and k-dramaEthan Cho (19:57)Hmm.Grace Shao (20:16)would that potentially be a vertical where they can tap into basically creating, you know, like companion bots, based on existing celebrities.Ethan Cho (20:25)I mean, I think yes on both is the short answer. I think the boyfriend bots, girlfriend bots are very popular in Korea. I think my son is also using one. I haven’t talked about that openly, but I think so. And yes, I think that Kakao and Neighbor would be very cautious about adopting that technology into their existing platform. There can be some...many opportunities to abuse that. People, as you know, once these bots are online, the first thing that everybody tries to do is abuse it one way or another. So I think a cow and neighbor would probably shy away from that and go into commerce, which is always what they wanted. We’re seeing more specialized startups that are just doing this boyfriend, girlfriend bots.like some are more adult focused, some are more teenager focused. So there’s definitely kind of a breed that’s coming out of that. On the K-pop and K-drama bots, I think that’s something that a lot of companies have worked on for quite a while. For instance, like Weverse, is the, it’s the entertainment company that basically ⁓ relates to all K-pop stars. They have their own like personas.So they actually provide not just only conversations on bots. I think they also give artificial voice calls. They’re already there. So you can have a conversation with your favorite star. It’s just not realistic enough yet. But I think they’re getting there. So ⁓ that’s definitely already happening. I think on the IP side, personally, think it’s more about how can you make these into really long-lasting legacies? Even for BTS and like...girls generation, which are the you know, the idols of the day. I think they’ve only been around for 10, 20 years. Like, can we make this into like decades long, right? Like a legend, like can we actually make this into something that goes through generations, not just decades? That’s a big homework for us to figure out and make them kind of timeless.Grace Shao (22:24)No, I totally see that. It feels like a Black Mirror episode with the Miley Cyrus ⁓ kind of fake doll as well. But I think to your point, there’s the IP issue, there’s obviously the security issue. There’s the psychosis issue. This is like a much bigger issue. I think that we require regulators to work with businesses, right? I want to kind of move our lens to the enterprise side of things. You you mentioned just now like Naver and Kakao wereGrace Shao (22:50)looking into maybe going to agentic AI and maybe even to commerce. Are we looking at something like what Alibaba is trying to do where you have agenda commerce through a one entry point, you you interface with a chat bot next thing you know, like a bubble teas at your door. Like, is that the future of commerce you think or what are we talking about here?Ethan Cho (23:11)I think, I think, neighbor and Kakao are both in fierce competition with coupon coupon is the dominant e-commerce player in Korea. ⁓ I think it’s a real headache for both of them because coupon was kind of non-existent. They didn’t have a lot of user interface and neighbor and Kakao were kind of self satisfied that they dominated the user interface. But, here comes coupon and they just basically just crushed every.aspect of e-commerce and is by far the number one player in Korea. So that’s something that Naver and Kakao are trying to battle. The only difference, as far as I can see, that they can make is real-time purchases. If you want milk at your home the next day, coupang is much easier and much better. That’s kind of a fact. But for Naver and Kakao, because they have basically 24-7 access to your daily life, if they can actually monetize on that, I thinkThat’s their way to go. The competition there is also not non-existent. That’s a problem. YouTube’s there. TikTok’s coming along in Korea. TikTok’s still small in Korea, but it’s growing rapidly. that space is also... I think people think... Koreans are just so big YouTube fans. The YouTube dominance is so... They used to. Now they’re getting more used to short forms now.Grace Shao (24:15)Why is it small? Curious. Why is it small?So they like long form.Ethan Cho (24:30)And the TikTok trend is definitely coming along. But I think most YouTubers, so-called YouTubers in Korea, are long-form originated. So the trend is changing now. So yeah, it’s a little bit slower to adopt. yeah, sure. Oh, yeah.Grace Shao (24:40)And I wanted some context. So coupon is like an Amazon or like it sounds like a DoorDashand like how do we understand this just for our American audience or Western audience?Ethan Cho (24:51)Yeah,Coupang is, they actually literally say that they want to be the Amazon of Korea. So they are the Amazon equivalent in Korea. Their main business is e-commerce. They have Coupang Eats, which is DoorDash or Uber Eats. They have Coupang Play, which is the Amazon Prime. So they’re basically, Coupang people would hate me saying this, but it’s kind of like the Amazon replica in Korea.But they’re doing a fascinating job. Their killer feature is next day dawn delivery. So if you deliver, if you place your order by midnight, they’ll get the item to your doorstep before 5 a.m. So it’s marvelous for. It’s not just groceries. Yeah, so they’re very good at demand expectations. So they have a lot of warehouses in Korea. So they fulfill them in advance so that they can basically distribute almost within like four or fiveGrace Shao (25:29)It’s not just groceries, it could be anything.Ethan Cho (25:43)hours window, which is also honestly possible because Korea is not so big as a country.Grace Shao (25:49)But it sounds kind of like almost a JD.com business model as well. It’s like, but more high, high, more expensive. Excuse me.Ethan Cho (25:52)True. Yeah, yeah, I think I think that’s a fair. Yeah, that’s a fair comparison. Yes.Grace Shao (25:57)It’s a bit more expensive, right?Yeah, so I think looking at that, then, you know, there’s also this rumor, or I guess it’s actually been verified that South Korea was, fact, OpenAI’s largest enterprise market outside of the US, which is crazy, because like you just said, South Korea is not exactly that big of a country. Why is that? Who are the people buying up all these tokens?Ethan Cho (26:12)Hmm. I don’t have exact numbers, but when I first heard that I wasn’t too surprised because if you look at like Koreans are very used to buying tokens online. Like that’s why Korean gaming has beenor what used to be so big, especially in the mobile era, because people were just fine with buying items online, like purchasing it like crazy, which was kind of now it’s kind of standard, but like back in the days, like early in 2000s, like it was a very weird phenomenon if you look at it from a global standard. So because of that, think people are really, really fine with just buying tokens and buying memberships, which costs 100, $200. I think that’s kind of whatthe base layer, so the willingness to pay the first layer. The second layer is that Koreans love to build things. Just trying to build things so much with their own hands is one tendency that we strongly have. So because of that, I think most of that building tokens go to my stock portfolio optimizer or kind of tools like that for personal use. But I think people just like to try out a lot of things that kind of led to that consumption.Grace Shao (27:26)Mm-hmm.Ethan Cho (27:33)On the B2B side, think as far as I know, because the head of OpenAI Korea used to be one of my kind of bosses at Google Korea, he was, well, OpenAI was very aggressive making contracts with Samsung and SK very early on. So I don’t know how many tokens they’re using, but just thinking about how many employees they have, if they struck a good deal on a B2B,business, I think that would be a very significant portion of tokens being burned in Korea as well.Grace Shao (28:03)That’s pretty crazy. So basically you’re saying the the enterprise side, people have pretty strong connections and reasons to buy and the consumer side are just willing to shell out subscription dollars. That’s very different from, would say, the Chinese market where there’s just like not a lot of willingness to pay from consumers. And hence why we saw all the consumer apps in China were all free. So I actually want to ask on enterprise end. So what are we seeing people spend money on in terms of AI thatEthan Cho (28:14)Yep.Grace Shao (28:30)What are people trying to build? Are we looking at like also like on the enterprise and are they trying to solve co-pilot like solutions? Are they trying to serve customer service issues, manufacturing optimization? Like what are people really focused on?Ethan Cho (28:43)So just based on my experience with the companies, I think one thing is the banks are very serious about building the CS layer via AI. So they want to substitute a lot of that labor force into AI. I’m not sure whether that’s like how fast that can be optimized just because people are very demanding in Korea. know, when even if you use the traditional kind of phone CS, peopleend up basically talking to people. They demand to talk to an actual person instead of going through the automated call. So we’ll have to see how the ROI comes out on that part. I know that there’s a lot of AI being used for the semiconductor processing and producing process, but that’s just not public information. So we really don’t know how much is being used there. So that’s on the enterprise side. On the consumer side, think because of everybody’s⁓ entrepreneurship program that I mentioned earlier. I think there’s a lot of people that are trying to use a lot of cloud code, for instance, or codecs from OpenAI to build programs. There’s a lot of events actually held in Korea. Maybe every week there’s an event from OpenAI or Anthropic basically, which is like the cloud ambassadors night or the OpenAI something, something night. people, lot of... ⁓the AI builders are actually encouraging Korean builders to use their own tools by giving out a lot of free tokens actually, like thousands of dollars are given out as tokens just to nudge them into building. So there’s gonna be a lot more activity in that space for sure in Korea. And hopefully there’s gonna be something that’s really interesting coming out from that.Grace Shao (30:25)That’s interesting. I did want to ask, you kind of mentioned this earlier that you guys even invested in a drone company. South Korea obviously has a very strong manufacturing sector, home appliances, phones, cars. How are we seeing this whole, we have generalized this whole sector kind of lean into AI? Are we seeing physical AI being prioritized? Are we going to see? more robotics coming out of South Korea. How do I understand that?Ethan Cho (30:55)I think there’s still some uncertainty there because of the all of the among all the Korean robotics companies, I think the most technologically advanced one is Boston Dynamics, but that’s not a Korean Korean company, to be honest, right? Because it used to be an American company acquired by Hyundai Motor Company. So there’s that. There are quite a few robotic startups that are starting in Korea.Just because we have Samsung, Hynix, and Hyundai, think the manufacturing industry obviously is a great application area or a market to sell to. So we are seeing a lot of robotics company coming out from the university as well as startups. The big question here is will they scale? That’s kind of the pressing question. I mean, I think the companies, for instance, for Coupang,which does all the logistics. They’re heavily using robotics just as Amazon does. So those robotics are already deployed or are being deployed. But for instance, humanoid robots, which China is leading the way, I think that’s still a long way to go for us. And we’re trying to figure out what the application should be. So one interesting example, I think China has this too, but...We have all these little, really cute delivery robots going down the road and trying to get food to their neighbors. That’s an experiment that a lot of companies are running right now. We also have small police robots that are also running around just to do surveillance. I think it’s a cute initiative, but can this scale is going to be a big question for a lot of us. So I think this is also intertwined withautonomous driving landscape in Korea, which is still kind of not there yet. So I think there’s going to be a lot more of this going forward.Grace Shao (32:44)Yeah, no, actually on that note, I just was in Shenzhen last week and I saw one of these like, you know, street sweeping robots per se, stuck in a puddle. And it’s like, to your point, like they look cute or like, you know, you have little robots delivering your phone charger in hotels, but they’re not actually that scalable. And I don’t actually know if they’re that cost effective is the issue, right? Because, you know, sometimes hiring a person to sweep the floor, frankly, inEthan Cho (32:48)Hmm.Grace Shao (33:12)a market like China is actually not that costly compared to even deploying a robot like that and then having to, you know, maintain it. So I see your point. Okay, I think, you know, I want to shift our focus back to, you know, your bread and butter. And I really appreciate you patiently breaking down the ecosystem for me as an outsider who don’t understand South Korea that well. But as a venture capital investor right now in South Korea, what are your, I guess, most interested areas?What kind of founders do you really want to invest in? And are you looking at the founders more or the companies more? Let’s start with that.Ethan Cho (33:45)I am looking for founders. I’m looking for founders because I think there’s been a evolution of generation or a change of generation that I’m seeing. I see a lot of Korean.like in their 20s or their 30s who are educated abroad, come back to Korea and start working in Korea, not too happy about their job and trying to figure out what to do next. I just want those people to actually start something new and I want to kind of back them. I call that like global ambition, but local execution. I think that’s something that we need more.Until now, as you know, all the companies that we’ve mentioned throughout this conversation, like Naver, Kakao, Coupang, they’re all basically really focused on the Korean market, which was, you it’s good. But still, as we all know, Korea is not the largest of the countries. And, you know, just doing business in Korea doesn’t mean a lot, especially as we move towards AI more and more. And because of that, I just want those...⁓ kind of people who are ambitious to really change the world in a significant way, not just build the next chatbot or the next food delivery app, but something that kind of, you know, breaks around and just changes something very significantly. That’s something that I’m really looking for these days.Grace Shao (35:00)That’s really interesting. Do think that has anything to do with your upbringing, just being so internationally exposed?Ethan Cho (35:05)Maybe, actually, yeah. think, this is kind of another personal note. think Asians are really smart in a lot of settings, but we as Eastern Asians, were brought up to be kind of modest and humility was one of our very top priorities as we grew up. And because of that, we tend to be more humble in front of people. And as we know,The Westerners, like this is not a great word maybe, but the Europeans or the Americans are much more aggressive in PR, but we tend to be more careful about that. Back in the days, that was great when we were just living amongst ourselves, but now as we go into the global market, PR is really important and having big ambitions like shoot for the stars, land and the moon is the way to go. But sometimes we just focus on what we have. I think that’s a healthy way of living, but.For entrepreneurship, we really have to dream bigger dreams.Grace Shao (36:02)That’s really interesting. I think it’s some things that I’ve even really noticed within the just generating Chinese founders as well. It’s really different. Like you mentioned coupon. I think the founder was Harvard educated, but he returned to Korea focused only on the cream market, just like the last year. He’s like the JD.com Alibaba’s and the day these are Chinese market businesses. They have global footprint, but there no one’s thinking of them as a international business, right? At the core, they’re Chinese company. But if you look at theEthan Cho (36:18)Yep.Grace Shao (36:28)whether it’s the LLM companies in China right now, or even some of the more consumer facing ones, or even the robotics ones in China, I kind of feel like there’s a shift in generational behavior. Exactly to your point, some of them are less educated than they’re not, but in general, people are not as modest. People are actually more, not in a bad way, but they’re much more open to doing PR for themselves. Well, not just PR, but actually flexing and going more ambitious, going global.like we said, like Kimi and Minimax, whatnot, Jiu-Jitsu, they’re used globally, right? And they’re not shying away from it. I think that’s really interesting. That’s like a phenomenon across East Asia right now. So I think for us to understand, what are some, I guess, misunderstandings or things that foreigners who are trying to invest in Korea often...you know, get wrong or not completely get correctly because, know, obviously there’s so much societal nuances. Well, in South Korea is a country where I find, like you said, it’s not not only not that immigration friendly, but actually in some ways a bit more closed off, Much like East Asia in general, like if you’re not from there, you don’t speak the language, don’t understand formality, especially South Korea has a lot of formalities. It’s really hard to do business, right? So how do what are things that you see that foreigners might be getting wrong that they could do better?Ethan Cho (37:44)Hmm, I think, well, I mean, first of all, think Koreans are just a lot of time. I wouldn’t say everybody, but a lot of Koreans are just shy. They’re, they’re friendly, but they’re shy. That’s kind of our kind of default mode. So, you know, if somebody comes to Korea and nobody wants to talk to you, that’s the norm. But once you try talking to any random Korean person, he or she will definitely help you out. That’s kind of the Korean kind of way. They’re being shy because they want to be polite. That’s kind of an Asian thing, right? So.There’s that. think because Korea has been such a small country, think people, some people think of Korea as just being focused on that very regional kind of market. We’re not, obviously. Like we want to also go global, but we just didn’t have enough chance to actually show off that.I mean, if you look at, for instance, the Koreans working in the States, they can show you what a Korean can do if they’re put in the right setting. So I think if you’re an investor and want to work with a Korean firm, think as long as you put the resources and the human talent in the right settings, they will perform. of course, I can’t guarantee everybody will, but in general, that’s how we’re formulated.I think one interesting factoid that I also always kind of want to emphasize is I think China is also similar to this, but because we have this crazy, crazy education system that’s like overly competitive, although we have been really stressed out throughout our teenage years, that actually made us very, very competitive when we just, you we’re put in the right settings. Like we will strive to become number one in whichever setting that we are put into. So just.like, you know, help us get to the right market and get to the right country or what right settings we will perform. So that’s, think, the expectation that you should kind of have for a lot of Koreans and Asians in general.Grace Shao (39:39)just like whoever can go through the national like university exams, like they have resilience. These buddies don’t like they don’t mess up. So I think on that note, I guess I want to ask what are what should we expect Korea to be exporting that if you’re saying that you want to back companies that are going global, you want to back ambitious internationally minded creates, what should we expect? Because no one expected it. Well, not no one. But a lot of people did not expect China to suddenly be exporting LLMs, right? As like one of their hottest new technology right now. I think a lot of times robotics maybe, EVs maybe were more in the expectation over the last five to 10 years because of the strength and the slow momentum it was gaining, right? But yeah, for Korea, what should we be looking at? Like you said, obviously hardware, chips, there were a lot of synergy there. You’re trying to build on top of that, but beyond that.Ethan Cho (40:29)I think the low hanging fruit or the easy pick is K beauty and K fashion. That’s definitely gonna come in the next ⁓ coming three to five years. I personally think there’s a lot of interesting angle in the Korea defense industry combined with AI because honestly speaking, there’s a lot of, how should I call this? Like confusion around the American diplomacy.policy recently because of all these international tensions. And because Korea has always been at war technically with North Korea, I think there is a lot of advancement in Korea technology wise. We have the best semiconductor in the world. I think we are one of the most flexible countries when it comes to, are you going to use US LLMs versus Chinese LLMs? Like we can do both. think we are.We see the pros and cons there. very flexible there, so we can optimize. I think because of that, the defense industry is not only growing very fast. I think it’s a very good place to kind of experiment the new warfare technology without going into actual war. And because of that, I think the Korean defense industry will benefit a lot from AI evolution.Grace Shao (41:41)I see. It’s an interesting area which I’m not like I’m not familiar with at all. But it’s like kind of like you said, it’s kind of one of those areas where you don’t really hope it being really used, right? ⁓ But it’s definitely a very hot space in terms of VC investment and in the US, especially with Palantir driving over the last couple years. Again, not really to AI, but I kind of want to double click on K beauty and K fashion. Why is it like what what is it that you know, over the last week? SoEthan Cho (41:50)Mm. Yeah.Grace Shao (42:08)My husband and were trying to talk about this very casually that day. We’re like, wow, we live in Hong Kong. In the 80s and 90s, everyone was obsessed with Hong Kong pop stars and Hong Kong movie stars across Asia and then even globally. They had all the kung fu shows and then all the police shows. And then in the early 2000s, we definitely had the Taiwan wave, the Taiwan pop stars coming out of East Asia. And even I was in Canada. I was growing up Canada and people loved Jay Chow, right?Ethan Cho (42:29)Mm.Grace Shao (42:36)Nowadays, obviously, it’s all about Blackpink, right? So how does this move around? Why is it going around? And how does one society kind of, I guess, nurture or incubate a global pop star? Is it tied to geopolitical reasons or economic reasons? Or do you think aesthetics?Ethan Cho (42:58)I don’t know exactly the reason if I knew I wouldn’t be working in V.C. I would be another producer. But I think there’s two reasons that I think is the biggest reasons. One is we have a massive farming system, you know. Koreans train boys and girls in their early teens.Grace Shao (43:04)Yeah.Ethan Cho (43:18)to become the K-pop stars and you have to go through vicious vicious competition to actually get there. So because of that, I think there’s so much talent that’s going through that pipeline, which is a blessing and occurs at the same time to society, obviously, I think so. But there’s that. The second part is I think Korea had a mix of...American culture very early on because of the Korean War and the Korean forces, sorry, American forces staying in Korea. So if you look at Blackpink’s music, because you quoted Blackpink or even BTS, there’s a lot of African-American music, ⁓ like features within embedded in that music line, in the melodies. It’s very some of it is reggae, some of it is very hip hop. And those kind of cultural fragments were embedded very early on because we hadmore exposure to African American or hip hop music versus let’s say China, which didn’t have American troops staying in China. So there was that. Then somebody might ask, what about Japan? They also have a huge American troop there. I think Korea was because, maybe because we were a smaller country, we were more open to actually getting into and using those vibes. And because of that, think.The Korean kind K-pop or K-beauty, K-fashion factory has become a little bit more westernized early on and that kind of made the entrance barrier a bit lower for the American market. That’s kind of my hypothesis.Grace Shao (44:49)Yeah, because if anything, kind of going back to your point on like South Korean and East Asian companies and people don’t really do a lot of marketing and PR, I would say ⁓ Korean cultural export has been extremely successful and has been a really, really strong soft power export. So I want to end the conversation on again, back to AI. What are some things that you think we might have not covered today? You think we’re missing? Like, what are some trends?Or say like if we really spoke again, let’s hope not two years later, but let’s say we spoke two years later, what would be true for you to think of how society has evolved, what Korea has maybe contributed in a global AI supply chain ecosystem, how to understand how you view the future.Ethan Cho (45:32)One thing that we haven’t touched that I’m personally passionate about and interested in is the mental health industry. It’s going to be very different from now versus three to five years down the road. As we know, the fitness industry, physical fitness industry, has become a huge industry ⁓ after the Industrial Revolution because people started to use less and less of their muscles. I think that’s exactly going to happen for our minds and brains.⁓ And because of that, this is not going to be driven by AI, but it’s going to be kind of a side effect or a secondary industry from the AI revolution. To keep everybody healthy, think this is something that we as a society and company as country has to work on. there’s going to be, I don’t think this, I don’t necessarily think of this as a dark scenario. I think as we go to the gym, we can go to this mental gym or something.very on a regular basis to keep ourselves healthy mentally. I think that’s gonna be something very huge. Until now, I think we’ve focused a lot on the hows, like how are we gonna do this? How are we gonna do that? The answer to that has been AI and robotics. There’s gonna be more and more questions about what are we gonna build with this? And after that, there’s definitely gonna be questions about why, why are we doing this? I think that’s not just gonna be philosophical, but it’s gonna be a very practical question.that will lead to a lot of business opportunities. So I think that’s something that we’ll have to question ourselves and answer and discuss on a very regular basis down the road to reach something meaningful either as an entrepreneur or an investor.Grace Shao (47:07)I think that’s really, really meaningful. And I think, you know, we kind of touched on like companion bots and even your you mentioned your son might be even using a companion bot himself. I don’t want to probe on a personal level, but actually on this note, then how do you view that? Like, do you ever fear that he’ll be too dependent on it or, you know, I could be creating a false reality?Ethan Cho (47:27)I think it really depends, right? I know this is not the best answer, but like I’m a big fan of the movie, Her. I think it was a very, very good example of how things can evolve. The ending was kind of sad and happy at the same time, but until then, he was very happy with Samantha. So it seems like there’s definitely a scenario where we can be more happy about the world, be more thankful about the world, thanks to this.Grace Shao (47:33)Mm.Ethan Cho (47:52)maybe emotional buffer that we create with our AI companion. There’s definitely that. But there’s also going to be a downside because the companion will feel real, but it’s not going to be real. So how can we cope with that? It’s going to be something. I still think it’s going to be very similar to the fitness industry just because when we do like, you know, bench presses or, you know, like all these like that pull downs, those are not actual resistances. We’re creating them artificially to strengthen our muscles.So I think our minds should also be strengthened in that way so that we can cope with all these scenarios that we’re not gonna be able to actually experience down the road because we’re gonna live in our own world, which is gonna be safe and creepy at the same time. you know, a lot of factors that will change down the road. So kind of excited and horrified at the same time.Grace Shao (48:42)No, 100%. I think your point on mental health, you use like a general term, but there’s obviously the obvious fear, like what we just talked about, like psychosis and dependency, but there’s also kind of like you mentioned, touched on like, you know, if we don’t really use our brain that way, you don’t really know how to do it anymore. Just kind of like languages, you know, when you move to a country and you don’t use that language for a while, you lose it. Math, I like literally don’t know how to do math anymore. It’s pretty sad. But you know what I mean? Like if these are skills where like you kind of havepush yourself and the gym is something quite, if you think about it, very arbitrarily created for our modern day lifestyle, which obviously didn’t exist even like two generations ago. But yeah, like I think that’s a really interesting take. I don’t know if that’s actually your differentiated view, but you know, I usually always ask one last question to every single guest that comes on the show,what is one differentiated view you hold? So something that might be a bit non-consensus, it could be provocative, it could be not, know, it could be about industry, it could be about life. Honestly, I think what you just said earlier was a bit, it’s quite insightful. It’s something not talked about in the mainstream enough, but if you have another one.Ethan Cho (49:31)differentiate the view. huh. I can make really dangerous comments here. ⁓ but I think,Grace Shao (49:55)No worries.Ethan Cho (49:56)yeah, this is one thing that I always think about. So I think that the creator cannot make something that the creator has not experienced. That is something that I think deeply about, and that is my personal view on the limitations of AI. How we think about AI is to become this everlasting thing that works 24-7, does only good things for humanity. Buthave human beings actually ever experienced that? I don’t think so. And that’s going to be a big question because we’ve never, we don’t know how to work 24 seven. Well, of course we’ve, you know, we’ve done all nighters for sure, but can we actually think of a process that can continuously work 24 seven by thinking, not just operating machinery and also can we think of a kind of standard that is always only helpful to human beings? Like we haven’t really done that.So I mean, I think that’s gonna be a big challenge. Like, however we construct the system or the standards for AI and robotics and all these systems going forward, there’s gonna be a loophole there. And that’s something that we’re gonna have to figure out as a society as a whole. So I think that’s gonna be something that it’s gonna be very interesting down the road.Grace Shao (51:08)Do you kind of, are you kind of alluding to what we’re seeing right now? A lot of people have AI fatigue where they actually make the agents just work 24 seven for them. So essentially the moment they just stops doing a task, they repeat, like they let’s restart it. That’s kind of the work, right?Ethan Cho (51:20)Yeah, I think so.that definitely shows what the problem is. Because we don’t know how to operate these. So I’m facing the same. I don’t use it as much as I used to like a month ago because of that. Because I’m feeling that, this is controlling me, not me controlling that. So there’s this reverse effect. So I think it’s a good thing that a lot of people are already kind of figuring that out. people are kind of.trying to like healthily distance themselves from all these agents. So that’s, I think, a positive sign. But I think as a society as a whole, that there’s going to be more and more things that we’ll have to think about.Grace Shao (51:55)No, I totally agree. And I think there’s certain things. There’s a lot of value in stopping and thinking about the action before the action respoots again. Obviously, there’s certain repetitive work that can be streamlined. so much of accessing knowledge work. mean, this discussion can go another hour, but so much of the whole argument on knowledge work being completely replaced just seems a bit I feel naive for me. Like, I feel like so much of the knowledge work actually requires us.creating things and I don’t know maybe I don’t understand technology well enough so who knows maybe it can create things on its own. Ambanao, I really really want to thank you for yourEthan Cho (52:26)Yeah, that’s another. Yeah,thank you. Thank you, that’s another hour of conversation so we can do it next time.Grace Shao (52:34)Yes, please. Thank you.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Assembled co-founder John Wang on building a AI native support system for enterprises 04.05.2026 43min

  In this episode, I sit down with John Wang, the co-founder of Assembled, to explore how AI is revolutionizing customer support. Having transitioned from a Stripe engineer to an AI startup founder, John shares his unique insights into the evolution of support tools. We delve into how these tools have shifted from being mere cost centers to becoming strategic assets that enhance customer experiences. John and I discuss the impact of AI on support volumes and staffing, highlighting how integration is reshaping the landscape. He emphasizes the importance of talent density and assembling high-caliber teams to drive success in the tech industry. Through his experiences, John provides practical insights into AI's current capabilities and limitations in support operations.We also explore the strategic considerations for future AI support ecosystems. John shares his thoughts on the role of support in driving revenue and customer satisfaction, and how AI can orchestrate with human support agents to create a seamless experience. His perspective on building high-performing support organizations offers valuable lessons for anyone looking to innovate in this space.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, founders, and product managers.For more information on the podcast series, see here.To find the previous episodes of Differentiated Understanding, see here.Chapters00:00 The Journey from Stripe to Assembled02:25 Understanding the Importance of Customer Support05:29 Lessons Learned from Stripe10:25 AI in Customer Support: Current State and Future16:04 The Economic Impact of Support Operations18:25 The Role of AI in Transforming Support Jobs24:30 The Future of Support Organizations26:58 Guardrails Against Fraud in AI Support32:42 Navigating the AI Ecosystem38:00 The Value of Long-Term Commitment in CareersAI-generated transcriptGrace Shao (00:00)Hey, John, thank you so much for joining us. I just recorded your bio already. It’s extremely impressive. And you’ve done quite a, you’ve had quite a few different roles now as the co-founder of assembled, right? To start, can you just tell us about your story? Like what inspired you to leave Stripe, you know, go into, you know, right now what you guys are doing, which is a software for people who run customer service support operations. You know, now you guys are pivoting into AI as well, or at least leaning into AI. Tell us about all of this.John Wang (00:28)Yeah, great question. When we, well, when my co-founders and I started, we were all at Stripe. We worked on a bunch of different things at Stripe. And one of the last things that my two other co-founders worked on was a support tool, an internal support tool. And I remember pretty clearly that they were making a bunch of headway. It was really, really cool. And...They had gone to this really, really high up person and product. And this person was basically like, why are you guys wasting your time on this? Like you guys are kind of like, you’ve been at Stripe for so long, you know all these things and you’re doing support. Like I’ve got this really cool Bitcoin project that I would love for you to work on instead. And I remember my co-founder coming to me and being like, hey, like pretty bummed this is what happened.And then I was like, wait, you just saved Stripe, you know, quite a few million dollars, increased customer satisfaction by 40%. And still they don’t understand the value of this. And that’s when we were like, hey, ⁓ there’s something here where there’s a market opportunity. So that’s what got us really, really excited about support. We were doing it at Stripe. We knew it was an undervalued place. We didn’t see any very good tools out there to do support well.And so we decided to go build something really, really great in the support space and just like make transform and elevate support is our mission. Yeah.Grace Shao (01:50)Do you think it was just that stripe was too rich? They were just, and they just didn’t care about saving a couple million dollars? Or do you think it was actually a blind spot for people?John Wang (01:59)I think Stripe was definitely very rich at the time. think it was also a blind, it was a combination, right? Because most people, you think of support, you think of it as just a cost center. And I think recently that started to change in the sense that like, hey, this is actually a really important part of your business. But for a lot of companies, like if you look at FinTech, if you look at like a lot of health tech companies, their entire product is their relationship with their customers.And so support’s actually really, really important for that. And I think a lot of people underappreciated that for quite a while. And now I think people are starting to understand again, hey, if you piss off your customers every time they come and talk to you, that’s not going to be a very good thing. You better be a monopoly. Otherwise, you know, they might not be coming back.Grace Shao (02:46)Yeah, definitely. I think I want to kind of lean into that later in our conversation as well. It’s like people are trying to replace support and customer service AI first. But if anything, it’s not the best experience when you’re frustrated with a product and you keep on getting a robot, right? But I want to kind of talk more about your experience at Stripe. You were there quite early. What do you think it taught you, you know, as a very early employee at such a successful startup now?if even considered still startup and then like what were things that you think you learned there lessons even if soft skills that you kind of took away to to your current role like as a founder.John Wang (03:21)Yeah, it’s a great question. You know, it’s really funny actually. I just met up with someone where, so when I was starting out of college, I had applied for all these jobs. I was able to get a lot of them, except for this one company that I really, really wanted to go to. It was called Meteor Development Group. They built open source software. In college, I had built open source software at Ruby on Rails.I was really big in that community. was like, wow, it’d be awesome to go and make this something I do day to day. And I didn’t get the job. I was really bummed about it. And then I was like, I’ll just fall back on my second here, which is Stripe. And Stripe was the obvious second choice because just the people were really, good. And now like 10 years later, I like think about that and I’m like, the business model is really important.because Meteor was not a good business. Like open source frameworks is not a good business, but Stripe, really boring. Honestly, it’s just like payments. You process payments, you go talk to Visa. You literally have to like, we had a server in the server room that would send like a specific file with specific tabs and spaces in order to get it out to Visa. Really boring. Really, really core infrastructure too.And so like the big overarching thing that I learned was like one, business model is unbelievably important because if you can just make a good product when the kind of like market is there and when there’s a really big need, then this can scale like unbelievably fast. Two was the people. I remember talking actually to a few people, Greg Brockman was maybe the second or third person I talked to.who’s now the co-founder of OpenAI. And I remember just talking to him and being like, wow, this person is so, so smart. This is awesome. And I would talk to kind of like, I would go to the lunchroom and be talking to people at Stripe. that was just, people were talking about all sorts of things. And I think like talent density was a really, really big part of like what made Stripe successful. AndIt wasn’t any one thing over time. was one, Stripe was in a great market. And then two, it iterated really, really fast on a lot of little things over and over and over again. So I thought that was a really good place to learn a lot about like what makes a company great.Grace Shao (05:49)Yeah, I think it’s interesting you’re talking about talent density and a lot of the AI labs I speak to actually also talk about that. But I’m curious, what does it mean when you have really strong talent? Is it like that they are technologically superior, like they can code better? Or does it mean actually that they can think outside the box, they’re more creative, they can pivot faster? Like what does it really mean to have really high caliber talent on your team?John Wang (06:11)I think it depends on what company or like what you’re trying to solve, right? Like talent density for Los Alamos national, like Los Alamos, like building the atomic bomb is like very different than talent density for like Bell Labs, which is very different than talent density at early Stripe, which is very different also than talent density at OpenAI Research. Like I think for Stripe in particular, the type of talent density that was there was really high curiosity.Grace Shao (06:31)Right.John Wang (06:38)really high product thinking, really technical people, and people that could dive deep on certain problems and weren’t afraid to go talk to a bunch of customers. You saw so many conversations about like, how do we make this particular API parameter better for everyone? And like hours and hours and hours of like making sure it was a really, really good product. And people who weren’t afraid to like, you know, take a week of work and just like dump it away because it wasn’t quite there. So it was like,This combination of like, they worked really hard, they’re really smart, and they care a lot about the end result and have a high quality bar. That was Stripe’s version of kind of like talent density. But I think like, you know, if you look at the labs, if you look at different research institutions, maybe it’s just, you know, I don’t know, the raw ability. Yeah. But.Grace Shao (07:26)research capabilities or whatnot, right? No, that makes a lot of sense. Yeah, I wanted to ask you earlier on in our conversation, you said, you know, look, a lot of people overlook support. It’s not that glamorous. People kind of think it’s like a back office thing. But, you know, is that was that your view back then? How does you kind of, I guess, lean into this? And did your perspective or support change over the years? Now you say it’s very important, right? Did you understand the category correctly? Do you think?John Wang (07:52)You know, I think that when we looked at the category, we went at it from like kind of the lens of, Stripe was this company that worked in this unsexy space and did really, really great things. And we thought very similar things about support. It took us a long time to really grock support. And we talked to hundreds and hundreds of different people across different parts of the support stack. AndI think early on, honestly, it was good and bad in certain ways. It was like, we thought we could build a piece of software really quickly that solved everything. Or like, you have that problem, we can build that in two weeks. Not a big deal. And we could solve the specific problems that they had in two weeks. And I remember talking to actually a few people, which was like, the system that you’re trying to do, which is called Workforce Management for Support.that’ll take you seven years. And we’re like, no way. Like we can do this. We can do this so fast. It’s going to be done soon. And now like seven and eight years later, we’re still working on it. We’re still uncovering more and more things. And that was probably the right, you know, that was probably the right call. But also there’s like some importance to naivety, which is like, if we had known that we wouldn’t have started. like we, yeah, like.Grace Shao (09:06)That’s why a lot of people say, yeah, as founders, right?John Wang (09:09)Yeah, so I think it was the right thing to do, which is just like start building stuff.Grace Shao (09:14)It’s amazing. ⁓ Why don’t we pivot into actually understanding your product bit better? So for someone who has never worked in support ops, what is the simplest way to explain to them what Assemble does? Because even between us, we had calls, we had back and forth emails. I was like, John, I don’t understand what you guys do. I’m trying to read through this material. I’ve listened to a few in the interviews. I don’t know what’s happening. Can you just dumb it down for me and explain to me what exactly you guys do?John Wang (09:37)Yeah, for sure. Let’s say you have like 10 people on your support team and you only do email, then you probably just staff them nine to five, right? Like there’s no big deal there. Once you start having a few more people on your support team, let’s say you have hundred people now and you might want to chat to your customers because AI chats, AI chat bots are a really big thing. Then you actually need to start thinking aboutwhen do these chats come in and how many people do I have in order to handle those chats, right? Because like, if you were talking to a chatbot, you’re getting instant responses back and forth, back and forth. And then you’re like, I have to wait 48 hours for the human response after I get handed off. That’s a really bad experience. So the problem is, you you’ve got a bunch of people who are calling in to support, writing in, who are chatting in.and they’re coming in at all different times of the day, they’re calling in for different types of problems, right? You, on the kind of like back end, you have a bunch of people and those people might be able to do different types of things. Like I might be a really good person to handle, you know, where’s my money kind of issues, but I might not be as good at like ⁓ fraud issues, right? Like if you’re having problems with fraud on your account.So there’s a lot of ways in which you can actually put people to the actual incoming tickets. And what our platform does is it tries to match those two things up. if you think about supply and demand, supply is the people that you have and demand is the people, like your customer is asking for questions. And if you don’t match those up well, you’re gonna either...spend way more money than you need to because you’re just going to staff everything way above what you need, or you’re going to have a terrible customer experience because it’s going to take you a really, really long time to get back to people. So it’s really an efficiency play. How do we make it really, really efficient for you to answer questions? In the last few years, we’ve also added AI agents, which is, you know, how do you actually respond instead of just with people, but also with AI togo and answer a chat or answer a phone call directly using AI.Grace Shao (11:50)That’s amazing. I really didn’t know there was so much like science kind of going behind that. I just thought kind of like you’re on a chatbot usually you have to have your frustrating like get me someone, get me someone. I’m one of those people who like no pages pressed zero all the time. I’m like, get me a human. But it makes sense. actually once you can match the talent with like the issue, it can be a lot more efficient in solving the issue and the customer experience will be much better as well. On the AI agent side.What’s the kind of, I guess, consensus right now? Like, are they really actually good at solving issues? Are customers complaining about them? Like, ⁓ how sophisticated are they at this point? He’s like, in my day to day, you know, obviously calling the banks or DHL for pickup or package returns, whatnot. None of those agents are really a pleasant experience, frankly.John Wang (12:35)Yeah, I think this depends pretty drastically on what tools you give these agents access to. I would say that the standard experience right now is fine. It will answer knowledge questions for you. And these can solve anywhere from 30 to 60 % of incoming issues, depending on how many knowledge questions you get.the place where it really is important is when you actually give it access to say your backend database and you can like make a refund or you can look up in order or you can identify why is my what’s going on with this error, right? And that is actually the hard part that prevents most of these banks and airlines and etc agents from being very good is because like thatAccess to data is a thing that they need to actually run and be able to perform actions. And then also the evals for that are really, really hard and not something that you just like launch without really thinking about it. So I’d say it’s in a progressive, like it’s in a progressing state, not at a place where it’s like, this is absolutely solved, but there are also some of our customers who have 90, 95 % of all issues who are able to be completely automated.because they’ve spent the time to give access to all of these systems and spent the time to validate that the agents are performing.Grace Shao (14:00)Very interesting. ⁓ I want to pivot into the to be kind of angle. Who are you guys actually selling to? Like who are the people inside companies that are managing this? Is it the head of support operations? And when they are buying and assessing your product like yours, is it really winning on price? Is it like over, you know, other maybe large softwares? Is it winning on speed and service? Like help us understand essentially how you guys are succeeding winning over customers.John Wang (14:26)Yeah, we generally sell into the head of support. Sometimes that person rolls up into the COO or there’s a head of operations or something like that. But generally there’s some group that is working on unsupport related things and that’s who we sell into usually. I think generally our differentiated, like the way that we actually go and sell this is one, we know all about workforce management, which is like a really, really nitty gritty detail about how youmake your systems really good. And it can save you millions and millions of dollars. Almost actually, and this is one of the things that’s really funny, it’s like using our AI agents versus using our workforce management, we actually see somewhat similar gains across those two. Because to use the AI agents, you’re usually doing it so that you can reduce head count, right?And in order to reduce headcount, you need to know how much can I reduce headcount without hurting my customer experience. And for that, you generally need something like Workforce Management. So what we do is we go in, usually we have Workforce Management helping you understand how your system is set up. And then what the AI agents that we can also bring in is a relatively easy sell becauseour AI agents are really, really connected to kind of like how you staff and when you pull in people. The thing that you were mentioning, which is like, hey, my bank still doesn’t have a very good experience, that’s true of a lot of places. And getting access to information is really hard. So escalating to a human actually happens pretty frequently, sometimes 20, 30, 40 % of the time. So getting to the right human or the rightor figuring out when to escalate to the right human is a really, really important skill to have. If you spend a million dollars a year with me, I should escalate you much more quickly than if you are a free user and you haven’t spent any money for me ever. Similarly, depending on the topic, depending on what kinds of things you’ve already previously talked to me about, I should be able to get to different types of agents and I should be able to have different levels of thresholds.that send me to a human. And I think because we have and handle the workforce management side, our ability to do the handle time and to make sure that you’re getting to the right person is much, better than a lot of our competitors.Grace Shao (16:45)So there’s an unspoken tier system then I guess with customer service as well that we don’t realize. How should we think about the economic importance of support operations? In terms of, we always think of it as like we said, back office support, but how much, do you have any proof that like basically better customer service equals better revenue?John Wang (16:52)There is, and sometimes it’s spoken. But yeah.You know, that’s a good question. should probably have some specific proof here. I guess the best anecdotes I can find are usually the kind of like medium to long-term anecdotes where companies that do not invest in their customer support tend to, you know, regress to the meat, right? Like if you are really trying to bare bones your way through customer support,⁓ Your customers will understand that and it’s not going to affect your revenue right now, but it will likely affect your revenue in 6, 12, 18 months the next time that purchase happens. have seen actually some of our customers, so in our AI agents, we have a configurable setting that’s like, do you want to be containment focused or do you want to be escalation focused? And how good of a customer experience do you want?And we’ve generally seen actually that there’s a strong correlation. Obviously we haven’t run a ⁓ natural experiment or a true A-B test with this because it’s pretty impossible. But you see a general correlation between the customers that spend more money on support, the customers that spend more on trying to have a high quality experience, and the revenue growth of those companies.actually most of the customers that we spend a lot of money on care so much about support that they actually have, you know, executive briefings every week about these, about what’s going on. And they’re the people who have the largest support teams and they’re the people who kind of like make the most, make the most changes with their team. Obviously this is a very biased perspective from our customer, like set of customers, but I think that there’s still something to that where if you spend money and if you want to make ayour support really, really good, that does tend to pay off with customers because they do tend to notice and it makes it easier from a product perspective to paper over all of the things that aren’t so great.Grace Shao (19:06)Yeah, no, totally makes sense. think even as consumers ourselves, we would be likely turned off by certain brands or experiences if the customer service really bad, right? Unless, like you said, they’re a monopoly and there’s nowhere you can go. All right, let’s talk about AI. You kind of touched on that earlier, but the naive view is that AI automate support, you know, a lot, a lot of the conversations right now about, my God, jobs are going be taken, especially the first batch is probably in roles like operationals and customer support roles.Second batch, people are saying are maybe in like more repetitive execution roles like junior consulting roles, a lot of junior training up roles, right? How do you see that? Because at least from where I sit in Hong Kong, a lot of stories are coming out saying markets like India, the Philippines, know, across Southeast Asia where they traditionally served as those telephone call centers or operational centers, they are getting caught. Is that going to be a trend forward?you know, how should we understand this?John Wang (20:02)Yeah. Yeah. I think there’s a few things. There’s like, like with all things, there’s a lot of nuance to this, which is I think your trend on seeing, you know, what we call tier one support, the first line of support who are traditionally humans outsourced. That is a place where we’re seeing a lot of change. And I don’t think that trend is going to slow down. That said, there’s a very interesting other trend thatwe’re seeing, which is that total spend on humans and headcount isn’t necessarily going down by that much. And it’s kind of like Jevin’s paradox where we see a lot of our customers and a lot of customers of other AI users ⁓ who have amazing resolution rates. They’re like answering so many questions, but that’s causing actually, or maybe there’s some correlation here ofthe number of tickets they’re getting and the number of chats they’re getting is like way, way, way higher than before. And I think there’s a few parts to this. One is you see way more ability for your AI agents to answer questions. And so obviously people are going to ask more questions because like, Hey, it used to be really hard for me because I had to literally type out an email to a human, wait a few days and get an answer. And now I can just like get an instant really good answer. Right. So I’m going to try asking more questions.The second thing is, as these companies do better and better, you actually just have this natural induced demand of increasing usage, numbers of people who are asking for support. So the higher amount of support that is automated is also, the general number of how much support is coming in is also very high.And so that actually offsets a very, very large portion of the head count. The head count is changing though. It’s not going to be the typical tier one support where it’s just like, answer an easy question. That is mostly going to go away to AI, think. The types of head count that is coming in are like, know, internal agents, people who are really good, people who can provide white glove support and like...actually go talk to people and provide like a human experience because like our companies still want and really crave giving that experience to people. And that’s just not what the kind of BPO standard really is. So I think it’s changing in the type of what you would see.Grace Shao (22:31)Yeah, I was actually going to ask about that, like, as in when AI agents start resolving more tickets, if we’re just going to see reduction of headcount. And I think you answered him when you once wait, whereas like, yes, in initial stages, but later on, there will be new jobs created, right? Essentially, people who will be managing more critical issues or even managing the agents. I want to understand. So for your company right now, essentially, are you a are you like a middleman between the human agents and the AI agents and becoming the orchestration layer, like you’re providing the service, the training and the orchestration. Like, how do we understand that?John Wang (23:05)Yeah, that’s a great question. So we think of ourselves as how do we get you to the right way to answer your question, right? In our view, there’s kind of three main types of people that can answer a support question. One, it’s AI. Two, it’s a tier one BPO’d outsourced agent. And three, it’s an internal agent who’s like super well-trained and like super, like, you really carrying about, like really trained on customer support. And what we are trying to do is make sure that you get placed at the right area, depending on what kind of issue you have and who you’re talking to and like what is the kind of like a cue that is backing up the set of people who need calls. So for us, what we’re trying to do is really provides you that ability to choose across a bunch of different options. So we don’t actually provide any, like we don’t provide any BPO agents, we don’t provide any internal agents. All we do is provide the software that routes you. And we also provide the software that can do the AI agents, or you can actually plug into a different piece of software if you want to have your own AI agents too.We’re trying to make sure that we are kind of third party and that we are making it really easy for you to optimize your support regardless of what specific providers you use.Grace Shao (24:30)So in your view, what does a well-run support organization look like in, let’s say, three years as AI adoption becomes mainstream or more more mass market?John Wang (24:38)I think you’ll probably want to have all of the different types of support using AI. So voice AI, chat AI, email AI. I think you’ll want to have a lot of nuance between the different types of customers that you have. You can’t generally provide the best level of support for literally everyone. Though this depends on also your customer base, right? Like a consumer customer base versus a super enterprise customer base with 100 very large customers is completely different. But let’s say for a standard company that might have ACVs that are in the, I don’t know, the 100 to couple tens of thousands range, then you’re probably going to have a combination of AI agents and human support. And you might have different tiers of human support, right? Some human support that’s really good at answering support questions and other tiers of human support, which is like, you’re just managing the agent. I think the other thing that’ll happen a lot is you’re gonna start to see more like, supporting agents acting in a simulation where right now, like the kind of typical flow is like a supporting agent gets a ticket and they answer it and it goes back. I think as the agents get more like, get more and more training data, get access to more information, really they’re only gonna come to humans for escalations. And similar to how Waymo works, if you’ve ever taken a Waymo, it’s a great experience, you’re like driving, driving, driving, and sometimes you get kicked out and a human operator in the Philippines is like, hey, I need to move you around this truck, right? And similar to support, That’s probably what’s going to happen. A human operator is going to come in and be like, hey, I can give you a refund right here. And then what’s going to happen is the AI agents are going to train on that, right? They’re going to like learn and get better. And you’re going to be able to use that whenever you have an interruption to understand like, why did I have this interruption? How do I make my model better for the future? And then you’ve got your closed loop. So I think in the future, you’re going to see much more of that happening than people who are just like, coming in and their job is to solve as many tickets as possible. I think the change is gonna be like, okay, people are gonna start to need to provide the best possible response in that particular instance so that the models can train on that and be as good as you are.Grace Shao (26:58)actually just on that, do you think then we’ll see more and more fraudulent activity or people trying to exploit that? like if say you know the models trained on, I say this one buzzword or one keyword and it triggers like refund. What if I just go on the call like on the phone all the time, just to be like keyword, keyword, you know, and then like how do we prevent something like that? Or do you guys kind of get involved in that building this guardrails as well?John Wang (27:21)Yeah, no, that is a age old question. think like, wouldn’t say there is going to be necessarily more or less of that, but I think like, it’s kind of like the cat and mouse game of like, everyone has always been doing that. And so like, and the methods always change every, every few months. I think the methods will change every few months here too. Our AI agents have a lot of guardrails put in place to automatically detect that. And we also have kind of like post-hoc guardrails which are like scanning through our logs and trying to identify situations where that might have happened. And we’re also training on those examples, right? So I think, yes, people will definitely start to exploit this and be like, hey, how do I get a refund faster? But there’s a ton of guardrails that you can put in place. For example, each account, you can have one or two, have like refunds without looking until that actually gets flagged and it needs to go to a human or.You can set good policies, for example, like, you know, if it is within policy of 30 to 40 days after purchase, like automatic refund, otherwise, you know, flag it and do something with it. So there’s a ton of stuff that you can do to actually like reduce the possibility of that. And I do think that it will end up being cat and mouse game like over and over again, as people get more sophisticated.Grace Shao (28:39)Right, right. And they’ll start using AI to trick AI. That’s what’s scary, right? So as we talked about, different gender standing the podcast does not have to interview anyone related to China or Asia, but we do have kind of an Asia angle to a lot of how we view the world. So my question for you really is because you’re like out in San Fran and like your company actually has no sales in China or anything. But I actually had a curious question. How does SFJohn Wang (28:43)Totally, yes.Grace Shao (29:05)as a whole, the startup ecosystem kind of view the current rise of a lot of Chinese AI. And have you guys yourself or your peers, you know, tried to use Chinese open source models over the years? Is there any view on the open source models given that, you know, you previously said you were very involved in open source and I think it’s part of your philosophical belief as well, right? So just kind of like the high level vibes.John Wang (29:28)Yeah, our vibes might be different than at the model, like the Frontier Labs, honestly. Our vibes, we love the Chinese open source models because it adds more competition. And I think the open source models are actually very, very good. I think from my friends at OpenAI Anthropic, they don’t like it quite as much because it’s competition. But for us, we have no allegiance really to any of the Frontier Labs.or any of the models that are out there, we want to provide the best possible experience to our customers at the best possible price. And that has meant, you know, over the years, like making changes in our models, making updates and to figure out what is that frontier of cost or performance. The Chinese models tend to perform really, really well on that, especiallyGrace Shao (30:12)Mm-hmm.John Wang (30:19)kind of like the latest series of models, we’ve actually spent a lot of time in the last six months kind of like pulling out a lot of our tokens. We have tens of billions of tokens per day. And a lot of it now goes to models like Quen or Kimi. And like that has actually started to really, really increase over time, mostly because you can find to them, you can do RL on them, you can...have better latency on them, you can run them on your own hardware. There’s just like so much more stuff that you can do with it. And also, you know, the cost performance latency trade off is really, really good. Now, most of our most of the like the strategy we take is actually one where we try to understand the use case and the problem and what type of model is necessary for that. So for kind of like the main model that’s actually answering questions. We’re actually usually using a frontier model for that. But actually the majority of our tokens come from out of secondary processes, processes like detecting if I need to escalate, detecting if there’s a fraud here, detecting if there’s an adversarial intent, making updates to large swathes of data in batch, like all this other stuff where you really don’t need frontier level intelligence and where if you have a a well-tuned prompt and an open source model or an open source model plus a fine-tuned model, you can get at or better in terms of frontier performance. We’ve really seen that and we’ve actually been able to save millions on our token costs in just the last two or three months by being very smart about how we use our models. And we’ve also seen a 15 to 20 % increase in quality.⁓ Just because like when you go and you have evals, you can make things much, much better more quickly with these open source models.Grace Shao (32:14)Yeah, I think that’s like the general sense I kind of get from a lot of startups, right? In a known day, it’s like, you guys are obviously more cost conscious. What is the best price to get to what you need? And there’s like a tier system where how you use the models, you might not use the most frontier models for everything. I think that makes a lot of sense, business sense, especially. Is there anything you would like to share with us that we haven’t touched on in terms of, just the overall AI ecosystem, any thoughts on, you know, where we’re going with this AI agentic push right now?⁓ you know, are we really going to see that, you a giant moment, like just kind of some high level thoughts.John Wang (32:49)Yeah, that’s a good question. Recently, I’ve been thinking a lot about Opus 4.7, which got launched a few days ago. And it’s actually kind of similar to what we were just talking about in terms of this price for performance ratio. And it seems like, based on my usage, based on our evals, based on other people’s usage on the coding side, that it’s a better model, but it is also more expensive.than before. like, you’re really it’s literally like a trade off in terms of dollars and intelligence. And it’s really interesting because, you know, a year ago, every model would just be like, this is strictly better, and it’s probably cheaper, and you’re to get more context and like, everything’s better. And you could basically just bet that you’re just going to like get better models across the board. And now actually, you’re just like kind of moving from this part of the like the frontier curve to the other part of the frontier curve without actually shifting the entire curve. And that’s happening with a few more model releases. You still see general increases in the frontier, but it’s less stark every single model release that you see that. And so I think it’s just an interesting area to look at because when you get into that world.Gross margins has become really important. Gross margins for ourselves as a startup, but also gross margins for Anthropic and OpenAI. One of the funny things that I’ve seen, just talking to people who are working at Anthropic and OpenAI, and also people who are trying to invest in those companies, gross margins are actually incredibly important. One of the OpenAI right now is becoming a much more...hand investment than before. And like, it used to be like six months ago, it’s like, you have it, you have shares of OpenAI, like, how do I get in? Now it’s completely different with, you have shares of Anthropic, how do I get in? And I think part of that’s because like, OpenAI wants to spend $100 billion on infrastructure. And Anthropic is a lot more measured in the way that they’re spending money. And I think gross margins actually do matter a lot right now. And that’s where I think actuallyChinese open source models are making a big difference because just at the end of the day, you still have to make money. And if you’re losing money on a per token basis, that’s really bad because if you go to infinity, you lose infinity money. And if you make money per token, great. Ramp usage up as high as you can.Grace Shao (35:03)Yeah. It’s just so crazy how the sentiment shifts like so every three months I feel like and then to your point like whenever I speak to investors like oh my god I got my hands on some anthropic shares and last three months earlier. Oh my god I got my hands on opening I like it’s just like and like oh no one would invest in opening right now like I don’t want to do that like people just completely go like black and white on these things it’s pretty crazy how the pendulum swings I do have a question actually on the infrastructure side doesn’t it actually make sense for open AI to eventually own their infrastructure because otherwise they have to becomecontinuously constantly pay the hyperscalers for all the infrastructure like so in the grand like scheme wouldn’t it make sense? I mean although obviously how much you’re spending is like absolutely crazy.John Wang (35:55)I think it actually does. And I think that’s like part of the problem, which is like, you know, if you think about what their compute costs are, I think actually doing all of these things that they were doing makes perfect sense. And it makes especially perfect sense if you have investors who are willing to bankroll this. But it’s almost like the, ⁓ what’s that paradox? It’s like the St. Petersburg paradox, something like that, where it’s like, you keep doing,your expected value is infinity and you keep doubling your money basically, but at some point you need to not double your money because you don’t have enough money.Grace Shao (36:32)That’s such a mo- I’m like, I’m still confused when you’re saying, go back. You keep on doubling your money.John Wang (36:36)Sorry, So I think the I think it’s like Let me let me look this up st. Petersburg paradox is Okay, it’s a coin flipping game and You start at two dollars and with every tails you double the pot and you can basically decide to like take your money at any time, right? and so you you’re doubling exponentially as you go up andIf you compute the expected value, you should basically just like, keep going forever because your expected value is like infinite, right? Like because the doubling of the pot is better than kind of like what your losses are. You just got to, you got to run. And I think OpenAI is in this St. Petersburg paradox where it’s like, well, in theory, double everything, keep going. But in practice, you don’t have enough money.Grace Shao (37:17)Yeah, I see what you mean.John Wang (37:25)and resources to be able to do that. I think that’s actually what’s happening is like, there’s not enough money in the world, not enough investors with liquid cash who are willing to invest in a business as big as OpenAI while the gains and the returns are still, yeah, having improvements. So I think it’s both rational, but also, you know, actually practically very hard to make what they’re doing.Grace Shao (37:41)haven’t been proven. Yeah. totally. Okay, I want to ask you one last question, which I ask every single guest. What is one differentiated view you have? It could be on your own sector, industry, life.John Wang (38:00)man, have a really like, I have one that like is very controversial. I don’t know if I should talk through that one. ⁓Grace Shao (38:07)You’re get doxxed and to hate it after this.Okay, tell me that one after, I wanna hear it.John Wang (38:17)Yeah, yeah, Let’s see. Like... I would say, I don’t know if this is differentiated now in the market or not, but the thing that I’ve been thinking about recently is that you should stay somewhere long enough where you see your mistakes through. And I think it’s like slightly differentiated right now, because like you’ve got in Silicon Valley, at least you’ve got people who are jumping between big labs, who are jumping between different startups where it’s like, Hey, I can make the next, you know, $5 million.by going to this next thing. And there’s just a whole huge amount of opportunity and there’s like a ton of opportunity costs to staying somewhere for a long time. And at the same time, think like long-term staying somewhere for a long time is actually one of the best things that you can do for your own learning. And it gives you that a better shot to make like the long-term massive gains that you could have like $5 million.is amazing for someone. But if you want to build your own startup, if you really want to like change everything, if you jump around between companies every year or two, like you’re probably not actually going to learn a lot. And you’re probably not in the position to make really hard decisions and then have to see those hard decisions through and then, you know, be able to learn and see that feedback from those hard decisions. Especially if you’re jumped likeEspecially if you’re like at OpenAI and you’re like, no, investors don’t want this anymore. You jump ship to Anthropic. That’s like, you know, I don’t think you’re going to get that, the learning that you really need to get.Grace Shao (39:46)Yeah, yeah. actually agree with that. think I also took some time and like experience to realize that because when we’re all young, like you’re really excited, right? It’s like, this looks cool. That looks cool. this person hates me. hate that person. Like you take everything very personally and then, you know, we’ve all heard these stories from peers, even ourselves. But what is the threshold though? Because then the other side of the argument is that like you see people who’ve been in a job for like a decade and clearly they’re frankly not.moving up in a very corporate structure way or even intellectually growing or even, you know, excited about their job anymore. You know, the joke is like you get the like, okay, this sounds on PC, but you know, like the 45 year old VP that’s been a VP for the last 15 years at banks, we have a lot of these. So what happens? Like when is it best for them to actually maybe jump or some say, that was like a lifestyle decision where they want to take it easy because they have some more time for kids. Fine.But taking that kind of considerate way, wouldn’t it sometimes be better that you jump to try something new to take risks?John Wang (40:50)I think if you are in a place where you’re unhappy with, so I will caveat this with, have to, you should only stay if you’re excited about what you’re doing and you’re learning continuously and you’re surrounded by great people. If those three things aren’t true, yeah, it’s really hard to fly.Grace Shao (41:05)Which is so hard to find. you were very lucky at Stripe, right? Like you said, you were just surrounded by very high caliber, high agency people, but not everyone can get all those things at the same time. ⁓ But no, that’s great. Thank you so much, Sean. ⁓ I had a lovely time chatting with you. I still wanna follow up on what was the unspoken differentiative you later. All right, thank you.John Wang (41:17)Yeah. ⁓ Let’s do it. Let’s do it. Thanks, guys. Get full access to AI Proem at aiproem.substack.com/subscribe
• Matt Sheehan on China’s AI Policies: Employment, Anxiety, Safety, and State Priorities 27.04.2026 1h 1min

  Today, I’m joined by Matt Sheehan who writes this insightful newsletter. Matt is a senior fellow in the Asia Program at the Carnegie Endowment for International Peace. He researches China’s AI ecosystem, Chinese tech policy, and how technology shapes the country’s political economy.Matt lived and worked in China from 2010 to 2016 and later led China tech research at the Paulson Institute’s MacroPolo. He’s the author of The Transpacific Experiment. He speaks Mandarin, and he turns complex policy into plain English.In this episode, he helps us understand China’s AI governance, about how Beijing is thinking through the social and political consequences of rapid AI adoption. We focus especially on a shift that became more visible in early 2025: rising concern inside China’s policy community about AI’s impact on jobs, worker anxiety, and social stability.Matt explains why China’s AI labor question is different from the Western debate. We also discuss how the Chinese government is trying to balance support for technological progress with the need to manage public anxiety, clarify labor rules, and avoid social instability as AI becomes more deeply embedded in the economy.He broke down the myths, explained the jargon, and the regulatory bodies in China. Our conversation started slow, but it became very, very heavy, what they call 干货满满 substance heavy. Also, a shoutout to Nathan Lambert’s work in helping us better understand the open-source ecosystem and Rui Ma’s for helping us understand investing in China AI!Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, founders, and product managers.For more information on the podcast series, see here.To find the previous episodes of Differentiated Understanding, see here.Chapters00:00 Introduction to AI Policy in China03:10 Matt Sheehan’s Journey into Chinese Tech Policy05:55 Shifting Perspectives on AI and Labor09:02 Public Concerns Over Job Security and Government Responses15:09 Education and AI: Preparing for the Future17:50 Regulatory Landscape of AI in China34:00 Navigating China’s AI Regulatory Landscape40:58 Misconceptions About Chinese AI and Government Funding43:57 Understanding AI Safety and Security in China52:03 Global AI Governance: Cooperation or Parallel Paths?AI-generated transcript Grace Shao (00:01)Hi Matt, thank you so much for joining us today. I’m so, so happy to finally have you on the pod for people who are listening. We’ve been trying to make this happen for like six months, but between us, there are like three little children running around with a bunch of viruses and have just not been able to make this happen. I’m really excited. ⁓ A few months ago, what really caught my attention about your work again is that you shared something on WeChat saying you were dissecting the new Chinese AI safety paper, like the big national one. ⁓like verbatim in Chinese. And I was like, wow, this is extremely impressive. It’s not an easy task. I commend you for doing that. So I really wanted you to help us understand the nuances of the AI policy world, especially how people are perceiving AI in China. I think there’s more more interest in how China’s governing AI ⁓ while we were hearing the backdrop of how the Chinese government is trying to push on AI diffusion, right? And then on top of all of this, like where areas where China’s AI governance seem to be leading, because in many ways it seems likeChina’s AI regulators are much faster to respond to how fast technology is evolving. But to start, we would love to hear about your personal story. Tell us about how you ended up studying China, studying Chinese tech policy. We met in Beijing years ago, maybe a decade ago. ⁓ Yeah, so tell us about that.Matt Sheehan (01:11)Sure.Yeah.Yeah, sure. Sort of stumbled into China stuff. I hadn’t taken Chinese or really knew anything about China until about halfway through college when I ended up getting a summer job in Beijing. I was just kind of like instantly fascinated and knew I wanted to move back there after I graduated. So took a little bit of Chinese my senior year, moved to Xi’an, taught English, kind of followed what at the time was a very like typicalknow, trajectory of like, go there, teach English and then go study Chinese at university and then get a job and get a slightly better job. And eventually I was able to kind of wiggle my way into journalism. And so I was a China correspondent for a publication called The World Post at the time. And that took me up. was there from 2010 to 2016. So kind of like the hinge period before and after she came to power. Pretty interesting thing to see. AndWhen I moved back to California in 2016, I started working on a book about China-California ties. I’m from California and this was like the period of kind of explosion in cross-border investment and Chinese students come into California in the Silicon Valley-China relationship getting even more like twisted and complicated. China-Hollywood. So I wrote a book about that and as I was doing it, the kind of the tech section, the China-Silicon Valley, China-U.S. tech connections kept growing bigger and bigger and I ended upworking a little bit with Kai-Fu Lee on his book, AI Superpowers, which was kind of my turn from like, it was like all things China, China, California, China, Silicon Valley, China AI. And since 2017, I’ve been working almost exclusively on AI issues in China. Maybe the first three years of that, like 2017 to 2020, was very focused on comparative capabilities. This was kind of a period right after the National AI Plan in China when there’s a big explosion in activity. And I think...This is kind of was like the first time America kind of got freaked out about Chinese AI capabilities. And so I spent a few years being like, okay, let’s try to like ground these assessments in some data. Let’s get like an actual grounded sense of where the countries are with each other. ⁓ And then starting in 2021, I sort of turned into focusing on Chinese AI governance, Chinese AI regulations. That’s when they first started rolling out their regulations or recommendation algorithms and sort of deep fakes. And I was kind of making a bet that I thinkIf China continues to be at or near the frontier of AI, then how they choose to regulate it domestically is going to have huge implications for China’s own ecosystem. And then it’s going to really ripple out internationally on safety, security, growth, all this stuff. So I spent the last, now it’s like five years, ⁓ just deep in the weeds of Chinese AI policy and regulation.Grace Shao (04:04)Oh, great. I think I definitely want to double click on all the algorithm security and the kind of, you coined the answer versus what’s called security and versus what’s the other Chinese word? Yes, yes.Matt Sheehan (04:17)Anshun safety security. ⁓Jeff Ding was talking about this very early on, but yeah, it’s a, it’s a constant thing that we have to negotiate for people who don’t know it’s the word Chinese, the Chinese word Anshun ⁓ means both safety and security. whenever you’re kind of translating documents on this front, you have to know, you talking about AI safety, which is kind of a different thing versus AI security, right. I think both you and Jeff definitely are some of the more nuanced scholars I follow. And I do want to kind of double click on that later on. But to start, I think I want to talk about something that’s top of mind for a lot of people. You just wrote a piece that you said was not super serious. It was just your scattered thinking put together on subset. I thought it was very well written about the growing anxiety around potential job losses. ⁓your perspective is that you know there are more and more people voicing this kind of concern and I wanted to hear a perspective on that and I kind of wanted to share a bit of my different share my different perspective on this and what I’m hearing on the ground and kind of have a conversation around that as well. Yeah why don’t you start with sharing like what you found yeahMatt Sheehan (05:22)Yeah, sounds great. Yeah. So, I’m not an AI and labor person. That’s not been my focus for a long time, but I’ve been monitoring it for a long time and just lightly. And starting around, I guess it was early 2025, I just started to hear a lot more out of the Chinese policy community about worries about AI’s impact on labor and jobs. And this was kind of a surprise to me because ⁓ just a little bit prior to this, say early 2024,I had, I sometimes ⁓ in my job, I run these kind of like informal surveys or almost like a, what do call it, focus group of American and Chinese AI policy people and asking them like, you how would you rank these different risks? How concerned are you about ⁓ job risks versus privacy versus military AI? And we have both sides that like rank the risks and then talk about the results. And when I ran one of these in early 2024, it was very striking that the Chinese sideI think it was at the time we had seven different risks and the Chinese side ranked labor impacts a second to last as six out of seven. And so my sort of baseline was like, OK, for a variety of reasons, this isn’t really too on the radar of China’s ⁓ policy community or wider policy community. And then starting around early 2025, some of those same people who I had been talking to about this before had really changed their thinking. They were saying that there was a big change in thinking withinChina, maybe especially within China’s of policy and government circles, but then I think also a little bit wider. And so that sort of sparked my curiosity. And for the past, now it’s like over a year, I’ve been just sort of tracking when does this AI and labor question show up in state media? When does it show up in kind of the online discourse? When does it show up in policy documents? And sort of the TLDR is like, I think this is really, really ramped up a lot. Over the past 18 months. It’s been maybe the single biggest change in how China perceives different sort of risks as it relates to AI. And I think I’m looking forward to sort of discussing how maybe like the policy world or the government’s perception of this differs from ordinary people or certain categories of people. ⁓ But I think it, from my perspective, it’s sort of it’s infused into both. think there’s been a fair amount of public concern.the policy community picks up on that and they want to both respond to like the actual problem, know, actual job losses, but they also really want to respond to people worrying about job losses. That’s kind of maybe the thing that actually made me write this piece now was discovering an interesting piece in state media. I think it was in science and technology daily. That was the headline was like ⁓ AI must be controllable, but people’s ⁓ anxiety about AI must also be controlled.And it was all about how sort of OpenClaw has triggered a lot of anxiety and a lot of people about, are they going to get replaced? You need to be building your own AI agent in order to not be left behind. And they’re sort of trying to tamp down those concerns in a few ways. So that’s what sparked the piece itself.Grace Shao (08:33)Yeah, I think definitely what you saw and you wrote about is like definitely kind of playing out in the China AI policy ecosystem that I see as well. And I think for sure, the open-claw frenzy have kind of opened the eyes to lot of the even average people what AI could potentially do. However, I guess my argument, not against it, but it’s just like, you know, we kind of cite each other’s work on subset. But my point was kind of saying, you know, this is a reflection of a relatively elite group of people end of the day, because the knowledge work economy in China end of the day is only like only 30 % of the workforce actually are the knowledge working economy. And end of the day, even though it’s 30%, because China has such a huge population, the mass, the sheer scale, it feels really large. However, I want to bring it back to the idea that like anyone who’s lived in China understands that the government’s like top top priority really is about social stability which leads to what they call social harmony, right? And I just think that, you know, the rising anxiety of job control a lot of times maybe is because there’s a fear of if there’s a lot of disruption to jobs then people will lead to social unrest which obviously gets a bit more sensitive but you know, a lot of what they do comes from that I guess thinking so I agree with you top down definitely have to understand what’s happening with technology and how advanced AI has become in 18 months.⁓ have given them kind of, I guess even fear mongered a little bit internally, right? ⁓ But the nuance here is that I think ⁓ the rest of the 70 % of the Chinese workforce actually don’t work in anything structured that we know. I think even probably even 80%, you know, people in China, they most of them are actually like, you know, service providers ⁓ from, you know, rural areas and urban areas. A lot of people work in factories, even the entrepreneurs, right? They run like say hospitals, clinics, factories, bottle cleaning, like factories, whatever, right? ⁓ Car logistic rental businesses, these people aren’t actually ⁓ trained in the way that maybe the West by default think they are. They actually just run it from a grassroots way and they don’t have very, very streamlined processes. They don’t have documentation. They don’t run like what we think a corporate has run. So in that sense, I think it’s very hard for AI to replace any of their workflow.because it’s actually not a, we can’t really provide context and a lot of things, business is done is through one C, is through a wink, through a look, through a gesture, through, you know. So a lot of that, I think, in fact, will be harder to replace than even maybe some of the more mature businesses in the West where there are structured processes and everything. So that’s kind of my, I guess, a more nuance, I think, push on that. Yeah, wonder what you think of it.Matt Sheehan (11:20)Yeah. Lots of thoughts. And I think the sort of the fundamental distinction that you’re pointing at is very valid in that like, you know, companies in the West, in the United States, they’ve been like big companies have been running sort of digital databases for decades. They have like decades worth of data. They have pretty advanced like enterprise software. It’s just a much more I want to say something like a bias, but it’s a little more like put together sort of official structured⁓ technological backend and not just technological, but like a process backend. Whereas in China, it’s just things have developed really quickly. A lot of it’s on the fly. lot of it is, know, enterprise software is just not, there’s not really a market for that in China in the same way. It’s a lot of stuff is pirated or they’re just not digitized in the same way. And it’s actually very interesting. This is in kind of the early days of the like China, US who’s ahead. ⁓ know, a lot of the debate focused on data.Matt Sheehan (12:18)And there was this idea of China has a billion people, so it must be this huge advantage in data. But my pushback on that was always kind of what you’re arguing, is like the US actually has very structured data, and it’s owned by corporations, it’s deployed by them, they’re already doing type of sort of lower end market intelligence type stuff. ⁓ So I think that that sort of backdrop is very real. think ⁓ maybe from there I’d like differentiate out to potential risks or debates. And it’s kind of what you were pointing out as well. There’s like the actual question of how many jobs are going to be impacted. How many people are, what is it going to do to people’s wages? What’s it going to do to aggregate employment? And then there’s the question of like, ⁓ how do people think about that? What are the fears due to sort of the Chinese social stability, even if the things haven’t manifested. So I think separating those out, definitely the government is... ⁓their sort of initial response is related to public worry about this. So in the piece, I detailed the way that ⁓ sort of a robo taxi incident in Wuhan was in many ways the spark that really like ramped up the government thinking on this. And this is something that I heard from a couple of different Chinese policy people who both pointed to this incident, which I had totally missed at the time and wasn’t like major international news, but that had a big impact. And basically what it was is thatBaidu was rolling out its sort of fully autonomous robot taxis throughout the city of Wuhan. There was this kind of like public letter, ⁓ open letter released by a taxi company that was kind of railing against, you know, both ride hailing platforms and autonomous vehicles as, you know, stealing the iron rice bowl or just smashing the rice bowl of taxi drivers and of companies. And even though it was a kind of like a small thing, a couple of days later, a Baidu taxi actually hit pedestrian, I don’t think they were seriously injured, but it kind of fed into this overall, a big kind of online reaction and discussion about like, what’s going on with AI? Is it going to take people’s jobs? And it, it’s one of those things, it’s funny to explain to people because it sounds like nothing, but it did lead to a pretty significant ⁓ imprint on the way that the Chinese government is thinking about it. So it was coming from, in many ways, public discussion of it. Like the discussion was happening online. This discussion might be happening among, you know, elites.chronically online people. ⁓ But it’s something that the government definitely picks up on. So I that’s one element. They’re worried about the worries and they want to, like among their sort of policy reactions in a ways, thus recently has been sort of directing platforms to say like, you kind of need to tamp down these articles or these viral videos that are telling everybody, like if you don’t adopt open claw, you’re going to be left behind. There’s been this kind of rash, both in China and here in the US of like,you know, kind of like fear mongering people into clicking and taking your course on building agents or just subscribing, whatever. And one thing the government is doing is telling you like, chill on that. Like, don’t be putting that narrative out there. ⁓ And so that’s part of this kind of like public opinion management thing. In terms of the actual impact on jobs and who will it hit?It’s a huge open question that I personally have gone back and forth on for years. I first kind of did a deep dive on this way back in 2017 when everything was still so speculative. At the time, was pretty not worried in part because of the reason you described it. I’m like, there’s just a lot of friction. There’s just so much friction in this economy. And just because an AI system can theoretically execute a test doesn’t mean it’s taking a person’s job. And for me personally, my thinking on this has changed a lot in the lastyear to 18 months, mostly because of how capable agents have proven to be. I expected agents to essentially be hitting a lot more roadblocks while they’re being deployed online. They haven’t been. They’ve been operating much smoother, or just they’re more relentless, and they can break through these bottlenecks. ⁓ It’s interesting that in China, the inciting incident was not about white-collar workers. It was about taxi drivers. ⁓I don’t know this as a fact, but if I had to guess, I would guess that a much larger portion of the Chinese population’s job is driving a car or driving a scooter or something like that. And that maybe that’s a vulnerability that they might face depending on how self-driving vehicles roll out or delivery robots and stuff like that. The knowledge workers, yeah, it’s such a messy and unclear thing, but I think the government has at least started to take it seriously because it’s not, their policy responses are not just this like,public opinion management stuff. They’re also talking about, ⁓ like one of the more interesting pieces that I highlighted is, and maybe the most concrete thing they’ve done so far is, ⁓ according to Chinese labor law, there’s sort of reasons why you can and cannot fire a person. There’s like legitimate and illegal reasons to fire someone. And when someone is fired and they object, this gets taken to like a labor law mediation ⁓ body that’s under the Ministry of Human Resources and...forget what the second part social ⁓ security. ⁓ Yeah, Ren Li, Ziyuan, Shouhui Bao Zhang. Yeah, that’s what it is. ⁓ And in the last year, one of the things that really made a lot of made kind of a big splash is that those mediation bodies declared and it was echoed in like the biggest state media that saying that you replace someone with AI that AI can now do this person’s job is not a legitimate reason to fire someone and those people have to be reinstituted into their jobs.That’s a concrete policy thing that’s actually directed quite clearly at like actual impacts. ⁓ Is it going to work? I really don’t know. It might just be a little bit of friction and, you know, maybe China’s kind of doing what it always does, which is like, we’ll figure this out. We’ll kind of muddle through this. We’ll put some friction here. We’ll grow a little more here. But I think the concerns are real, whether they bear out, ⁓ whether they hit faster in the United States or China, which country is better positioned to sort of roll out a more redistributive welfare system. think these are all open questions, but I think the concerns at least are real.Grace Shao (18:24)Yeah, and I think you hit something that I feel like it’s being kind of missed in the headlines, which is the government actually cares more about the general mass, which are the people who are driving the scooters and the like the DDS cars more than the knowledge workers, which is kind of different from the Western kind of conversation right now, where a lot of the whether it’s fact, frankly, the power that can lobby and the power that the voice that have the voice are all really concentrated in.the white collar elite jobs that are very much concentrated in Silicon Valley and whatnot, right? And I think it really reminds me of the time when, you know, during the Hulianmang Shidae, like the internet era, you know, like the big tech only really got clamped down when the average consumers felt like they were really being pushed to R-Shrine Egypt 2-1. So that’s when they had to add the monopoly to probes. And then soon after, only maybe two years after the probes happened, there was a common prosperity rollout, which basically all the big tech in some capacity had to like showcase that they had a CSR aspect to them. I think this is something that we don’t really see in the US as much with all the big techs, because it’s kind of like they’re doing what they need to do. They have their profit driven interests. And then of course, everyone has a CSR, but it’s not really allegiance to the government CSR mission. It’s more like, we believe in ESG. We believe in climate. Amazon is going to have some like carbon footprint reduction plan, right? Whereas like the common prosperity thing rolled out. ⁓you know, it kind of died on its own, like no one really talks about it anymore. However, during that phase, when it did get rolled out, it was like an understanding where, OK, if the government wants the, frankly, the poor or the middle lower class to feel protected, then you as a very large ⁓ moneymaker in the economy need to showcase that you are somehow ⁓ part of this kind of support. So I wonder how this will play out for the big tech in China when, the job protection policies really get rolled out in practice, like what you’ve mentioned. And obviously they can’t really say, you’re being replaced by AI. At least there’s that superficial guardrail there, I think.Matt Sheehan (20:24)Yeah, I think the sort of the political economy of these questions is going to be super interesting in both countries. know, essentially like business in many ways, it inverts the sort of technological impacts on employment that have been around for so long. Normally, like greater technologies integrated into the workforce, it hits sort of people working maybe low end manufacturing jobs or jobs that would be considered sort of repetitive and, you know, quote unquote, low skilled jobs, even if they’re not. ⁓And, you know, in the United States, we’ve seen like three, four decades of this. And the people who are concerned about that basically didn’t get hurt because they are not, like you say, part of these influential classes. AI is going to, yeah, in the United States is going to be very different. This is going to be the first time that you have the way that my sort of mental model for it is like, if you’re a senator and you have kids or nieces and nephews or your friends, kids like what, what are their problems?And like how close do those feel to you? And you know, if you’re a 60 year old senator and you’ve got like a 23 year old niece and she just graduated from college and she got a degree in, you know, something that’s like is legitimately employable normally like in marketing or something like that, and those jobs just aren’t there, that’s just going to feel very close to home for people in power in the U.S. in the ways that it hasn’t felt in past waves of technology impacts. In China, I...partially agree with what you’re saying, but I think there is also going to be a significant element of this sort of the same dynamic as the United States. mean, yes, common, you know, Xi, common prosperity. He’s focused a lot on sort of eliminating extreme poverty and, you know, the CCP, it’s in its bones that like the rural, the working class are in many ways kind of the long term support base of them. But I mean, also if you look back at Chinese history, like a lot of the biggest and for the government most dangerous protest movements came out of elite schools, came out of students at elite schools who ⁓ either couldn’t find jobs or were facing inflation issues or had, for ⁓ ideological reasons. I think that stuff does hit close to home. think some of those same dynamics, if you’re a deputy director at the National Development and Reform Commission.your family, the people that are close to you, are going to be the type of knowledge workers that are going to be impacted by this. And I think that just can’t help but kind of like compress in on the thinking on this. ⁓ You know, how AIs can... Yeah, yeah, and like how...Grace Shao (22:58)Yeah.everything becomes personal in the end. Like in the end, it’s likepolitics is still personal. Yeah. Sorry.Matt Sheehan (23:08)Politicsis personal and yeah, mean, like cities are where social instability is the most dangerous. Like cities are where people gather and you can have potentially dangerous incidents. These are the people who are very online and are sort of sparking or leading the conversation as much as that can be controlled and manipulated via censorship regimes or public opinion guidance. Like these people are gonna be vocal. yeah, I think if I was at the CSPI, I’d be concerned aboutGrace Shao (23:39)I want to kind of go into on education. Like you kind of touched on it, right? Like, you know, there’s been draft rules about children’s interaction with AI in China as well. There seems to be more guidance and obviously concerns around that ⁓ from at least from the top down ⁓ about their mental state, their dependency, or even what constitutes as an AI companion, how we should draw the line on that. We know likefamously a couple years ago, China installed this rule where like, you know, kids under 16 cannot actually play online games on their own without the parents consent. However, that you know, there’s obviously loopholes in practice. But again, it goes back to there are, you know, rules and laws in place to try to protect minors. How do you view all of this? ⁓ Because in the with the backdrop of China trying really hard to diffuse AI into the real economy. And then there is this pushback like you just mentioned onconcerns about AI taking jobs. I feel like there’s also almost like a ironic kind of contradiction happening where, you know, Tiger moms are like, okay, now we don’t need to learn math, we know how to learn AI. And Tiger moms are like saying, how do we optimize getting into, I don’t know, Harvard with AI’s help? And how do we get AI into the education, education, ASAP? I mean, honestly, we don’t even know what the education system might look like in like two decades.from our kids, but at this point, seems like there is like embrace. I don’t know. How do feel about that?Matt Sheehan (25:13)Yeah, a couple strands there. One just on the sort of regulatory side, like this is a long term strand in Chinese tech policy and tech regulation. They always put a pretty heavy emphasis on like how are kids using technology. They have, ⁓ they sort of mandated having like a minors mode on various ⁓ apps. ⁓ This is the regulation I think that you’re referring to as the newly passed. ⁓I translate as anthropomorphic AI. That’s the word that’s the official translation. ⁓ So it basically means AI that, you know, behaves like a human. This could include AI companions that are, you know, literally like a character pretending to be your friend. They could also include, you know, the way that people interact with chat, GBT or Kimmy or whatever, you know, the phenomenon of having AI boyfriends and girlfriends and all this stuff. So there was a new regulation on this that was just finalized, I think, last week andIt has some protections for everybody, specifically around ⁓ self-harm, addiction, and stuff like that. But it has really ramped up protections for minors and for elderly users. So there’s all these kind of specific add-on requirements. For ⁓ minors, it involves permission from parents. Parents can review at least some. They can set limitations on how the child uses the system. They can review.conversation that might have got toned down a little bit in the new version. ⁓ But I think, this is many ways it’s the same concern that surfaces in the US and elsewhere, like California just passed a just passed last year, passed a similar regulation on AI chat bots that I think also had believe it had extra protections in there for kids ⁓ on the education side of things. I guess there’s a couple of things. One, there’s just like the yeah, you say the tiger mom’s like this is ⁓It’s a booming industry of like, I’m going to teach your, you know, four year old AI so that they can use it because this is going to be how they get a job and how they get into school and how they get a job. ⁓ you know, a lot of it is bogus. Maybe most of it is bogus, but it’s very attractive to parents who have grown up in a really, really cutthroat competitive education system where you’re looking for every single edge that you can find.And so that’s, that’s a piece of it on the, from the policy side, they have both sort of AI and education policies, AI plus education policies that they’re pushing in a bunch of ways. I have some friends who are working with teachers over there who are described to me pretty like sophisticated and interesting ways. The teacher that are using AI to lesson plan, to create like really interesting games that keep the kids engaged and learning stuff like that. So you have those, and then on the labor.the labor side of things, they’re also viewing AI, they’re also viewing education as something of an antidote ⁓ to AI fuel job disruption. This is in the, I in the five year version, it is in the five year plan, it’s in the AI plus plan, it’s in a few other places where they say, we’re really gonna prioritize lifelong education. So maybe you used to be an accountant, you lost that job and ⁓ you’re gonna retrain as something else. ⁓which I think is a good, you know, it’s a good attitude to have. If you’re a person, you should always, I’m always trying to learn, you know, books. Um, think they’re great, but I don’t know if at a totally like a, you know, macro population, I know if you’re going to get 500 million people to be constantly staying one step ahead of AI in terms of what jobs it can do now. I mean, a lot of the things that we would have told you go back like two, three years and say, what jobs are going to be disrupted by AI? A lot of the recommendations would have been totally backwards.People would have thought that, ⁓ coding jobs are great. jobs involving creativity, ⁓ illustration, ⁓ stuff like that. AI can’t do those things. It can’t be creative in that way. And it’s like, that’s actually kind of what it’s best at now in some ways. I mean, you can argue about the level of creativity, but like generation of content, generation of images, videos, language. So I’d say it’s a piece of the Chinese sort of response on labor concerns.a fad, but maybe like a ⁓ useful fad within like the sort of education industry. But I’m a little bit skeptical of this as like ⁓ an actual antidote to the disruption that I at least imagine is coming.Grace Shao (29:49)Yeah, it’ll be really hard to be like upskilling, like re-skilling like hundreds of millions of people. Like, it’s just, you don’t even have the capacity to do so if there’s actually mass disruption. alright. ⁓Matt Sheehan (30:00)mean, this was always the response, in the United States on coal miners. We’re going to teach them to code. everybody, all these manufacturing workers, we’ll offer like a job retraining program. I’m like, ⁓ maybe, yeah.Grace Shao (30:05)Yeah.I’ll take generations. I’ll take generations for things to shift, know, resources to shift, people’s mentality shift, you know, for a while, like, when many, when remember the first wave of like, a basic rural kids no longer wanted to work in factories and wanted to go to urban cities, there was a surplus essentially service providers and then like everyone eventually became a DD driver or a food delivery man and thenNow we’re seeing a reshuffle in that population again, where people want to move back to their rural, cities. so I think based on how society is evolving, opportunities will arise without even us realizing anything, hopefully in the best case scenario, where people will find opportunities to reskill. ⁓ But I want to talk about something that’s a bitGrace Shao (30:57)I guess not heavy, but actually not many people understand, even including myself. So you really look at the government and the policy structure of ⁓ China’s regulators in the cybersecurity space and whatnot. ⁓ There’s so many players. There’s the CAC, then there’s the NDRC, there’s the MMIT. I can keep on naming acronyms, but can you give us a really, really quick high level understanding of who’s regulating whom? ⁓how do they actually work with each other and are their KPIs aligned before we get into more about how you know, how China’s policy is shaping the technology and AI ecosystem.Matt Sheehan (31:36)Yeah, maybe I’ll do it. ⁓ I’ll introduce a couple of the players and I’ll do it somewhat chronologically in terms of like when have they become important or rise and fall in importance. So AI policy, like the first really big policy document was the 2017 National AI Plan. It was released by the State Council, effectively sort of China’s cabinet, sort of the highest level of government. But ⁓ people who are in the know say that that was largely sort of drafted and pushed by the Ministry of Science and Technology. So this isreally like the policy wave of like 2017 to 2020 more or less. And it’s the Ministry of Science and Technology and it’s also the Ministry of Industry and Information Technology, MIIT. So these are really the organizations whose job it is to promote science, promote innovation, and MIIT is more of like the industrial applications of the technology. So they were kind of in the driver’s seat in that period of time. They were the most relevant actors. They were the ones who were driving real activity.starting in 2020, 2021, you had the CAC, the cyberspace administration of China, really like come to the center and become the most important actor in AI policy. The CAC, it’s basically the internet regulator. It was created in 2014. It was largely created to kind of like get the Chinese internet under control from a sort of political content ideology perspective. They’re connected to the Ministry of Prop, or the propaganda department.publicity, as they say now. ⁓ So from 2021 through 2023, the CAC was the one rolling out these binding regulations on recommendation algorithms, on deepfakes, on generative AI. And these are the regulations that actually force companies to do things. They actually force companies to register their models, to do pre-deployment testing, at this point to label AI-generated images in different contexts.There’s for that 2017 to 2020. It’s kind like the go-go period. Let’s just like push this industry forward. You have the Ministry of Science Technology, MIIT. And then from 2021 to 2023, it’s really the CAC. This corresponds roughly with the tech crackdown of 2020 through the end of 2022. That was a period when the CAC, the CAC is kind of at least historically, it’s kind like the bad cop of tech policy. They’re the ones who are like telling companies like come in and drink tea and we’ll tell you what you’re doing wrong or, finding companies in different ways. Cyberspace Administration of China, yeah, CAC. ⁓ Some people call it CAC. ⁓ And then sort of one of the more significant changes from 2023 to now is the rise of the NDRC, the National Development and Reform Commission, Chinese Fagawei. ⁓ And they are a macroeconomic regulator. They are like what grew out of the sort of state planning apparatus. And they’reGrace Shao (34:01)This is a cyberspace administration of China, right?Matt Sheehan (34:30)really powerful, they’re kind of a super, super ministry within the bureaucracy, but they’re not sort of directly, there aren’t that many direct connections to AI. They deal with, they deal a lot with money. They have money to give out for projects that funnels into compute projects and stuff like that. ⁓ But they wouldn’t be like who you would think of as the go-to AI regulator. What I was told and what I feel pretty confident ⁓ is what happened is that in some time in, I think, 2023, maybe mid to late 2023 and then into 2024, the top leadership in China essentially said, hey, we need a little more balance in our AI policy. The last three years it’s been led by the CAC. They’re kind of a bad cop. They’re really focused on controlling the technology, controlling the sort of output, the content, the ideology from it. And that is important. That’s kind of their first priority. But we need to rebalance this a little bit. We need to move out of our total tech crackdown era. And now we realize like our economy isn’t doing great.We realized we’re behind the US after CHAT GPT came out, and we need to balance this out. And so they empowered the NDRC to be a of a coordinator across AI policy, someone who is intended to take the input from the various ministries, from Ministry of Science and Technology, MIIT, CAC, and to try to make it little more coherent and balanced. And so that’s kind of the role that they have played for the past few years. The details of how that works out areshrouded in secrecy, you know, you hear little tidbits here and there. But there have been like visible manifestations of it. They had not, when they released these regulations, usually there’s a sort of a lead regulator on it or a lead policy document person on it. And then various other ministries, they co-sign it and they’re like listed below. NDRC hadn’t been on any regulations prior to 2023. And then starting in 2023, they were listed second as like the second sort of most important organ.policy body on these things. essentially we have this kind of like 2017 to 2020 is this like go-go period. Let’s diffuse. Let’s push the technology. Let’s push innovation. 2020 to 2023 is this more constrictive crackdown. Let’s build the regulatory infrastructure for things. And then 2023 to today is just like, let’s balance this out. Let’s not be purely focused on the content and ideology concerns. Let’s also be thinking about development. Let’s be thinking about employment. The NDRC is actually allegedly one of the groups that is very concerned about the employment impacts. you know, tons more details that I will love to go in on, but maybe that’s a starting point.Grace Shao (37:06)No, I think it’s super, super helpful. I just understand the nuances of like what their actual KPIs even are and like, you know, who does what, how they work together. I think that’s really helpful for lot of listeners and even investors who are trying to follow the space and just confused by acronyms. But help me understand now, like you say that 2023 to now essentially is in the same kind of era. However, I feel like at least from the capital market perspective, you know, the last year might have seen a bit of ashift again, you know, it was a bit of a let’s go AI, big tech AI, all the labs, let’s go, let’s go IPO. Then obviously the deals, some of the deals didn’t come through, some of the IPOs didn’t come through. ⁓ There seems like you even said people are being told to tamper down their excitement a little bit. Is that aligned with what’s happening with the policy side of things? Or is that actually more a reflection of just, frankly, you know, the AI space not being that exciting right now, you know, since the Gentic ⁓ kind of breakthrough. We’ve not seen more consumer and breakthrough. Also, there’s a lot of talk about, you know, there’s no obvious proof ROI on all the spending from all the big tech right now. Help me understand all that, I guess.Matt Sheehan (38:15)Sure. Yeah. When I was breaking down those errors, is largely its policy, but it’s already kind of like government attitude towards it. It’s like which, you know, they’re always in some ways swinging back and forth, going back and forth on the seesaw between, you know, control development, control development. And that 2023 to now being one era is sort of in that sense. It’s the period of rebalancing more towards development. There’s tons of sort of wiggles in that process andthings they’re pushing more and retreating on. But from a positive perspective, that’s the overlay. ⁓ In terms of like the capital markets, investments, I mean, I think this is kind of at least for people in the United States, it’s kind of like the one of the most misunderstood things about the Chinese AI ecosystem is that it is really like cash constrained, that it is not like the United States where, you know, open AI is just like sucking in.the tens of billions of dollars from a huge variety of investors are just spending huge capital outlays, which people talk about, is it a bubble? Is this going to come back to bite them? That’s an open question. But in China, you don’t have the concern about that bubble because there just is not the same level of infusion of cash. when a couple of the companies did IPO recently, Z.ai, formerly Jerpool and Minimax IPO in Hong Kong, and I think I’m notGrace Shao (39:28)100%.Matt Sheehan (39:39)really an IPO guy. think the IPOs were like modestly successful, but the valuations are just, yeah, the valuations are, yeah, not even close. And ⁓ it reflects a lot of things, but it largely reflects like a funding environment, a business environment, a macro economic environment, and the general sort of attitude towards risk investment. think I was just reading something that ⁓ Ray Ma from ⁓ TechBuzz.Grace Shao (39:42)So that’s six to eight billion dollars. The valuation is tiny compared to American peers.Matt Sheehan (40:06)China was writing on this. She’s always very good on these topics. yeah, it’s just people kind of assume that there’s like infinite money in China. They’re like, yeah, the government, whenever they want to, they just like turn on the taps and then, you know, it’s like, no, that’s not how it works. And like the VC ecosystem is much smaller, much more new and immature. And so it’s a different story.Grace Shao (40:28)on that note you know I was just in SF like last month and I met with quite a lot of investors and people’s kind of I guess misunderstanding was often twofold. One is exactly your point, people are just like oh China’s so rich the government just gives money all these AI companies are backed by the Chinese government I was like 100 % no first of all like there’s some other issues happening in the background but like the government doesn’t even are you know it’s kind of cash constraint and not even that much right now second all these companies are definitely not being backed by the government in any sense in factMost of don’t want to take municipal governments or provincial government money because you get kind of tied into, you know, what we’re seeing is, you you get forced into working with the government and it constrains your profitability and commercial goals. On the other hand, another really big misconception was, I thought quite funny was that people often ask, was open-claw frenzy because the Chinese, average Chinese consumer or user were really, really concerned about privacy issues. So they wanted everything on edge.I was like, hmm, like again, it’s kind of like just not, a major conversation people have. Like I think I hate to generalize, but I think because of how the internet ecosystem is in China, people by default have kind of ceded to not thinking too much about privacy or personal data issues as much. So that definitely isn’t. So I kind of want to bring the conversation that this, you know, likeWhat are some biggest misconceptions you think people have and how do we help them understand and bridge that gap a little bit better?Matt Sheehan (41:59)Yeah, I think yeah some of the stuff that you point out is correct like If you’re if you’re a if you’re a startup if you’re like a small medium company You I’ve talked to these people they’re like actually like we do not want to take government money if we can avoid it not just because we get kind of in mesh but like Entrepreneurs are legitimately afraid that if they take government money and then their company doesn’t work out and they lose the government’s money like they could end up like on the hook like in jail thislegitimate fear that it was stated to me by someone. like, you know, is that happening to entrepreneurs everywhere? No, but it’s like you don’t. ⁓ The government. It does a certain amount of sort of VC-esque investing, but there’s not really that VC mentality of like high risk, high reward. Like we know that most of this is going to go under. It’s kind of local governments at least have been trained on like real estate investment, which is like 10 percent, 10 percent, 10 percent every year.And this idea that most of these companies that you invest in are going to fail is not really ⁓ deeply embedded there. I do think some of the companies do rely on government funding in different ways. ⁓ mean, Z.ai, Drupal, one of their biggest, maybe their biggest single revenue stream is from ⁓ building custom models, but custom applications for ⁓ state-owned enterprises, local government, stuff like that.Matt Sheehan (43:28)When you listen to them in interviews, they’re like, it’s not that big. Maybe it’s 40 % or something like that. But it’s a significant revenue. It’s part of their business model. So there’s that type of a connection to government. With DeepSeek, that’s a company that’s kind of quite mysterious. And we don’t know exactly where all their money comes from. Is it all earned? I think the government got more hands on with them in the sort of aftermath of the DeepSeek moment. You had reports about the government taking passports away frompeople who worked there to make sure like you guys stay local ⁓ or the government was like vetting investors was another this is reporting the information. ⁓ But the idea that like these companies are just they just have the kind of like the hose of government money just flowing in at all times and therefore they don’t have to think about anything else is just not it’s just not real. ⁓ They’re they’re much more constrained cash constrained. ⁓terms of like trying to misconception on Chinese AI regulation, AI policy, this is like my, you know, much of my job is like first getting across like, the trend does actually like seriously regulate the technology. And then, you know, the next layer being like, it’s not all people think, you know, it’s an authoritarian system. She didn’t think he must just kind of like sit down and just like write the regulation. So like nothing matters except what he thinks. And we don’t know what he thinks. It’s like, no, like he doesn’t. There’s this actually very complicated and sophisticated policy ecosystem of,legal scholars and the companies are doing their lobbying and their thought leadership and, you know, they’re responding to public outcry over things. And I think that’s a, know, you can get this across to people, but it’s certainly not the people’s default mental model of how China works on policy is ⁓ just does not reflect the kind of sophistication in this zone. And it’s somewhat understandable. Like there are policy areas where Xi Jinping just like makes a decision and that’s.that’s where things are going. Like I think we saw a lot of this in the kind of 2020 to 2022 era. But as an AI policy, COVID, AI policy, it’s not that way. certainly things are not, people are not gonna like, you know, actively push things that are totally against the will of the top leaders, but they are within the constraints of like,Matt Sheehan (45:52)the direction of travel, what the CCP is good with, what she wants to do within that kind of very wide lens. It’s really individual people, scholars, bureaucrats, companies that are filling in all the details on this. And it’s a very sophisticated system because they’ve just had a lot of ⁓ had a lot of bites at the apple. They have like passed, I don’t know, eight, nine different A.I. already.The regulators at the CAC have been getting documentation from AI companies for three, four years. They’ve been building evaluations. been like, and they kind of, got their reps in with AI policy and it leads to a more sophisticated ecosystem.Grace Shao (46:33)⁓ yeah, so the last kind of section I want to focus on is just getting into the nitty gritty about, you know, the policy and the security and safety kind of side of things we touched on in the beginning of our conversation. you are one of the few in the West, I think, and talk about the nuance of the word, which you just explained, it’s security, but also safety. ⁓ help us understand.how to interpret that when we read about that. It actually even helps us understand a little bit of what’s happening in the West. Like, I feel like there’s the governance people, the security people, the safety people, but from someone who might not be in that ecosystem, people are conflating it a little bit. And I just want to understand, you know, how do we understand each of their objectives, again, KPIs, or even their goals?Matt Sheehan (47:20)Yeah, yeah, basically, it’s really complicated. It’s very context dependent and it’s always changing. ⁓ I think maybe the first key thing to understand here is like the very particular meaning of AI safety in the West like that. The West AI safety ⁓ largely refers to kind of a specific camp ⁓ of AI development and policy people that are, you know, believe that AI is going to achieve human and superhuman capabilities.And this could pose like serious, maybe catastrophic risks to people. like, that’s a somewhat coherent community in the United States that has a certain amount of power. Their power kind of ebbs and flows depending on things. But like when you say AI safety in Washington, D.C., it means one quite specific thing. ⁓ In China, that community, it has started to emerge, but it’s much newer. It’s much more recent. It doesn’t have the deep roots that it has in the West.And ⁓ the way that the word is used in policy documents is both confusing and has changed over time. So a lot of times when ⁓ just to kind of put a little color on the terminology, Anquan, when ⁓ when you’re talking about cybersecurity in Chinese, you say Wang Luo Anquan. So it’s like network Anquan, network security and cybersecurity means something very different fromAI safety from super powerful AI systems posing risks. And so there’s one sort of category of mistakes, which is to be very naive and to read all the Chinese policy documents. And every time they say a word that’s translated as safety to believe, wow, they’re talking about AI safety, they really, really care about this. That is a very naive and incorrect reading of things. ⁓ But in the past, I would say, 18 months, two years,you have seen a pretty significant uptick in the way that people sort of in and around the system and to a certain extent in and around the companies, their level of attention to what we would call in the West, AI safety to these more kind of large scale, potentially catastrophic risks from powerful AI. I’d say this is, there’s like sort of levels and degrees of this. There’s people talking about this. There’s it showing up in government documents in one way or another.And then there’s actually implementing this either through like binding regulations or through sort of companies doing their own testing and evaluation to try to their own sort of safety research and their own safety testing. I’d say what we’ve seen so far is a large increase in rhetoric, a large increase in sort of awareness within the policy community about these safety issues. We’ve seen it to start to show up in more significant documents. I think the one you’re referring to early on that I was working working on analyzing is calledThey call in Chinese the AI safety and governance framework 2.0, which is in many ways put out by some organizations underneath the CAC, the internet regulator. And it’s kind of ⁓ their attempt to diagram and do an initial discussion of how they see different risks from AI ⁓ and how are they going to mitigate these risks. Oftentimes they’re focused on technical standards as a mitigation. And there was a AI safety governance framework 1.0 in 2010.fall of 2024 and there was a 2.0 in fall of 2025. And just between those two documents, you can see real increase in the frequency and to a certain extent, the sophistication of the discussion around these risks in China. I’d say it’s pretty significantly below the sort of the AI safety discussion in the United States, but it’s on the radar. will counter, they’re like, okay, that’s great that they’re talking about it, but are what, youAre they just trying to trick us? Are they trying to make us believe they believe in safety? Are they saying it but not doing it? And ⁓ they’re like, we have not seen much in the way of like, we certainly have not seen like concrete binding regulations that sort of implement safeguards on this front. And in terms of what the companies are doing, it’s quite opaque, but ⁓ we don’t think that they’re doing the, I’d say.pretty confident they’re not doing nearly the level of sophistication or intensity of safety testing as you see at places like OpenAI and Anthropic. To me, this seems somewhat normal. This is kind of a process. Chinese companies have been behind. The government has perceived itself as being behind. When you’re behind the frontier, you’re not as worried about frontier risks as you’re like other people are going to get to those first and we need to catch up. ⁓ So I see this as kind of like a long-term process. And I think that the sort of increase in discussion about this isyou encouraging if you’re concerned about these issues. But it’s a really don’t want to be just kind of reading the documents and say every time we see Anquan being like, wow, China cares about AI safety. Look at all this stuff. It’s much more ⁓ nuanced and evolving, ⁓ evolving quickly, I would say.Grace Shao (52:21)I know like when we spoke a couple months ago, just catching up, you were saying a big part of your job is also trying to help, you know, bring the two sides together. Obviously, you know, ⁓ it’s been challenging given the geopolitical backdrop, but how do you think the global AI governance space can work together? ⁓ Are we going to see, you know, kind of the two world superpowers and two super AI powers, ⁓ you know, guide?in different directions or do you think there are certain issues where they need to come together and they will come together and are coming together? ⁓ For example, to your point on safety issues around protecting humanity, protecting children, are these things that you are seeing collaboration?Matt Sheehan (53:08)Yeah, I’m kind of ⁓ both an optimist and a pessimist on this front in that, like I said, I have very, very low expectations for the United States and China to work together on anything. I have very, very low expectations of any type of a binding agreement or some type of detente where we both shake hands and kumbaya and we’re both going to be very safe with AI and we agree and it’s great. I just don’t expect that. ⁓So in that way, I’m pessimistic. I think attempts to try to sort of preemptively create these global governance structures that are going to bind both of the countries in advance so we never reach these dangerous thresholds. ⁓ That’s just not where I’m putting my bets. I think it’s good. We need to make all kinds of bets on this front, and it’s good that people are working on this, but that’s not where I’m putting my bets. Where I’m putting my bets is on a much morelimited kind of narrow bore, but I think potentially highly effective form of ⁓ engagement. wouldn’t even say cooperation. I wouldn’t even necessarily say coordination. ⁓ my sort of the term, my mental model for it is something I call AI safety in parallel, which is that like the two ecosystems are going to be moving somewhat in parallel. They’re both going to be pushing the technology forward. They’re both going to be working through safety issues from a technical perspective, from a policy perspective.And as we kind of move forward in parallel, we’re not going to be telling each other what to do. And we’re not going to be like, okay, I’ll do the safety thing because you are. You told me you’re going to do it, so I’m going to do it. We’re not like sort of moving in lockstep on this, moving in parallel. And we need to have these touch points. We need to have touch points where the two sides develop some form of mutual understanding of what the other side is doing. They understand how other side is thinking about the issues. They understand how they’re perceiving these risks. That’s one of the reasons I do the risk ranking.stuff and in some cases trying to share best practices, explain kind of explain what we’re doing and why we’re doing it and have the Chinese side explain what they’re doing and why they’re doing it and then where possible share good ideas that we think are sort of uniformly good in the U.S. and China. If we think that we have a policy intervention maybe it’s around ⁓ certain types of pre-deployment testing. ⁓ It’s good to communicate that.to the Chinese side. And it’s good to have the Chinese side communicate some of ⁓ the reasons and the sort of the specifics of what they’re doing on these fronts. We’re not here to just like trust each other. I think a lot of people are very worried that the Chinese side is going to, is they’re going to trick us. They’re going to say they’re doing it and they’re not, which is like legitimate concern. know, that’s, this is high stakes like geopolitics and powerful technology. So you don’t take anybody’s word for it. But when you have these conversations in talking with people, you can get a,a sense of their level of sophistication when talking about the issue. If someone is talking about AI safety and they’re like, yes, humanity first, protect the humans, control the machines, that’s our policy. And it’s like, okay, is there anything more to that? They don’t have more than you kind of know that they’re actually not really thinking about it. But if you can get into a more ⁓ deeply engaged discussion, you can see like, actually, yeah, they’re working through these problems themselves. You can see it in the way that they’re.discussing it. You can see it when they talk about their specific regulatory mechanisms. You can see sort of the connection between sort of action and outcome or thinking and action. And so my model for this is like, we’re not going to agree on things. We’re not going to sort of trust each other. But there are ways that we can both be moving forward at the same time and comparing notes, checking in, getting a sense of what the other side is thinking and doing that I think could contribute to safety in a meaningful way.Grace Shao (57:02)I think that’s fair and I think the word you kept on using trust is quite interesting because I feel like whenever I speak to people in the industry, ⁓ there’s just such a lack of trust even within whether you want to say countries or communities or beliefs and value systems and a very, very optimistic, naive way, I really hope that there can be a bit more consensus on certain things like that need to be protected in practice, like such as children, right? And how we go ahead with that. ⁓ But okay, I don’t want to end on a super somber note or anything, but... ⁓ The takeaway is trust nobody. That was...Matt Sheehan (57:34)It’s optimistic in a way. think this can’t... When you do see... Well, trust nobody, buttalk and see if you can share some good ideas along the way. think there is real... ⁓ I’ve seen some real sort of traction from these type of things and I think it’s limited. We’re not going to get some kind of hard guarantee that China is going to be perfectly safe or we’re both going to...Matt Sheehan (58:04)do the right thing. But within with those low expectations, with those kind of pessimistic expectations, there are there’s progress that can be made.Grace Shao (58:13)⁓ I do want to ask one question that’s kind of been happening around right now. That’s been kind of happening like the whole idea of Chinese open models seem to take a little bit of a sidetrack and starting to kind of only release their most frontier related models and close weights. ⁓ Obviously from a very like, you know, capital perspective, where I study it is I feel like it’s a lot of it is because they need to see our eye. They cannot keep doing this because they’re not making money. API sales not enough to, you know, sustain the kind oflong-term ⁓ business as well as research costs. Are you seeing anything from the policy side? Like do you think there’s been a policy shift? That’s also kind of why I asked earlier if this year somehow, know, last year there was a public embrace by the government saying we should open source our technology. Has there been a shift?Matt Sheehan (59:04)So ⁓ I’ve seen sort of little tidbits around ⁓ sort of public policy concern about open weight models, but not enough that I would call it a shift. ⁓ In that document, the AI safety governance framework 2.0, it was interesting because it was the first time that there was a fair amount of ink spent on potential risks from open source models. The risks they were primarily talking about wereessentially if there are vulnerabilities in these models in some way, either maliciously inserted or just a vulnerability mistake, those could proliferate throughout the ecosystem because you have all these downstream models and that could lead to impacts. There’s a little bit of a mention of like, maybe open models will be used by criminals and stuff like that. I certainly don’t see this shift in specifically Alibaba strategy asMatt Sheehan (1:00:04)in reaction to a significant policy shift. mean, it kind of makes, yeah, corporate decision. It makes if you’re going to be spending tens of billions of dollars building models or at least hundreds of millions, billions, billions of models, giving it away for free is a. That’s a choice. I think there’s reasons why it’s advantageous for China to do that, or at least it was for a stage like it was going to be pretty hard.Grace Shao (1:00:09)corporate decision then.Matt Sheehan (1:00:34)to get people around the world to kind of believe in ⁓ Chinese models if they were only going to be able to access them through API. ⁓ You know, a lot of American companies that are, you know, deploying ⁓ Quan, Alibaba’s model, I don’t think they’d be doing that if they, I don’t think they would have at least made that leap initially if they had to sign a contract with Alibaba and they believe that maybe their data was going back to China or, you know, the model was more of a black box relative to them. So maybe the open model wave was a very good period of publicity. ⁓that might pass with time, but I don’t know. I think it’s being more nuanced. I’m not the expert on this. Nathan Lambert, runs the interconnects sub stack, and then Kevin Shue, who runs the interconnected sub stack, are much, much better and more sophisticated on this, and they have good writing on the ecosystem. But from a policy perspective, I haven’t seen a shift. I think it just kind of makes sense.Grace Shao (1:01:35)Yeah. Nathan, I actually spoke about this ⁓ a week ago, and I think both of us kind of feel like it’s really more of a capital and business constraint that’s really driving this. ⁓ But I just wanted to hear from you if maybe there was some kind of a top down initiative as to but it doesn’t seem like it. Right. I have one last question for you, which is a question I ask every single guest that comes on. ⁓ What is one differentiated view you hold? And actually, I try not to limit the question to just about AI. However, most people do want to talk about that.Matt Sheehan (1:02:02)Mm.⁓ Differentiated view. ⁓I mean, is... I don’t know where most people fall on this, but I think one thing I’ve been thinking about lately is just language learning and what’s going to happen with language learning in the era of simultaneous really good translation either through hardware, devices, or software. And I’ve been reflecting, why have I been spending 16 years or more learning Chinese and just in the...Matt Sheehan (1:02:40)mud of trying to learn and remember this language. And I guess my maybe differentiated view on this is I think it still is really important. And I, I waver, you know, I don’t want to just be like, ⁓ you know, justifying to myself why I’ve spent all this time. And like, I, don’t know that I would tell like a 16 year old, like I want you to invest, you know, 10,000 hours on a language when, when you can have it all translated for you, but they’rethere just is something quite meaningfully different about reading these policy documents, about listening to people and hearing the original language and just knowing how that language is used in all these different contexts that gets flattened by translation. I use machine translation all the time. I throw documents into it to get either a first draft or to get something I can share with people. But I guess my view is that I really thinkWe need people to keep like actually learning Americans to actually keep learning Chinese. And I also just think it’s so much fun. just, was earlier today, I was thinking it was like, why is this, it’s been such a, like a joy in many ways, extremely painful, but kind of a joy to like really struggle with a language over time. And so that’s my take.Grace Shao (1:03:59)No, I actually agree with you and I think if anything, I’ve thought about this a lot. So I’m raising a trilingual child by nature because we speak English at home, our parents speak Mandarin, the environment she lives in, they live in speaks Cantonese, right? And I think to your point, in many ways I’m like, wow, it’s actually, it would be so easy for them to travel the world and communicate with people. But the reason why I pushed them to learn the language is really to communicate, to understand a culture and the people and more.nuanced way, even for myself, like my parents pushed me to learn Mandarin. It like to your point, it’s so painful. But the ability to speak to my mother in her native tongue and understand her is so much more complex and you appreciate much more when you’re older, even though my parents speak English, obviously, but when we were young, we would speak English to him. As I got older, I actually enjoyed speaking to them in Chinese much more because you hear aboutMatt Sheehan (1:04:55)Hmm.Grace Shao (1:04:57)It’s also your personality changes, right? Like you kind of get to the core of who they are in their native language and their native way of expressions. So I think for sure, I agree for certain languages, there’s still such value, if anything, even more so to understand a human connection, human connectivity. then for pragmatic reasons, like I took two years of German, I remember nothing. I probably wouldn’t do that again to myself, especially in my class. a bunch of third gen German kids where they spoke the language at home but they can get away with saying they were doing beginner’s German, you know? But yeah, so I appreciate that. Thank you so much Matt, thank you for your time, I really appreciate it, we finally got together to do this episode.Matt Sheehan (1:05:34)Yeah, thanks for having me. That was fun. Get full access to AI Proem at aiproem.substack.com/subscribe
• Tencent's QClaw goes global, aims to serve the average consumer user, with PM Shuyu Zhang 21.04.2026 45min

  Amid Anthropic’s success with coding products, many AI labs and companies have also tried to lean into that vertical. OpenAI has stepped back from courting consumers and shut down its video model division, Sora. Alibaba, meanwhile, has more recently begun releasing closed-weight proprietary models and is reportedly pushing the Qwen team to find clearer paths to monetization. The Chinese tech giant has also launched Qoder, a Cursor-like product under the Alibaba umbrella, which we interviewed last year.But despite all this, Tencent remains notably committed to the mass consumer market. The OpenClaw frenzy has already led to five different Clawbot-style products emerging across its ecosystem. Joining me today is Shuyu Zhang, Senior Product Manager of QClaw, to break down the thinking behind that frenzy, from the cultural logic to the business rationale to the product design choices shaping it all. QClaw is to be accessible to everyone on April 21. It is the first consumer-grade AI agent built on OpenClaw. No technical setup, scan a QR Code, and the agent will be live in 3 minutes. Product link: qclawsg.qq.comWaist list: https://docs.google.com/forms/d/e/1FAIpQLSeIfEzlOV8jq_tGMbV5mqTSALyufE0kZ933XqE3Fnha1_CRfA/viewform?usp=publish-editor (Founding Claw — limited 20,000 slots)Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently. Season two will host a series of guests from early-stage investing, as well as builders, founders, and product managers. For more information on the podcast series, see here.To find the previous episodes of Differentiated Understanding, see here.Chapters00:00 Introduction to OpenClaw Frenzy in China02:30 Shuyu Zhang’s Journey and Insights on AI Accessibility05:27 Cultural and Societal Factors Driving AI Adoption in China07:48 Understanding OpenClaw’s Popularity and Usage10:36 Exploring Tencent’s AI Product Ecosystem13:33 QClaw’s Integration and User Experience15:45 The Philosophy Behind QClaw’s Design18:14 Raising the Claw: The Concept of Personal AI22:48 Incremental Value in AI Products23:33 User Experience as a Priority27:55 Understanding User Needs and Safety Concerns33:07 Business Model and Global ExpansionTranscript (AI-generated)Grace Shao (00:00)Thank you so much for joining us today, Shuyu. I’m so excited to have you on the first day. I understand it’s such a pleasure to have you here. I’m really excited for today’s topic because it’s something that’s kind of been at the top of mind for a lot of people. Why was there a massive open call frenzy in China? Why did it take off? The thinking behind it all. How did WeChat go about opening up to this whole new era of agentic AI within WeChat development? And there’s no better person to talk about this topic than you. So first of all, for the audience, please start with telling us about yourself, your role. How did you end up here? Shuyu Zhang (00:35)OK.Hi, everyone. I’m Shuyu, product lead of Qclaw. Also, the architect behind it’s overnight growth story in China. I achieved breakout success with zero marketing spend. I got a master’s degree of finance from Washington University in San Luis, then worked for Alibaba Group as head of AI product in Sanyo for four years. We mostly crafted AI product for business there. Almost everything we did was about AI for work. But one day, I decided to do something different. I want to get more exposed to consumer side.to experiment with the chemistry of AI for common people. Tencent is famous for its consumer side products I want to work with and learn from consumer product experts here. So I came here last September. Yeah.Grace Shao (01:15)Really cool. And I think one thing that really kind of resonated with me is when we’re talking about AI agents right now, how a lot of times, you know, the products still don’t feel that intuitive for non-technical people. And you yourself joked and said, you know, look, you’re like a social science and liberal arts student. You’re not a technical person yourself, but you’re able to lead the product design for this. And your mission is really to make QClaw more accessible to non-technical people. Tell us about that and the thinking behind it all.Shuyu Zhang (01:43)Okay, okay. Actually, the story starts when I was working for Alibaba. Initially, we worked for the engineers to help them improve their working efficiency with AI. But later I found out that this group is overly served. They only account for like 1 % of all of the people, butEvery day there are lot of products designed exactly for them. And I think this is still the major theme of AI revolution since 2023, where people think the way to AGI lies. But these groups also are easily unsatisfied. They raise a lot of questions about the services that the AI provided, and at the same time, they’re afraid to be replaced. So I think the vibe is strange. So when I’m moving to a new environment to Shenzhen, when I was hanging around, I found a lot of interesting common people are using AI.And they’re getting a lot of happiness and convenience, even though the product capabilities for them are not the most advanced and designated for their needs. They’re still satisfied. There are two interesting stories. The first story is when I was getting home on the plane during spring festival, I met a 12 year old girl. This girl looks smart. She was playing with some toys. And after that, she was playing with a chatbot during waiting for the plane to take off. I was shocked.Because in the past, when I am younger, we usually play games during while we’re waiting for the plane to take off. But the youngest generations are playing with chatbots. This is interesting. And she’s using it to either cut the photos and even make phone calls with the chatbot. And when I talked to her, he also told me that when their friends are hanging out, the 12 years old girls are hanging out. They’re playing with the chatbot too. And I was really shocked by that story. And the second story is during the early January, Yang Liping was doing opera in Shenzhen. And I went there too. And a 50 year old lady sitting next to me, I cleared at her phone screen. The phone storage was actually out of use.But she didn’t delete the chat box applications in her phone as well. She only reserved WeChat, TikTok and the chat box applications in her phone, even though it’s already full because she might not use the very advanced phones. Yeah. And then I found out that for common people, the requirements or the needs for AI exists as well. And problems widespread for the problems, not widespread, for the problems widespread, but not complicated. But the supplies for them are far from enough. And I know there are a lot of people are chasing higher and stronger AI, but the gap between the bottom and the ceilings, 90 % of the people are between them. I want to make a product that can feel and satisfy these people’s needs. Yeah, that’s my story.Grace Shao (04:30)That’s really interesting. think it really ties into a theme that we’ve been writing a lot about AI-prone, which is also about how China is really embracing it as a full on mass market product versus I think in the West right now, AI is still really used by a select group of knowledge workers or certain kind of demographic. I do want to double click on that, which is what do you mean by the young girl is playing with AI? What was she doing actually?Was it interactive with their friends or were they trying to build products or what were they doing?Shuyu Zhang (05:00)Okay. The young girl, when she was playing with the chat bot applications, she actually sent a photo of her roughly taken, not in a very good light or in a very good background. And she just asked that, asked that application to curve it for me to make me look prettier or make me look funnier. She’s not actually a, because I asked, I also asked her a very interesting question. I asked, do you post TikTok shorts or Instagram?She said, I don’t because I don’t like to show enough my life to the public, but I just like to see my photos in the funny way. don’t even though I use AI to, you know, process my photos, I’m just enjoying it by myself. I don’t want to it to other people. And the happiness of the AI processing of the photos is already enough for her. And this is the first scenery she’s using. And the second scenery is that she actually stays at school all the time. She didn’t go home during the from Monday to Friday. So she told me when she missed her mother, because her mother worked in Shenzhen and is a working mother, her mother doesn’t have a lot of time to, you know, FaceTime with her or the teacher doesn’t allow that as well. Because when she go back to the dormitory, the roommates are silent. They can’t do that, but she can always talk to that chatbot. It’s like a companion. And that also makes me feel warm actually, but also little bit sad for her. And the third scenery for her is that she told me she would call the chatbot because the chatbot never blames her and the chatbot always holds her words because sometimes for I don’t know, for the young generations, a lot of their topics are hard to get for the friends, but She said, chatbot is always a good friend because the young generations, they don’t actually care about the or they don’t know about the appearance of people or the words behind the words. But the chatbot is always blunt and sincere and always happy to chat with. Yeah, this is the three scenarios she’s using it.Grace Shao (00:00)Following up on our previous conversation about why Chinese people seem to have a much more optimistic approach to AI and why did the open claw, why did open claw take off in China, like such like wildfire.Shuyu Zhang (00:15)Okay, so I think Chinese people in general, embrace technology with open arms. They have a strong, better self mindset. They believe that new tools can help you learn faster, work smarter and live better. People want to upgrade themselves. And the second point, and also this is a key, for over a decade, Chinese tech companies have been quietly lowering the barriers using technology. They make complex things simple. You don’t need to be an engineer to call a DD or take out on Meituan or buy anything online. It just works. So people naturally expect that new technology will make life easier, not harder. That’s exactly what Qclaw and Tencent’s lobster products did. They took OpenClaw’s powerful but geeky core and wrapped it into a simple IM plugin. The barrier to entry dropped from weeks of learning to 10 seconds. That’s why lobster caught fire in China. not because it was the most advanced AI in the world, but because someone finally made it useful for everyone.Grace Shao (01:13)I think that’s a really interesting take and I think in general it kind of ties together to the bigger kind of sentiment as well where overall the reputation of Chinese big tech such as Alibaba, dance and tents that still are perceived quite positively by the average person to be an employee there is something very prestigious, ⁓ very like sought after. Whereas in the US, I think in the last couple of years, there is a bit more contention or negativity around the big text, whether it’s monopoly or behavior or even the capital allocation that they’ve really received that’s unfair compared to the rest of the country. overall sentiment is a bit different. I think that really did contribute to this as well.Shuyu Zhang (01:58)Yeah, exactly.Grace Shao (07:00)I think that’s really eye-opening and I think I’m not trying to hijack this whole conversation, but to me when I hear that, I think I want to like you said, the companionship is really great and the ability to help the child feel more connected to her mother is great. But at the same time, I do feel like there is some concerns or worries about that. ⁓ I did notice that the Chinese regulators recently pushed out some regulations around actual child use of AI.which we’ll have an expert to join us one day to talk about this. But I think it’s interesting to showcase a phenomenon of China really embracing this at a mass scale. On that note, I don’t want to go too deep into the child use today, but on that note, I do want to ask you, why is it that China seems to be so amazed and enthused about AI, and especially this time with the open-claw embrace?Obviously at Tencent’s headquarters, saw pictures going viral where people were lining up and getting open-clawed saws. Various big tech, whether it’s Alibaba or Baidu pushed out similar products like yours, like Qclaw. Could you explain to us from the big picture, is it cultural? Is it societal reason? Is it a top-down policy reason? Is it commercialized as a business reason? Is it product design, like you said? What is it that really drove like everyone going AI.Shuyu Zhang (08:16)Okay, so the first background of Chinese AI is that since 2023, since the chat-chip goes out and after that Baidu, Alibaba and also Tencent and also most importantly DeepSeq, it’s widespread of AI, their LLM makes AI widespread in China already because during last, the one before the last Spring Festival, almost like 200 million people use DeepSeek every day. So the basic foundation of people knowing AI in China is already widespread. And why OpenClaw is also going wild in China this year? Because OpenClaw’s capability is quite different from other products that people are familiar with currently. So this is the first decision point. more about the culture thing.Firstly, China is a fast developing country and Chinese people are diligent by nature. And every generation, people of every age want to be a better self. So during this race, anxious middle-aged actually is a very big contributing factor. And also fear is also a big pusher. They always fear, you know, when they’re getting old, they will be lagging behind. So they want to, you know, learn more things.learn what the new generations or the techie guys are doing. Yeah, this is the first decision point. And the second point is that hiring a personal assistant or secretary here is not common, but everyone wants to be an emperor because there are so many operas and TV series and soap opera shows recently about how the past generations, how the ancient times, how the emperor times. Yeah, everyone wants to be an emperor. So texting a message, gets the people done your job. Everyone wants it. So as long as you make the product simple enough and convey strong similarity with being emperor by sending messages through WeChat, really suits people’s taste. Yeah, I think this is two big factors about culture and society reasons here.Grace Shao (10:18)It’s interesting because basically they’re saying involution itself needs you and has made everyone want to adopt a new technology faster. It’s something I’ve never thought about. On the second point, am curious, like jokes aside about the Emperor thing. What is it that like the average person, like what are they using OpenClaw for though?Shuyu Zhang (10:36)Okay, actually there are two big categories. The first category is still for the common people because they don’t, even though they know OpenClaw is wild, they don’t know what to use it about. So they still use it like Yuanbao or Doubao or the other phone. They just like asking, yes, yes, yes. They still use it like the chat bots. the second part, they already use it in a more advanced way. They used it to earn money.Grace Shao (10:53)which are chat bot products. Yeah.Shuyu Zhang (11:04)Like for example, ask QClaw to seek jobs for them, like scanning the boss or the Liepin website and apply for the job.Grace Shao (11:12)Which are LinkedIn, Chinese LinkedIn, Craigslist, or indeed kind of like websites. Yeah. Okay.Shuyu Zhang (11:18)Yes, yes. And the second part, they use it to operate the social media account like Red Note or they even use it to operate X account, get some posts from X and watch it to write on other social media platform. Yeah. And the third condition is that they actually, trying to use it for like investing suggestions, how to invest in some stocks, what’s the price to get in and should they keep it or sell it.Yeah, these are the main categories, but they are all about making money.Grace Shao (11:49)I see. That’s what fascinating. I want to get into case studies a bit more later. But before we get into this further, I want to help our listeners understand, can we, provide a base framework? Right now there are five claw bought like products within just Tencent ecosystem. And Qclaw is one of them, right? Which falls under WeChat, the product WeChat. Can you help us understand like these products first?Shuyu Zhang (12:13)Yeah, sure. The five claw product here is different for the users and are different between the target users. First, Qclaw and also WorkBuddy, we’re targeting at consumer and the Lighthouse, they’re targeting at enterprise side needs. And also there is a product called Claw Pro. They’re targeting at enterprise for the enterprise who want to make their stuff. Everyone has an ⁓ enterprise size claw and also the cloud desktop they’re targeting also at the enterprise side. Yeah.Grace Shao (12:44)I see that that’s just a good framework to have. So, okay, let’s get into the product pieces and your strategic intent then. OpenClaw is the open source framework ecosystem, while QClaw is Tencent’s package localized layer built on top of it. Can you actually help us understand how that works? What does it mean to have an OpenClaw integrated into a Tencent ecosystem?Shuyu Zhang (13:04)Okay, sure. Okay. So open cloud is actually a package of codes. If you install them on your laptop and connect it with LLM APIs, you can have a personal assistant already on your desktop. But that were required to handle like command lines, which is very technical skill. Even though I’m a product manager, I don’t know how to run command lines before after my engineers taught me to. Yeah. So purchasing and also purchasing APIs from providers is also not familiar for common people.and also connecting with WeChat channels or like the other channels is also not that so easy before we do it there. OK, so we made all of these coding execution into visible and simple product features, which are already educated to common people. For example, we made the channel connection by making it just scanning a QR code and you can already get a QR code onboard on WeChat.Grace Shao (13:33)Mm-hmm.Shuyu Zhang (13:59)And also we make all of the LLM API purchasing processing invisible. We don’t need them to purchase a game. We are reincarnate in the product. And also the installation part in the past, people might need to, know, NPM run open claw, but now they only need to download the applications and double click it’s on.Grace Shao (14:19)see that that’s really helpful for people who don’t understand technologies, understand how this works, including myself even, I was a bit troubled. So I want to understand the thinking behind the QClaw product, right? You kind of talked about it, the frankly, the more cultural aspect of like how this came about and your own personal mission. What was the business decision really for WeChat? Like why did WeChat push out QClaw?Shuyu Zhang (14:43)Okay, so the thinking behind QClaw, how did that come out? Actually, aside from the thinking that I want to build a product that is easy enough for common people to use it, I still have the following thinking. The first is, ⁓ what’s the vibe of the product we should use? Is it work or life or both? Because the chosen, the choice of the vibe will be different for different people. How do we categorize that the work it can do for us? Because in the past, I think all of the AI products categorization are hard to get because most of time you just categorize, for example, something like finance or.⁓ work usage or something like information gathering. my God, who knows that? So, and also there are a lot of times that work and life are mingled together. So if you’re designing an agent or a product features for everyone, if you’re trying to do that, that’d be hard. For example, if I want to build a finance agent, common people might just want to, you know, like search the stock price of something for me.or recommend whether I buy or sell. But if for the professional users, their requirements will be higher. For example, they want, they wanted to, for example, write a quantitative trading strategy for me or something like that. Actually, the depth of capabilities providing would be hard to define here. And also it may discourage users if not handled properly because ror example, if for the pro users you’re designing it too easy, they would think, this is useless. This is far from replacing my interns or something like that. And if it’s designed to be too complicated, the common users would think, my God, this problem is not just designed for me. I don’t deserve to use that. And people are born to have fearness towards their unfamiliar domains. So I think technology or products shouldundermine these fears or lower the barriers here. So in that way, we actually divide the categories of QClaw in three ways, which will be in international version. We categorize in three ways. First, QClaw it up. For the things you don’t want to do, but I have to let QClaw do it for me. That’s QClaw it. And also,QClaw daily for the things I need to do every day, but I don’t want to forget a break. And the third QClaw up for the things I can achieve by myself and the expertise support. We divide the categories in these three ways. So everyone would have the it daily and up requirements. And we will also be more flexible or more concentrated focus on what types and what level of capabilities we’re providing.This is the first thinking behind that. Yeah, because I think that the categories are complex and I want to make it simple and direct and focused to the people’s needs. They will know this, oh, this is for me. This is not too hard or something I don’t deserve. This is something I deserve to use and it would really help me. And the second thoughts behind that is, it’s a line we draw in selecting the building clause for agent.since we already raised a lot of full grown claw. Because in China, a lot of people find it hard to raise a claw. They have to educate it. They have to do a lot of configs. So it’s hard for them to raise. But we...Grace Shao (18:05)Sorry, one moment. Explain thecontext of what raising a claw mean. Like in Chinese right now, the buzzword is 養龍蝦. Explain to people what that means.Shuyu Zhang (18:14)Yes. So raising the club, a lot of people for common people, they were thinking like feeding all of my knowledge, what I know, who I am, what I want, what I like to it. They take into your input and they know what to generate in this mind. And also there is a mechanism called dream during the dream. They were, they were rethink about everything you, you told it today and they would generate something called memory.And in the later usage, they will use this memory to know you better. So you will know that after daily’s inputs, after daily’s talk with it, your claw will know you better. And every instructions you give it will be better than the common AI products that don’t know you. That is called Yang Longxia or raising the claw.Grace Shao (18:59)It’s so funny. It’sbasically providing the technology, the context, but then when it gets better, better people say they’re like lobsters are growing, growing. It’s a funny analogy. I don’t know how it caught up, but it’s hilarious. Yeah. But yes, please continue. Thank you for that context.Shuyu Zhang (19:12)Yeah, sure.Yeah, yeah. And I also think Yang Longxia or raising the claw is interesting because it’s like raising a kid or raising a pet. Because in the past, I think everyone still remembers there is a product called QQ Pet on QQ of Tencent. And I think back in my days, I was like seven to nine years old. I also raised a pet by myself, even though it died twice. Yeah.Yeah, I really enjoyed raising a pet, like feeding it every day or just bring it up on the website and see it on my desktop every day. I think that is interesting for me. And I think a lot of people would want that this kind of companion of AI and they also enjoy the feeling that something is getting smarter or clever because of them. There is connections between them and the AI. They will bring a great joyness here.not just something, for example, something that is already very supreme, high end package well and bring it to you and you just use it. You don’t feel connections with it. I think this is a very different feelings here. OK, so keep going with my point. ⁓Grace Shao (20:22)No, I think it’s funnybecause I do think you touch on something basically like how Chinese tech companies gamified as well. So that’s how it also helped the mass market adoption that we were talking about earlier. And you just reminded me like when I was young, we all had Neil pets. think any millennial people in the West would know that. And like, even though they’re a virtual pet, you actually had a strong emotional connection with it. So I kind of see what you mean by this whereA lot of people might have not even found a purpose or use case originally, but even just building that context, that relationship with the AI that actually helped, you know, adoption rate. then sooner or later you try to find ways to make it more useful. Right. But yeah, please continue.Shuyu Zhang (21:03)Yes.Yeah, exactly. OK, so after I saw that trend and also I saw the problem here, I think if we’re giving them some ground-claw, we should bring incremental value or incremental user experience to these people. And because this is important for the product perception, because if we’re the same like the chatbot about what we can do, people wouldn’t think or people wouldn’t take it seriously.people will still think, okay, this is just another chatbot, but I want to bring incremental value to them. For example, before, in the past, when people want to make some travel plans using AI chatbot, they can only say like, I’m going to like Shenzhen for three days trip. Can you design a trip for me? Because in the past, like even though AI might be different in every answer, there are 80 % of the answers are similar.or told you like to go to some park or some supermarket or some something like that or the hotels to say but if you’re using claw it would told you okay based on your fondness based on your habits based on the things you told me before i think some blah blah blah hotels would be better for you and some restaurant which is closer for example to your hometown or the flavor is similar to what you have told me that you likeThis would be the incremental value. also the claw can also like book the flights, book the hotels, or just, you know, make transactions with the restaurants and ask for example, I’m going to spend my birthday there. Can you arrange something for me? All of the things is incremental value of QClaw AI can bring to people compared with the common air products. And I think this is the second thought behind it.I want to do something that brings incremental value. Yeah. And I think these are the considerations here.Grace Shao (22:48)So I think one thing that’s quite interesting is Qclaw is obviously built out by the Tencent team. However, it can be accessed not only through your own WeChat Wecom, which is the WeChat Enterprise product. It can also be accessed through ByteDance Lark, which is like the Slack product within ByteDance. What is the thinking behind that? Why did you open up to your competitors essentially?Shuyu Zhang (23:06)Okay, so these are all channels, channels where people are already living in. We want QQL to keep company with users either in life or work or any interface. So limiting any channel will bring inconvenience to users. For us, the user experience is ultimate mode. So we don’t really care about the other, you know, so-called the business consideration. I think the user experience is the most important thing for us. Yeah.Grace Shao (23:33)That’s a very WeChat answer. feel like Alan Zhang has been kind of known to always prioritize user experience over any other kind of thinking, whether it’s commercialization or even, you know, sometimes functional adjacency within other products within the Tense Umbrella. So interesting. OK, so I have more questions. Something we talked about prior to recording was that you said WeChat is the default entry point for QClaw.And that’s still a huge mode or advantage for you guys, especially certain features that you guys introduced, such as like scanning the QR code for downloading or installing a claw has been a big selling point. Walk us through kind of why is that and why WeChat, QCOP being built within WeChat is something so powerful.Shuyu Zhang (24:22)Okay, so actually there is a very simple reason before that and after that I will explain a more complicated one. The very first simple reason is that during the initial launch, we only had five engineers and me in the very first place. And we started to develop this product after spring festival, but we launched it in March 9th, I remember. So the time is very limited, but there are so many things to be productized.of OpenClaw. What is the choice here? Because I want to make it simple. I want to make it user friendly. I want to make it widespread. So the first two things need to be adjusted by us. We need to do deletions. And the third part, we need to do adding items. Because before that, actually WeChat is not supported by OpenClaw officially. So we justwatch all of the files on the WeChat open platform and found out that there is actually a way to connect WeChat to OpenClaw. Then we just implemented it. And actually we don’t have more time to do that. So I think, okay, if I only have one time to make a decision or I only have one chance to select the channels, what would I choose? Of course I would choose WeChat because not everyone use some other working messaging applications.But almost everyone in China use WeChat, even though whether you are like four or five years old even, or you’re 70, 80 years old, everyone use that. so that is why I choose it for the first channels. And the second reason is that people are mostly adjusted to use WeChat. And WeChat can be also connected through scanning QR code, which is the user experience other products can give them.because other products, you see it on the light, like the tutorial, you need to go to the open platform of that product. need to copy your user ID, which is a very long link, and you need to copy the token or pin or something that is so technical terms. People would get scared by that. But we already have a very simple user experience, ⁓ user interface method that is getting QR code.Why don’t we just do that? And also, also through the past years, mobile payment through scanning QR code is also widespread by Tencent. Tencent made scanning QR code and making payments widespread in China. And everyone, either they’re like the merchandising, the business, big business or small business, they can have their own QR code. And everyone isused to scanning the QR code and connect everything. That is why we choose WeChat as the default selection.Grace Shao (27:04)I think that’s interesting because it’s like basically what headlines been missing. A lot of it is also just the native user experience and the ease of people to even access this kind of new technology because I think like you said, a lot of people are not maybe not scared, but it’s intimidating, right? It’s intimidating to try out new things and a lot of what’s out there in the market right now feels very technical. And if you’re not a technical person, you’re not following the progression of AI like closely. It’sIt feels intimidating to even try these new products out. Yeah, so I want to bring the conversation back to the real life usage, right? Because that’s a thread we kind of been talking about throughout our conversation. You really focus on bringing AI agents to the average show, the average person to the mass market. What is it that people really, really want out of this? Is it?purely for gimmicky use, it’s for fun. You know, the case that he brought up with the young girl is obviously very interesting, but I would assume that’s not the mass demographic, right? Is it for consumer convenience, prosumer productivity? Is it for small medium sized businesses automation? I wanted to understand that. And then I want to expand beyond that, which is, are they not concerned at all about safety or privacy when they’re using these products?Shuyu Zhang (28:22)⁓ actually I think, yeah. Okay. So, ⁓ the, the first question I’ll explain here is what is a real target use case here? Either it’s consumer convenience, prosumer productivity or small business office of my automation there, right?Grace Shao (28:23)So two parts.Shuyu Zhang (28:36)Okay, so for me, it’s still consumer convenience, or we don’t categorize in this way, because we look at people as the subject, what the people need to do and how we can make it smooth and convenient. People can have different requirements in different conditions, either it’s on life or it’s on productivity, or some more serious or related to the business. We make integrations and push the ecosystem to provide the rest.We will, like for example, if you see QClaw international version, you will see already put like some, I mentioned before the QClaw and QClaw Daily about for example, either you’re seeking jobs or you are operating your ex account or you are, for example, your career pop fan and you want to search for the concert tickets or chasing all of the information behind the hero. This is the conditions we provide. And we will also push the ecosystem.For example, we already have a lot of ecosystem supporters here who provide the doc intelligence, like providing, for example, scanning contracts, the receipts, scanning something like that. And also there are ecosystem friends who already provide video generation, something relevant to, for example, creative parts or the designer shop. Yeah, we have a lot of ecosystem friends.providing here. So we are still starting from what the people needs in aggregate, what people needs aggregately. so this will, the second question about the safety problem here is that we actually provide an incarnate safety features called the AI gateway or in Chinese Longxia Guanjia. That is something relevant to our team because my team is a Tencent PC manager.which is a very, very, very old product. I think many people who knows this product might be like 30 or 40 or even older. Yeah, this product was built in 2004. remember, might not be correct. So it’s still in the computer-sized format, but in the past, it actually used a lot of safety capabilities.Either it’s like preventing prompt injection or skill security or a lot of file security. It already has a lot of capabilities in it. And it’s also vetting the possible cyber attack over the internet. So we will know what’s the risks here. So we already provide a gateway in the product that everyone can be protected under this gateway.they will not be exposed to the risks on the internet already. And also there’s point that, yeah, and actually.Grace Shao (31:18)I see. that actually protectsthem more than someone directly installing an open claw themselves, right? Like going through cue claw is a lot safer in that sense.Shuyu Zhang (31:28)pardon?Grace Shao (31:29)So it’s basically a lot safer for the average person to use Qclaw than installing their own open claw on a Mac Mini or whatever. Because there’s not that kind of safety guard rail built around it, is what I’m trying to say, yeah.Shuyu Zhang (31:37)Yes, yes.Yes, yes, because common people actually initially if they’re as long as they’re on the internet, they’re exposed to these risks. That is worse when they’re using open cloth without any protection. ButGrace Shao (31:49)Mm-hmm. Yes.Shuyu Zhang (31:55)There’s also interesting part is your computer actually don’t have a lot of information and you don’t have a lot of money. So you’re not actually a target. Yeah. But we provide enough to safeguard for these people. And for the more important issues or more severe issues, part of this will provide higher levels of security. Yeah.Grace Shao (32:04)Yeah.I see, Yeah, I think, you know, that’s a really good big picture on just QClaw’s build out and why China really took on like QClaw at a mass scale. I want to understand your business model and long-term implications. Obviously, I understand, you know, you guys fall under Tencent. It’s extremely lucrative business in the cloud side, the gaming side, obviously, we chat advertisement, etc.You guys might not face the pressure to make money from this product, but I still want to understand how does it work? Will QClaw be free basically forever? Will you guys introduce a subscription model? know, I know recently you’ve even talked about you’ve been traveling around the world a lot. You’re in the States. You’ve been in Europe for some time. Are you trying to go global with this product? Are you trying to sell globally? Is it to enterprise? What is the kind of business behind thinking behind this?Shuyu Zhang (33:08)Okay, so the firstly, it will not be free forever, but we will always provide some free tokens for the first time users because they deserve to know what it can bring you, what increment value it can bring you. So we will provide some free tokens here, but we also provide different subscription plan here to support different layers of requirements. But I don’t think the token fee will be the only monetization methods here if we bring people’s whole life here, just like the...WeChat strategy as well. Yeah, because selling tokens, tokens currently is really expensive. Even it’s a huge company, will still face the pressure here. Yeah, but a lot other modernization methods here, but we will also be very cautious here to trading between the experience and the modernization here. And this is theGrace Shao (33:46)Mm-hmm.Shuyu Zhang (33:57)the answer about the subscription plan and the charging plan. And also we are traveling abroad. We are trying to go in global. And this month, exactly April, we’ll be launching internationally and we will be starting from some main regions, North America, of course, and Asia Pacific. And then we were spending into more areas. Why? Because I traveled a lot in the last years. Everywhere I went, I would stay there for like a month.And I would thoroughly experience the real life there and talk to the people there. I found out that even though there are differences between people’s life, of course, but people everywhere share similarities in needs, and they also have curiosity towards others. So people are bringing a better way to live with themselves with QQLO, like QQLO A, QQLO WAP, and QQLO Daily. So they can share it with others.on Qthaw and benefit from other people’s sharing. That is why we are bringing it globally. And every states we go, we will co-create with people there and pass on merit to more areas.Grace Shao (35:00)Yeah, so I think it’s really interesting because obviously QQLA has a huge advantage in China because it leans into the WeChat ecosystem we talked about a lot today. But what is your advantage when you are going global? How do you compete with international competitors?Shuyu Zhang (35:16)Okay. I think the first important part is still the use experience because I use a lot of global applications because I, my job is AI product manager. I would find a lot of product even complicated for me. I think the product experience is not, hasn’t been done very good yet. There are still a lot of bugs. There are still a lot of, you know, complex, complex items, complex terms or complex workflows here.I want to make it all simple. And integration would also be a great part here. And the third part is the community advantage. would, you know, because we are not so intimidating in the image, we will not be like, we are the tech guys. We are the advanced ones. If you don’t know it, you’re the stupid one. You’re going to learn from us. No, we will not do it like that. Yeah. Yeah. We would co-work with the creators. For example,Grace Shao (36:05)YouShuyu Zhang (36:11)⁓ For example, when we’re spending to Japan, will co-create with some local, like fan, big fans there, or the ones who knows well about the food there, or the one who knows about the job marketing there. And we will make it more localized and make the people who really use it create that. We will co-create with them and bring it.to their community because that would be the way that the community gets the concept or gets the usage of AI in fastest way. Yeah, I think so. basically.Grace Shao (36:44)Yeah, I see what you mean. Like you marketed a much more accessible product than other peers maybe on the market right now, which are targeted for again, relatively niche demographic. So that brings me to the next question, which is like, do you think then the messaging apps with a chatbot like kind of interface will become, will remain the default or will we see a new kind of interface layer for agentic AI and how we call on them.Shuyu Zhang (37:16)think, actually, I don’t know about this answer because messaging apps can be, they can evolve as well because they have a lot of engineers as well and they have a lot of ⁓ intelligent people there and they care about the user experience here and Asian products can also evolve. The true interface layer, I think is dynamic and there is no fall or lamb in the current arena. So I actually don’t know the real answer of the, you know, who are the final interface layer, who is the true one? I don’t have the answer yet because I see a lot of interesting apps evolving from AI agents, but they’re trying to, you know, cutting into the messaging apps interface. And also there are a lot of messaging apps there. for someone like the X or I don’t know what’s his future plan, but I also know there are a lot of messaging apps who are cutting to an agent domain. the answer is dynamic, I think.Grace Shao (38:13)I appreciate the honesty and the humility actually. I think no one really knows the future right now. The speed of evolution of industry is insane right now. And I always hear people who really like, you know, like even the Ben Evans and the world, they’re like saying, if you think you really know the industry, you don’t really know the industry because you can’t possibly, you know, have a strong grip on what’s happening because things are changing so fast and so much happening constantly.Grace Shao (39:05)I have the last question, which is a question I ask every single guest that come on the pod. What is one differentiative view you hold? This can be anything about the product we talked about today. It could be about the industry. It could be about anything in life.Shuyu Zhang (39:08)Yeah.Okay, one differentiated view I hold is that I think the most profound function of a superior AI, like the clock, is not to solve problems, but to reveal them, to reflect back to us the questions we’ve been unwilling to ask ourselves. Think of it as an archeologist of behavior. We narrate our lives, edit our memories, even lie to ourselves without knowing. But something like the clock observeswhat we actually do, what we choose or what we linger on. It doesn’t judge. It mirrors. In the end, what it shows us isn’t its own intelligence. It’s us, our contradictions, our desires, the fractures in our collective consciousness. So the most meaningful conversation isn’t whether AI is becoming human, but whether we are brave enough to look clearly into this vast, mirror it holds up. And finally, see...ourselves.Grace Shao (40:18)I think that’s super interesting think it helps us. You touch on something, it’s like really helping us recognize blind spots. I just want to share a personal anecdote as well, which is like, think, you know, when I started AI Pro nearly two years ago, it was really hard for me to sometimes seek help from other people’s opinion because it takes people’s time, right? And then for reviewing of my work and I didn’t simply want a grammar.like, you know, copy editor kind of grammar fix. So now what I did was I built an editor council and what I did is train the council basically to have certain perspectives, follow certain guidelines or, you know, ⁓ angles of the world and then critique my work and really help me recognize ⁓ blind spots I’ve been missing or my logic or my thinking that are not, you know, synthesized clearly or not.flowing smoothly, things like that, that I just found it so helpful. In fact, in some ways more helpful than a human editor at certain tasks, because exactly to your point, humans have biases, humans have judgment, and not like intentionally, but just by nature, we all hold biases based on our own knowledge, whatever. But the machines basically are just can be very critical if you tell it to be critical and can kind of show you a 360 view of your thinking. that’s super interesting and I appreciate your sharing on that.Okay, thank you so much for your time today, Shuyu. If anyone wants to reach out to you, how should they find you? If they want to learn more about QClaw, where should they go?Shuyu Zhang (41:50)my LinkedIn. The name is Shu Yuzhang. And they can also follow me on my X account, which I can share with you later. It’s also called Shu Yuzhang. Okay.Grace Shao (41:59)Perfect. Thank you so much. We wonderful conversation with you. Thanks again.Shuyu Zhang (42:03)Thank you, thank you Grace. Bye.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Sovereign AI, Open Source, and the Gulf’s Big Bet with Interconnected Kevin Xu 14.01.2026 58min

  Every panel on AI and geopolitics seems to default to the same cliché: “the US–China race.” In this episode of Differential Understanding, I wanted to sit with someone who has actually lived inside DC, Silicon Valley, and the US–China tech corridor, and ask whether that framing still makes sense.My guest is Kevin Xu, founder of Interconnected Capital – a global hedge fund focused on the picks and shovels of AI – and author of the Interconnected newsletter, which sits at the intersection of tech, business, and geopolitics. Kevin’s path runs from Obama campaign staffer and White House / Commerce Department comms to GitHub’s international expansion lead, and now to full-time investor–writer with a very explicit geopolitical lens.We start with why he insists on “thinking in public” as an investor, and why he believes ideas soulocking in a vault. From there, we dive into his critique of the “race” narrative and his alternative concept of US–China co-opetition – a messy mix of competition, cooperation, and outright co-opting of each other’s models and research. That leads naturally into China’s open-source AI ecosystem, the Manus–Meta deal, and what he would need to see before feeling comfortable owning the upcoming MiniMax and Zhipu IPOs in Hong Kong.In the second half, we zoom out to sovereign AI: why South Korea might be one of the few countries outside the US and China with a shot at true full-stack AI sovereignty; how to read OpenAI’s Stargate initiative as an explicit American export play; and why the Gulf – particularly the UAE – is emerging as an AI “swing vote”, combining abundant energy, sovereign wealth, and a 1.5 million-strong construction workforce into a potential global compute hub. We close with Kevin’s differentiated view on China: AI diffusion is far more visible there, but the economic impact is not necessarily greater, and Beijing may end up being the first government forced to confront AI’s social implications.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Timestamps (chapters):* 00:57 – From DC to GitHub to Interconnected Capital* 02:39 – Why Kevin “thinks in public” and writes Interconnected* 04:44 – US–China AI is not a “race”: co-opetition explained* 09:27 – Open-source / open-weight AI as last bastion of global cooperation* 12:14 – Capital flows, decoupling and why “capital finds a way”* 15:36 – Manus x Meta: product quality, viral growth, rationality in AI* 19:40 – MiniMax & Zhipu IPOs: revenue reality vs AI lab hype* 24:39 – Can Chinese labs win the Global South with cheaper AI?* 27:09 – Sovereign AI 101 and why South Korea looks uniquely powerful* 32:03 – Stargate as de-facto US sovereign AI and export strategy* 34:44 – Kevin’s trip to UAE: Gulf AI strategies and the “swing vote” thesis* 40:06 – Sovereign funds, MGX, and attracting talent from Hong Kong & beyond* 44:18 – Non-consensus bet: UAE Stargate as a global compute hub* 46:57 – Differentiated views on China AI diffusion and economic impact* 51:29 – Embodied AI, “aunties pressing elevator buttons” and social risk* 55:01 – Robotaxis, delivery drivers, and why China may go slower than expectedAI-generated transcriptGrace Shao (00:00)Hey everyone, welcome back to another episode of Differential Understanding. This is your host, Grace Shao And joining me today is Kevin XuKevin Xu is the founder of Interconnected Capital, a global hedge fund focused on the picks and shovels of AI. He writes the Interconnected newsletter on SubSac, which covers tech, business, and geopolitics.His insights have been frequently cited by the New York Times, Bloomberg, Economist, CNBC Information, Financial Times, Wall Street Journal, among many other media outlets. He previously worked as a senior executive at GitHub, the world’s largest developer platform, and served in the White House and Commerce Department during the Obama administration. He studied international relations at Brown University and law and computer science at Stanford.Grace Shao (00:40) Hey Kevin, thank you so much for joining us today. I already introduced you, but for listeners who may not know you that well, can you introduce yourself, the different hats you wear today, running Interconnected Capital, writing Interconnected Newsletter, and operating at the intersection of US, Asia, tech, geopolitics, and investing.Kevin Xu (00:57) Yeah, first of all, thank you for having me. So as you mentioned, what I do currently during the day is I write the interconnected newsletter on the intersections of geopolitics, technology, and business. I also run my own long only fund called Interconnected Capital, focused on the picks and shovels of AI, both hardware and software. Prior to that, I actually work as an operator inside multiple Silicon Valley tech startups. The most recent one is GitHub, which is the Microsoft-owned developerplatform. I was their lead for international expansion strategy. That was my most recent real job, if you will. I also spent a bunch of time at different startups of varying sizes. And before that, actually started my career in politics. So I joined the first Obama administration’s campaign back in 2008. I was a campaign staffer. That was my first job out of college and then moved with the campaign team to DC, ⁓ worked in a few different roles.in the Commerce Department as well as the White House doing mostly press and communications work. So that is ⁓ my sort of all over the place background that led me to what I’m doing today, which is investing and writing ⁓ in technology, but with a very heavy geopolitical lens to the process.Grace Shao (02:13) I think that’s really interesting and explains to why you have a geopolitical lens, given that you actually have a DC background, right? But you run a fund and you actually keep most of your thinking, public, So instead of just keeping it mostly all private, which is what most investors do, why do you publish a very opinionated, very insightful sub stack and, share it very, I think generously with the public?What kind of conversation gap are you trying to fill when you start the interconnected newsletter?Kevin Xu (02:39) I think there are two elements to that. First is that this is more of my Silicon Valley ethos, which is that no idea is that worth keeping in secret. It’s all about the execution. Like I’m not, I didn’t come from a Wall Street finance background, right? Where a proprietary trading algorithm or some secret information you got from a meeting is this big trade secret that you want to lock into a vault inside Goldman Sachs or whatever. And that’s going to make you billions of dollars. That is not my approach to investing.I think thinking in public, sharing in public, and really getting the feedback that I get from writing is much more valuable than keeping all these thoughts in my head as if they’re the next best thing since sliced bread. When you actually write it down, when you put it out into the internet, half of them are good, half of them are actually crap. And I use writing and writing in public specifically basically as a canvas for meto think better, to hold my thoughts more clearly. I think knowing how to think is the most important skill for any investor to be able to succeed for the long term. And if any idea that I shared out there benefits somebody else, and you made some money off of it for free, so be it. Good for you that you actually understood some of the value from the writing, even perhaps more than I did as the writer. But for me, that’s not something that I keep very possessively as a trade secret.Grace Shao (04:00) I really relate to that and I think exactly to your point you’re like writing everything down is the way of thinking through your thoughts sometimes it’s all jumbled up in there and then also people ask me why do you keep AI Pro all free? I was like well if it really benefits you you know I don’t really mind I’m not trying to make money off of like selling you know just my content but really the content is my thinking and to your point sometimes I put in so much work and then the result and feedback is so bad and then some things I just kind of like throw out thereAnd then it actually really sticks with people you never know. It’s really good to get the feedback from the public as well. Well, I think now I want to ask you about your journey into investing and really covering China and US. So you are based in the US, but you are Chinese.birth, right? ⁓ So does that play into why you cover China-U.S. related work right now? And I do want to talk about your recent article, which you said you think calling the China-U.S. relationship in tech and AI a race is quite lazy. You said instead of seeing it as pure competition, you think it’s more of a competition. So cooperation plus competition. Do walk us through that and how you kind of came out with this frame.Kevin Xu (04:44) Correct.So just to put a finer print on it, as far as the personal history is concerned, I was born in China. I moved to Canada when I was little, similar to you, think, Grace. And I moved to US later on. So obviously, I work in the US government. So I’m a US citizen. So I’m actually a card-carrying Canadian as well as an American right now. And I think having had a very global citizen-ishGrace Shao (05:22) Mm-hmm.Kevin Xu (05:30) upbringing and life experience was the lens that wanted to bring to my newsletter when I first started writing it roughly five and a half, six years ago. Some of it has to do with the US China. Some of it actually just has to do with a specific industry trend in the software, in hardware, in the nerdy techie stuff, in open source that I like to talk about. And I think only the US China stuff got picked up for one reason or another. People started to pay more attention toGrace Shao (05:31) Mm-hmm.Kevin Xu (05:58) tothe stuff they’re writing when there is a China-US lens. And maybe it’s just because there’s a dearth of content out there that actually brings a level of nuance to the conversation. And that brings to what you asked me about, which is this notion of a US-China AI co-opetition, is how I like to call it, as opposed to calling it a race, which is the kind of intellectually lazy approach that I have fallen into multiple times.Throughout my own writing just calling the race calling it a race But not really thinking what that implies which is that one a race implies that there is an end point There is a finish line to this race But for AI there really isn’tEven the most fervent believer of what an AGI is does not believe that is a static endpoint in which once you reach it, you’re done. And of course, there is the implication that this whole thing is a very zero sum dynamic if you call it a race. But in reality, if you look at everything that’s actually happening on the ground between the US and China on AI, it is a manifestation of co-op petition, which is that there’s a lot of competition between different firms.between different labs, both within China and between the US and China, lots of startups. There is also a lot of cooperation. The cooperation stuff gets probably shoved over to the side or doesn’t get mentioned as much because of the geopolitical toxicity of the conversation. But there are lots of papers, academic institutions, adjoined productions in terms of research and collaboration that is still happening both in academia and frankly in a lot of startups where the approach to all this is much more pragmaticandless geopolitical. And then the last element I actually want to introduce to this fake word is co-opting. There’s a lot of co-opting between leading AI labs from both sides. When the initial chat GBT moment happened three years ago, every single Chinese lab more or less used Lama as their basic building block.to advance their ⁓ model building. Every single hyperseal in China used Lama as one of their leading cloud services to get things going, right? That is a co-opting of an American, I guess, production, if you will, of a model, just to use model as an example. And then as Chinese open source became much more well-known, much more prevalent, much more popular from B-seq to Quinn to whatever, now we have Airbnb being one of the biggest users of Quinn.We have a UiPath being one of the biggest users of Quinn and a bunch of startups that they don’t want to talk about using Chinese models to really bring down their own costs so they can run a profitable startup, co-opting each other’s work. So I think co-op petition is the most accurate way to talk about it, but I also understand it’s probably not the easiest way to say the word. And so I’m not...counting on the word catching on at all, but at least for my own intellectual honesty sake, that is the word or the way that I plan to talk about this dynamic going forward because I think it’s the most accurate way to reflect reality on theGrace Shao (08:56) I think definitely your writing is one of the more nuanced kind of work that I’ve come across on the internet where it does touch on China, US, where it talks about the cooperation as well as competition and give the audience a geopolitical background ⁓ but still focus on the business, the society and offer that more neutral un biased, I think, analysis of the businesses,But from where you sit, where do you think the founders, engineers, investors actually feel like they are really collaborating? Give me some more concrete examples.Kevin Xu (09:27) I think open source AI, open weight AI, the rise of that is probably the best and most concrete example of collaboration and cooperation happening despite all the resistance, the challenges to cooperating, right? There is a lot of resistance to cooperating on anything. And the natural way is to kind of go towards the path of least resistance. But something that is happening that is, I think, probably the biggestin 2025 is the rise of China’s open AI ecosystem becoming all of a sudden leading the world. Not just pretty good, not just, oh, it’s also happening, but is flooding the zone as far as models are concerned. And the nature of open source is open collaboration. There is no deep-seek open model.without the lineage of all the innovation that came out of GPT-2 that was actually open source back in the days, or Lama, or whatever the open things that the US lab...was feeling comfortable doing until it no longer felt comfortable doing. And then you need a lot of Chinese labs to give back to the whole ecosystem as well, entirely without charge. That is the other thing about open source is that you can do whatever you want with open source product for the most part. And DeepSeek and Quinn really led the way from not just opening it, but also having the legal license to permitjust proliferation everywhere. You can do anything with a Quinn model. You don’t have to tell Alibaba you’re doing something. You don’t have to really pay Alibaba a cent. You don’t have to even give credit to the Alibaba team. Just kind of go forth and prosper, right? Now it’s very hard to track.⁓ what that diffusion really looks like. Having worked at GitHub, for example, which is the home of all open source code for the entire internet pretty much prior to AI, I know how hard it is to track. We’ve tried to do that internally with our data. We have some rough sense of which country is contributing on GitHub more than other country, which company, but we don’t ever get too deep into the people behind that for privacy reasons and whatnot.But you know from a institution perspective and an intuitive sense that it’s gonna proliferate everywhere, right? And the only surprising thing is that this came out of China, which shocked a lot of people. I don’t know why it should shock a lot of people, but it did. But.But that’s kind of where the big story comes from. So I think cooperation is happening regardless. And open source is probably sort of this last bastion of global collaboration as the world splinters into its own camp as geopolitics and AI kind of co-mingle together to make everything feel more cagey. This is still the last kind of remaining source of cooperation.Grace Shao (12:14) I think you talk about the technology being much more cooperative than people expect or want to admit. But what about capital? Over the years, we’ve seen that. first for context, think people need to understand in the 90s and early 2000s, US capital were the predominant capital that were actually behind a lot of the Chinese big tech we see today. But today, now we know there is a decoupling in terms of US investment into China, especially in the sensitive areas such as AI, robotics, andsemiconductors, right? So do you think this is something structural or cyclical? Like, are we going to see more opening up from the US government to allow these US funds to invest in China again? Because a lot of them are obviously still interested in doing so.Kevin Xu (12:57) I think the rumor is it is loosening up. I think there’s a lot of chatter that Chinese VCs who for a period of time just could not raise any USD fundfor probably like five to six years or so is starting to do so again and I think that spigot is slowly but surely going to open up and it probably won’t be like as wide open as it used to be before but my personal feeling is that capital finds a way it’s just like water it’s going to flow towards whatever the final destination it needs to go to even if it has to go around mountains it has to go through a bunch of rocks it’s going to grind that rock to a smooth edgethat being geopolitics sooner or later, but it will probably take more time than most people have the patience for. And you know, to come back to what I talked about open source real quick, if we can double click on that, I think the contrast between capital flow and source code flow in terms of open source is that ⁓ engineers, doesn’t matter which country you come from, want to work on⁓ the most open piece of software or code that is open source and you can collaborate with the rest of world, right? Like that’s why GitHub became so popular because engineers, whether you’re from China or the US or Germany, you identify with the code that you build. You don’t necessarily identify as much with the nationality that you were born into. That isn’t really a big part of your work at all.Right? Even calling something a Chinese open model is a bit of an anathema because like what is it? What part does it really is Chinese versus when it’s out in the open, it’s just like this piece of common good in the internet now. Right? Like no one can really control it. So what’s the point of calling a Chinese or American or whatever? And that’s how engineers like to operate.So that’s why there’s this tug and pull between the geopolitics force and really the engineering and the builder force that is by definition very global.Grace Shao (14:51) Yeah, I think the engineers and scientists you speak to definitely are not geopolitically driven or as ideology driven as I think sometimes the business people because they need the support of their government for certain policies. So I think the business people who seem to sound geopolitically driven are not actually geopolitically driven. They just need to do so before for their business survival. And that’s just the reality of how the businesses work. Right. So I want to put you on the spot. We touch on this quickly before we start recording.Manus, speaking of the most famous US injection into a Chinese AI company is Manus. And I just woke up to the news, ⁓ day of recording is December 30th, that, you know, Manus was just bought out by MetaI I used to manage this really good product. How do you view this whole thing?Kevin Xu (15:36) I also use Manus. I think I got an early access code actually before it even launched. I was going to say back in the days, but that was only like a few months ago. It was actually like less than a year ago, right? And this company ⁓ went from zero to a hundred million dollar ARR in about eight months, which is just astronomical.Grace Shao (15:40) Yeah.It’s so crazy. Yeah.Kevin Xu (15:57) ⁓ growth on the back of essentially its product quality. And I think that is one of the most interesting takeaway for me as an investor, as a technologist, which is that we talk all about like geopolitics and, you know, this and that none of this is actually about the product or the tool, right? About AI, but Manas, this little bitty startup, basically proved all of us wrong, which is that ⁓ product quality still matters.if you have goodQuality product people will share you people will talk about you people will you know? Do word-of-mouth to tell other people to use your product I think one of their more famous element is their ability to kind of crack this black magic of viral marketing without spending any money Right back in the days when they first shared their first version, you know, Jack Dorsey tweeted about it all these like Silicon Valley Luminary started sharing about it and it’s because their product actually spokefor itself. And it continued to evolve very, very quickly to capture not just attention, but actually revenue, which is very, very hard in this current climate of AI kind of bubble-licious noisiness that we are living through. And on the outcome itself, first of all, congratulations to the entire team. I think it’s very impressive, this outcome to be bought out by Meta. At this moment, we don’t know how much it actually paid for. Maybe by the time this app was released, we actually know how much Metapaid for, but the last round they raised that was $500 million valuation, right? Which in AI land is actually really, really cheap because we have, you know, 10 to $20 billion startups being funded in the U.S. right now that has zero product, zero revenue, and more or less a bet on a very impressive team, which could still come out okay, but we will see what happens. But I think this Manus dealto me is a very I want to say it’s evidence that rationality still matters. It’s evidence that like economic kind of pragmatism still has its moment in the day and doesn’t have to be whiplashed by geopolitical consideration. So I find that the deal very heartwarming as an investor who really just hopes for more economic rationality for everybody who’s involved.Grace Shao (18:18) Yeah, I think ⁓ to your point, like it’s definitely like a positive signal because it means that people are evaluating the products how good they are instead of just the narrative of the geopolitical kind of cloud above it. And I think it’s really interesting. Like when I was speaking to people like about this deal this morning, they’re saying actually, you know, people overestimate the PR they done back in the day when they first released it. It wasn’t because, you know, they did some black magic PR.It was simply because they didn’t even have the compute capability to actually serve too many people. So they sent it to people to try first. And I think I have a lot of startups that come to me being like, how do I achieve madness PR? I was like, it’s not just the PR. The best PR you can possibly do is to have a really, really strong product and have the product speak for itself. Right. So yeah, it’s, think it’s a very interesting time and it’s, and it’s interesting to see probably one of the first Chinese homegrown.company in AI being completely separated from the Chinese market now and operating in the West per se and now being bought up by American US company. Okay, talking about startups, I want to ask your opinion on MiniMax and Zhipu They both submitted their prospectus now. We are expecting them to go public in Hong Kong.What would you need to see before you feel comfortable owning one of these IPOs and how do you evaluate these companies as they go public?Kevin Xu (19:40) So first of all, I am actually a public market investor. So I don’t do any VC at this moment. So I’m very, very interested in how the Zhipu and Minimax listing happen. Even though as a rule, I don’t invest in IPOs because they’re quite frothy and confusing. I’m happy to wait it out. I think there are a couple of signals. And this is actually interesting as a comparison to Manus. If you look at the revenue numbers that Zhipu and Minimax have shared, they’re both in theGrace Shao (19:45) Okay.Kevin Xu (20:08) double digit USD million range, right, which is very modest. And it’s even more modest compared to their losses, which is all in the hundreds of millions of USD as far as how much money they’re losing right now as companies. And you compare that to Manus, which probably is like reasonably profitable at this point as like a hundred million dollar ARR company, not revenue ARR, but still they’re small, they’re growing and they’re probably managing their costs.because they’re not model trainers, right? Like I think Manus was very clear that we don’t build models. We don’t really have expertise in that, but we are very good at wrapping a model into a very good, trusting user experience. But Drupal and Minimax both became or started out as the model makers, which is a very expensive endeavor. So as far as what I look for as an investor is concerned,It’s very, hard. I think a path to profitability, and specifically, think EBIT profitability, so earning before interest in taxes, is going to be key for me to see how does a business ⁓ like this, which has a...I don’t know. I feel like they’re limited to the China market, which is big, but not huge, I would say. And I think MiniMax does have some consumer product similar to ChatGPT, which is going to be how they can maybe justify their higher evaluation, even though most of the revenue comes from serving up their model as a form of APIs, which is a B2B play. How do they balance those two, which are two very different go-to-market motion? It’s going to be interesting.pretty clear path or lane at this point, which is I make my models and I’m very good at serving large, older legacy enterprises and governments that is a very specific type of customer with a very specific taste, if you will. And you have to really orient your whole company to cater to that kind of customer. And Drupal kind of has cornered that market for now, at least. So that could work really well from a profitability perspective over time, even though those are very tough customers.customers to track. But the big takeaway, I think, is that the revenue is still very modest and certainly very modest compared to the large labs that we just sort of talk about willy-nilly in the US, like OpenAI, which is going to have $20 billion.in ARR by the end of this year, probably, right? Like Anthropic is projected to have five, $6 billion in ARR. These are two orders of magnitude larger than Whatchupu and Minimax has shared to the public. But the enthusiasm for investing in AI pure play is still very high in the public market. And I know Hong Kong’s IPO market has been doing very well this year and probably will continue. So that energy can be kepthopefully by these two companies because they actually need the money, right? That goes to what you were mentioning before, which is that I think if we had done this, Chad, in 2018, there will probably be multiple rounds of VC in China with USD backing that are readily available to fund the Gipu and Minimax for maybe two, three more rounds. So they don’t have to go public.Grace Shao (22:53) They need the money.Kevin Xu (23:14) They can still operate as a private company, raise more money, mag around, just like what we have been doing here in the US. But that option has basically run out of this course.right now for any Chinese VC-backed company. So they kind of have to touch the public market earlier in their life cycle for fundraising, which may not be a bad thing for organizational perspective, because you do become a more disciplined, well-run company for the most part, I think, when you become public. But it does expose you also to the public market. But they need the funding clearly, so that’s why they’re doing it.Grace Shao (23:47) Yeah, I actually just interviewed one of the leaders at Zhipu recently for the podcast and he was saying candidly, for them, it’s really about survival at this point because they’ve just run out of money. And if they don’t want to be swallowed by someone else and if there even is a desire to swallow them, because given that, you know, all the BATs we see have very, very strong labs themselves, they don’t really need to acquire a talent, new talent pool. So then there’s no way to keep going unless they go public because they need that money.But on this point on them going public and you know, actually it’s, it coincide with them trying to go global, right? A lot of them, like you said, they’re currently serving China as a market, but they are selling their model as a service to the global South, maybe for a much cheaper price than a lot of the US labs and the peers out there. How do you view that? Do you think that’s something that could potentially work out for them just by selling cheaper services compared to maybe the open AIs of the world?Kevin Xu (24:39) I think it could.Yeah, I think it could. mean, I think USAI, American AI is very expensive. Like the quality may or may not justify the premium, but it’s very expensive, right? Like we have like thousands of dollars.Grace Shao (24:44) Mm.Kevin Xu (24:51) I think the max chat GBD plan is like 200 bucks. People want like $1,000, know, no rate limit, chat GBD plans. And we’re spending a lot of money. And that’s partly goes into these revenue numbers, right? The billions that we’re talking about. Like you can think that is like an inflation almost of AI product costs here in the US for the most part. But we know that Chinese entrepreneurs are very good at reducing costs, right? They’re already released their models because the models are commodities. They’re open source. There’s not a whole lot of value capture really that happens at theGrace Shao (25:14) Hmm.Kevin Xu (25:20) autolayer. And if you can wrap that around with really good services for just throughout random examples of like a city government in Malaysia, right, or a hospital in Thailand, for example, right, these are all the kind of unsexy industries in very unsexy countries when it comes to AI adoption that we don’t ever really think about. But if they have a strategy to go after them, and I do think listing Hong Kong as opposed to on theyou know, Shanghai market, which I think could have done as well. But choosing Hong Kong is very smart because it increases their name recognition, their exposure ⁓ in that part of the world. As much as you and I talk about these companies like everybody should have heard about them eons ago, most people don’t know what these companies are. They don’t know what the differences are. They have no idea. They probably have heard of Chachi PT, but that’s about it. They probably don’t even know what Ethlopiq really is. Right. But if you can reallytap into that capital market and use the public listing as a way to raise your profile for these second tier market and second tier countries, then I think there’s a decent business to be made there.How much will it fetch a premium in the public market? I will never know. But I think that’s been a playbook for a lot of Chinese companies that were shut out from what’s called the premium markets globally, which is the US, Canada, and Western Europe. And they have to go to the so-called global south to make a living. And they’ve been able to make it work. And there’s no reason to just assume that these companies can’t make it work either.Grace Shao (26:40) Yeah.Yeah, for sure. Okay. I want to talk about sovereign AI. You’ve written a lot about sovereign AI and you’ve used South Korea as an example.Why is South Korea a champion basically in the region as for sovereign AI?Kevin Xu (27:09) I think to back up a little bit, sovereign AI is one of these things that ⁓ I’ve been really fascinated with for a better part of this year, ⁓ which is, it’s the first time I’ve seen where a major technological transformational period has beenaggressively embraced by national governments everywhere, right? Part of that has to do with Jason Huang of Nvidia just being the incredibly charismatic salesman that he is, right? Like sovereign AI, he did not come up with the term. I think it came from the EU in 2019 or something, but he really embraced it as the next wave of AI adoption. So more countries can have their own AI, which initially I thought, oh, this is just like a clever sales pitch, you know, to kind of sell more chips. But if you really think ofabout it, ⁓ all these AI models do have a way of encoding culture. Encoding not just your mainstream culture, but also your minority culture, your different languages and whatnot. And the countries have learned, I think, their lesson by being really hands-off during the first wave of the internet and especially social media, but not caring about how does technology impact their domesticsituation, if you will. You can talk about in terms in the context of Arab Springs or, you know, violence in Myanmar or just generally speaking data privacy, social media, all this sort of stuff that countries used to have just by definition a lot of control over by having sovereignty and they’re actually losing sovereignty.to the wave of technology. So with this AI coming together, this wave, more and more countries are actually exerting that notion of sovereignty without really knowing what it means, but they’re exerting it right now more aggressively than ever before. Now I picked on South Korea because sovereignty is kind of this big fuzzy word that means different things to different people. But if you use sovereignty as a proxy to talk about control, South Korea actually has probably one of the betterset of tools to exert more control over their own AI future more than other countries. Because if you really think about full stack AI from top to bottom, from land power chip models and then applications, only the US and China really have a grasp of every single layer of that stack.Grace Shao (29:24) component, yeah.⁓Kevin Xu (29:25) in diff todifferent degree, obviously, but you know that they have control over every single step, right? Every other country for the most part is a customer of one of those stacks coming from the US or coming from China, except I think for a handful of countries, South Korea being one of them because it has a very strong memory.⁓ ecosystem for chip fabrication, not for logical chip, but for memory. And high bandwidth memory is basically an exclusive South Korean national export at this point coming out of SK Hynix and Samsung. Yeah, we have some Micron over here in the US too, but the two thirds of the market is dominated by two Korean players.Grace Shao (29:46) Yeah.Mm-hmm.Kevin Xu (30:03) And then they have their pretty cool little internet ecosystem as well with Naver, with KakaoTalk. They’re all very Korea-centric. They don’t do so well outside of Korea, but inside Korea, just like how we go to China, we have to install WeChat. If you go to Korea, you have to install KakaoTalk. You have to install Naver for your map. Otherwise, you just can’t get anywhere, right? So you kind of have...Grace Shao (30:23) There is Google Map.Yeah.Kevin Xu (30:24) Exactly.So they have that set up cone coming into it. So they actually have a bunch of different good dominant controls nationally throughout all that layer. So when Jensen visited South Korea recently to sign a bunch of deals and allocated a bunch of black wall chips to different major players among these Chibos, I just thought this is like an actual manifestation of a South Korean sovereign AI at play. Now they’re still using American chips, but part of that American chip is fused with South Korea made memory.And that gives them a lot more say, at least, to sovereignty of the AI application that they’re hoping to adopt. And South Korea is just very digitally forward, I think, in general. It’s one of the most digitally connected society, period, of any country in the world. And so ⁓ I think they have a good shot at actually making sovereignty real in the AI era.Grace Shao (31:03) Yeah.It’s interesting because I just went to Korea I think earlier in the year and I was talking to investors on the ground and they were saying that South Korean startups are actually a lot more, again going back to our point, non-geopolitically driven or minded and a lot more agnostic about which kind of, what countries models they use. However, for the country itself right now, the government, they’re still pushing US models forward. And I think it’s really interesting to see to your point likeThey actually have such a small but closed ecosystem in the digital infrastructure. Like everything is with, they don’t use American apps like for social media. They don’t use Chinese apps for social media. They’re actually completely independent. So it would be an interesting kind of case studies to follow through with, I think. When we look at sovereign AI and I look at, know, projects like Stargate, is that something like a de facto US sovereign AI project? Like, how do we understand that?Kevin Xu (32:03) That’s how I understand it. I think Stargate is, first of all, for people who don’t follow this stuff as closely, is this brainchild from OpenAI to build these massive multi-gigawatt data center, not just in the United States anymore, but actually throughout the world, to support its global multi-trillion dollar ambition. And it’s in countries that are willing to be on Team America. So in a way, it’s a sovereign extensionGrace Shao (32:05) Mm-hmm.Kevin Xu (32:31) of American AI in a way is also a reduction of sovereignty in whichever country is willing to receive American AI and be a proxy of American AI, right? And we have a few different sites already announced. We have ones in Argentina, in the UAE, in Norway. I think these are the ones outside the US Stargate projects, maybe perhaps India as well. And that is the most aggressive.expression of American sovereign AI and the most explicit one as well. And the US government is very honest about this as well. Like they want to promote and literally sell the American stack to countries around the world that want to buy American projects. It’s basically like a bigknow, White House driven go to market strategy, right? Where the content of the product is actually open AI, Anthropic, Nvidia chips, Oracle, construction, and all the American kind of major companies that come together into literally a package, right? And then we want to sell that abroad to different countries around the world, including the global South as well. And I think that’s one of the things that a little bit of a shift geopolitically is that the US is no longergiving up the global south as this also ran that it is no longer paying attention to in the way that China has been paying very close attention to for two decades at this point. It’s no longer willing to surrender that part of the world commercially. And Stargate and AI export program is actually a way to express that re-interest in those regions, if we will. And Stargate is just kind of the tip of the iceberg there.Grace Shao (33:56) Mm-hmm.I think to talk about sovereign AI, we have to talk about Middle East and it’s something I really know nothing about. I was really fascinated by your recent article and your series in sovereign AI. So first for listeners, can you tell us about your trip? You just got back from Abu Dhabi, I think a week or two ago, you wrote a really insightful long piece on just how the Middle East is building out their AI strategy. And you talked about it as a region, also kind of breaking it down the whole, looking at the Gulf separately, the UAE, the Saudi Arabia, Qatar, Bahrain, each of the...AI strategies, right? Can you kind of walk us through, first of all, why did you go? What was the event for? And then just some of the high level takeaways from that trip.Kevin Xu (34:44) So I went to that trip from the exact same position that you are now, which is that I’ve never been to the region. I’ve heard a lot of things about the region. Just in the AI conversation alone, we’ve had major announcements and deals being signed by Saudi Arabia, by the UAE, with the United States. We know Chinese tech have been in that region for a very long time as well. There a lot of robot taxi Chinese companies that are deploying their self-driving vehicles on the ground as we speak. So it’s a region that I’ve been really wantingto go if I get the chance to go for a long time. And just by happenstance, I was invited to be part of a delegation among other Washington, D.C. think tankers to go to visit the UAE for a week. So we are an American delegation, right? So that’s important context for you to know. If you were to read the post that I wrote and understand what I was trying to convey and how I learned things, we spent a whole week in both Abu Dhabi and in Dubai meeting with pretty much everybodythat has a hand in its AI future, from government officials to investors to all the funds that you have heard of. And the major takeaway for me was, first of all, just to see stuff on the ground, which is thatThey are very, ⁓ from an AI perspective in particular, just take away the other stuff for now, from an AI perspective, they’re very much in Team America’s camp. They really want to be building UAE Stargate. That’s one of the very few Stargate projects outside the United States that has actually broken ground. Like there are actual buildings that have been built in the desert.⁓ ready to receive NVIDIA BlackWall GPUs if expert control were to be permitted from the US side to let them buy as many as they would like to buy. So that’s number one. I think number two, they are in this very interesting geopolitical position as a very tiny country of 10 million people where they don’t want to actually be Switzerland. They’re not neutral. That was the message that I received from a lot of people that they have a point of view on where they want to be in this globala big game of AI, of geopolitical influence, which is that they can vote for one country or one side, but they also have their opinion to build a society of their own. That’s a very modern Arabic.you know, society, I think there’s a lot of stereotypes to, know, how does it, what is it like to be in a be a woman in the Middle East? What is it like to operate in the Middle East? Lots of cultural stereotypes that they want to debunk. It’s Dubai is one of the most modern cities I’ve ever ever been to. Right. And that is kind of the cultural takeaway that they want us to have. And then lastly, when it comes to this US China conversation, frankly, they’re a little bit tired that they always get mentioned in that context. Right. Every time the US official goes to the UAE is aboutWhat are you doing with China? then, know, presumably when the Chinese official visits, they’re also like, what are you all doing with the Americans? But they want to be seen on their own term. And they certainly have the wealth to do so.as well. So it’s a fascinating region and I think they’re playing both sides very well. I call them the swing vote of the global AI competition. They can swing one way, can swing the other, but they have a lot of leverage in this conversation because they need to have the best of both worlds to feel an economy that is in the desert that literally grows nothing.So they have to export import, sorry, they have to import basically everything from talent, from food, from vegetables, from, you know, the only thing they have is energy coming out of the ground, oil, but they’re trying to diversify away from that, which is the only point where they’re investing all this technology stuff in the first place. And that has been happening for 20 years at this point. So it’s not like a Chad GBT moment thing per se. So a lot of takeaway there, but happy to answer more questions because it’s a trip that I’m still processing, to be honest, because that was first time in the region, had a lot ofcoming at me and I’m trying to still come to terms with ⁓ what I understand now but also what I still don’t understand even though I just went there.Grace Shao (38:40) Yeah, I think that that’s super interesting. And I’ve been really fascinated by the region as well. Actually, we were just talking about this offline. A lot of people in Hong Kong are now being recruited over on the point of talent. And I think, you know, as a lot of these countries have huge sovereign funds, they’re looking for top tier investor talent to go to whether it’s UAE or Saudi or Qatar to really deploy that capital, whether it’s an AI or not. And it’s really interesting kind of to see howYou mentioned they have, what, UAE has 10 million people, but 90 % of that is actually foreign workers, including laborers, as well as knowledge workers. And people kind of forget that actually these are extremely wealthy countries per GB per capita. So I think it’s interesting to hear that they don’t really want to be put in the middle as a China camp or a US camp country now. Similarly to Singapore, where we also talked about, know, like Singapore isTiny small, you know peninsula has actually really made it work from themselves and Pretty much have to import everything from groceries and labor is sometimes from Malaysia and even energy to talent from around the world and now mostly China It’s kind of like in that sense not a Switzerland like Singapore right like you said in your article. I want to understand better actually How do we understand the sovereign funds behind?these investment funds are investing in AI because it actually is so different from private capital in the US and even how Chinese capital is structured.Kevin Xu (40:06) The way I-the sovereign fund in the UAE in particular, know, just that part of the Middle East have not been to Saudi Arabia or anything. Obviously Saudi Arabia is a major, major player as well. So is Qatar, which actually announced their own AI initiative while we were in the UAE as part of the Doha Forum. So there’s a lot of, let’s also call it co-op petition as well, among the Middle Eastern countries as well. Like they’re presented as this sort of monolith sometimes, but there’s actually a lotGrace Shao (40:23) Exactly,Kevin Xu (40:37) of rivalry or friendly competition between, in particular, these three Middle Eastern golf countries, Saudi Arabia, UAE, and Qatar. Now, the UAE sovereign wealth fund in particular, again, we met with everybody there. so their strategic purpose is, of course, to diversify away from oil wealth.Right, but that’s easier said than done. What do you do when you have all this money? From selling oil that you know, it’s gonna run out at some point or you don’t want to be overly dependent on this one source of wealth, right? So Mubarak is their kind of marquee sovereign wealth fund that plays very actively in the world of technology investing and they’ve been Investing for 20 plus years around the world. They’ve had offices in China in South Korea in Brazilobviously in the United States, in Europe for many, many, years. They’ve been placing bets and serving mostly as LPs to local VC firms fora long time. They’ve also bought a global foundry, which is the chip manufacturing plant, similar to TSMC. But you know, that was kind of a spin out out of AMD, I believe, back in the days. So they’ve kind of placed their bet in the chip ecosystem, again, long before AI was a thing. Now, that doesn’t mean they’re all successful, because the diversification justification is very different from ABC, who is motivated to generate the largestfinancial outcome right per fund per fund and they actually did understand more recently why that’s not such a good model which is directly related to your point about Hong Kong professionals finance professionals being recruitedto go to the UAE because they started this new fund called MGX, which is basically more of classic VC fund that has all the incentive structures of a Silicon Valley VC firm like a Sequoia or a 16Z. Mubadala is one of the anchor GPs, but they’re raising money from around the world just like a normal VC would because they need to attract the best talent, which they actually could not if you just run asovereign wealth fund because sovereign wealth fund is kind of like a quasi government institution, right? They’re still kind of government employees at the end of the day. They don’t get a huge carry or a payout because one of their funds hit it out of the park and got less than the NASDAQ. They’re just kind of collecting their paycheck, right? They’re more like a pension fund manager. And that doesn’t get you the best, most hungry, I don’t know, money.making talent from London or Hong Kong or wherever. So they’re just very recently started to restructure that because it’s an evolution of sovereign wealth fund being managed, one, to diversify and then to get into the best technology and then to actually generate a good return and get the best talent, which is really a long-term play because if they can lock down the best talent from Hong Kong for a decade or two to live in Dubai, to live in Abu Dhabi, then that is a long game that they can, again, supplant this 10 million people that needs to be constantly replenished.with better talent and more diversified talent. So the sovereign wealth fund, the game is very complex, I think, and they probably played it better than most people that I’ve seen coming out of any sovereign wealth fund. Singapore sovereign wealth fund is very sophisticated as well, but that took a long time to evolve GIC and Tomasic.Grace Shao (43:49) Yeah.Yes, yes. That’s really interesting context. I think I have one last question for you on Middle East, just given the time constraint, but I would love to talk more about this offline. If you had one non-consensus bet on the Middle East and how it may shape AI globally in the next few years, what would it be? Like, how do we understand the Middle East role going forward, especially amongst this US-China co-petition?Kevin Xu (44:18) I think I was skeptical going into the trip that it’s going to be a region that actually would matter because there so many data centers being built everywhere. But coming out of it, ⁓ I think there is a good chance that the UAE Stargate will house a significant amount of compute for not just that region, but for the entire world.First, because its energy is abundant. Second, its construction force, which is something that we did not talk about explicitly. They have 1.5 million construction workers. So 15 % of the population in the UAE is constructing something. They wake up, they’re building something. It could be a hotel, it could be a resort, or it could be a data center. That is something that we’re actually very...Grace Shao (44:59) these are mostly workers from abroad, right? From India, Pakistan, Philippines. Yeah.Kevin Xu (45:02) These are almost, these are entirely workers from abroad, right? These are Pakistanis,a lot of South Asians who are there on workers visa. So they’re not, you know, living some glamorous life. They’re just a construction worker life, right? And there are a lot of kind of like issues with that approach, if you think about it. But as far as the capacity is concerned, they’re able to really build stuff faster than just about any country that I’ve seen. And as the United States hits its challenges, I think, when it comes to labor,when it comes to energy capacity and I think will also become a domestic political issue very very soon especially this upcoming year with the midterm election that could grind a lot of the pace to a bit of a halt and the UAE is ready to risk kind ofreceive all that chips that are being made in Taiwan. And I think that will really be something that people haven’t really thought about as far as where their computer will actually physically live, which really will bring again the sovereign AI story of the UAE toto life because it’s not just another talking point anymore. They actually have a significant amount of compute that could be used for training models and it can also service a bunch of the region over there because the telecommunication cable between the UAE and say India, for example, or Pakistani, the speed there is like 30 or sub 30 milliseconds, which is super, super fast. So you can actually serve a bunch of users from the UAE to India if you’re okay with that kind of, you know, data center set up.So that’s something that I think people are probably still sleeping on. We may see that becoming a more real just in another 12 months or so.Grace Shao (46:39) Interesting. Kevin, you’re so knowledgeable and everything. I love reading your work and I just really enjoy this conversation. I have one last question for you, which is a question I ask every single guest that comes on the podcast. What is one differentiated view you hold? Non-consensus, something maybe even controversial that you truly believe in that maybe your peers don’t?Kevin Xu (46:57) I think, I’ll share two, but they’re interconnected. Obviously they’re really one, but in two parts. One is that I think there’s a consensus that China AI, AI in China is diffusing better than the US. I think from an economic perspective, from an economic impact perspective, that is actually not true. If you just compare the revenue number,between Gipu and Minimax to any lab that we have here in the US. It’s peanuts, right? Now you can say we have a bit of a token price inflation over here in the US, as I’ve admittedly mentioned during our conversation, but it’s not 100x premium as far as like that delta is concerned. So there’s actually a lot of economic ⁓ value being captured here in the US just by the diffusion pace that we’ve been able to push out here alone.So that’s sort of a non-consensus thing, the one. And the other thing that is related is that because the pace of diffusion in China is a bit more up and down the stack, you know, not just in knowledge worker, but in factories, in on the road with self-driving and in robotics and whatnot, let’s just assume all these will just kind of continue at pace faster than any other country in the world. Then China will also be the one countryThat has to deal with all the social ramifications of AI before any other country in the world So this is a very interesting moment where the Chinese government and regulators will have to lead the world Into this kind of dark space as we’re all filling out what the hell this AI is gonna do Before anybody else and I’m really interested to wait to see how much they’re willing to share their learningtheir failures, their successes from a rulemaking perspective? And also, how humble will the European regulators and the American regulators be willing to learn from the Chinese failures so we don’t screw up too much in our own backyard?That will be something that I think will happen for sure, but it could really determine the direction of where all this is going. And we kind of saw a little bit of that with the most recent regulation coming out of China when it comes to regulating the chatbots. It’s much more prescriptive than the usual list of harms when it comes to data privacy and whatnot. It touched very specific use cases, like if a chatbot is going to talk a lot about, you know,Grace Shao (49:02) What’s you say?Kevin Xu (49:16) the giving mental health advice or all these much more personal use cases that could lead to self-harm the regulators in China is having a very particular point of view on how this should be Diffused in its society whether that lesson good or bad gets learned here in the US and elsewhere in the world is Gonna be interesting but China will have to lead on this front ⁓ Which is a position that I don’t think the Chinese regulators are used toGrace Shao (49:42) Even expected,Kevin Xu (49:42) ⁓ up to this point.Yeah, they’re used to learning from outside. They’re very good at absorbing the best rules from Europe and the US to bulk up their own regulatory capacity and knowledge. But this could be the one thing where they will be the first to step into the abyss and they have to help us get out of it.Grace Shao (49:59) I think a really, really interesting point. And I actually been thinking about this as well. To your point, I diffusion in China is so much more obvious to the human naked eye because it’s seen through consumer usage, through just the rampant digital infrastructure buildup that we’ve seen in China. So everyone, like you said, from random auntie to knowledge workers will be using AI. But the actual capital gain, the real money has not been proven to be greater than...than the US and already we can see that from just IPOs like MiniMax and Zhipu And I think the regulation that you were talking about actually came out interestingly right after MiniMax and Jhipu actually released their prospectus to the public. So it’s like, I think regulators are really keeping a keen eye and a hand on it and trying to see what could potentially happen. We speak to people in China practicing AI, like the actual builders and the scientists, they say, there’s less of a discussion about this like.Doomerism kind of view people are taking more pragmatic view, you know, people are really focused on technological advancements less about societal implications Yes, I kind of believe that being probably the case given that you know in China last 20 years people really saw technology as a Path to economic prosperity, but however, I think what your point is is really interesting is that actually this time they can’t see what happens how the US regulates by techthey have to do and start themselves, right? So that’s a really interesting point. I actually will think about that a bit more as well. ⁓ Thank you, Kevin.Kevin Xu (51:29) Yeah, there’s a 100% chance that China will have to be the first country to lay off a bunch of delivery drivers and, know, ride sharing drivers if robot taxi becomes a thing, right? What will Wuhan do? I think everybody else in the world will want to know when that happens. Yeah.Grace Shao (51:46) Yeah, yeah, especially whenembodied AI becomes more of a but okay on this point I wonder your thoughts on this because When we go to China, it’s really interesting you have these random jobs that are like placed for sure not for like actual practical reason like you know those aunties who sit in elevators and press the button for you or Like a uncle who sits there like an older kind of larger man who sits outside the parking lot who just pressed the toll button for you likeKevin Xu (52:03) Mm-hmm.That’s Right. That’s right.Grace Shao (52:13) Thesejobs frankly are not needed, but they’re implemented I think for societal harmony purposes because you need employment. You need to give these frankly not very skilled laborers a job. So if you’re gonna push for embodied AI in China and these physical, whether it’s robots or whatnot, are gonna replace a lot of these lower skilled jobs, what’s gonna happen to society? Do you think they’ll actually?implemented at mass or do you think they would actually take a more cautious decision?Kevin Xu (52:43) My read on that is they will be very, very cautious, which again goes to the non-consensus view that I just shared on the diffusion narrative about China and AI. Right now, the consensus is that, oh, China’s diffusion is so much faster. They’re going to push all this AI. We’re screwed here in the US. But really, there is a very good human reason to not do that.⁓ You know, this is not exactly public knowledge, so I won’t cite it. But if you look at the pace of deployment of the self-driving companies operating in China alone, right? know, you Baidu, you have Pony, you have Rewrite, you have some of smaller players. It is, they’re all born there. They have very good regulatory environment to experiment and develop their technology. But they’re actually throttled by local permit capacity.on a city by city basis as far as how many of these cars can they actually deploy on the road? Because it’s not a free for all at all. Every city is looking at the numbers and be like, okay, if we actually do this tomorrow, like let the floodgate open because the technology is actually really, really good already. And we already know the Chinese, yeah, well the Chinese OEMs can pump them out really quickly, right? I think that the safety concerns actually getting really, really good. But what would the delivery drivers do?Grace Shao (53:49) It’s not a safety concern.Kevin Xu (53:59) What would the DD drivers do? So there is this toggling already between how much are the government willing to let this technology loose versus taking care of the aunties who pressing the buttons and the dachu who’s letting you into the parking lot because there has to be a pathway there. It’s just not obviously a subsidy program, but it’s clearly a government funded economically irrational employment program.Right? Like the only corollary we have in the US is the greeters at Walmart stores. I’ve never been to a Walmart super center. There’s like this person who just says hi to you and you walk in and you get your Walmart stuff. Like does that person need to exist? It’s of Walmart’s premium user experience for shopping there. But we don’t have as much of that here in the US, but we certainly have a little bit of that too. Right? So again, China is going to hit that at scale.Grace Shao (54:23) Yeah.It’s part of user experience, Kevin. They want you to feel welcome.Kevin Xu (54:45) before any other country. And they’re trying to figure out the right balance right now as we speak, but we don’t really have a good sense, at least from the outside, of what are the rationales, can we learn from that, can they share more of the thinking, so we can all kind of benefit from that, from a rulemaking perspective.Grace Shao (55:01) Yeah. And you saw that with the urbanization demand, like what, 10 years ago, we saw a huge rise of young men from rural areas moved to urban cities to become delivery workers, whether food delivery or package delivery courier, that created a lot of economic gain for the country. And then when COVID hit, it was crazy. A lot of people got laid off from their white collar jobs. And then you saw a huge increase of essentially Chinese Uber ride drivers.Kevin Xu (55:27) That’s right.Grace Shao (55:29) So all of a sudden people all became drivers and there’s a huge oversupply of riders and now you can call a DD and any car would shut up within like a minute. It will be interesting where would these people go if you’re gonna introduce all these self-driving cars, self-driving delivery man, whatnot. It will be interesting because that makes up a huge part of the urban economy right now. Yeah.Kevin Xu (55:43) Yeah.That’s right. That’s right.And you know, one approach is just that you don’t, right? You just say, okay, we know we have the tech, you can export to the UAE all you want, which though they’re doing really well in the UAE, the Chinese are all with taxi companies. But at home, you’re going to pace yourself because we have a lot of people who are going to get really, really upset if this thing gets unleashed tomorrow, which it can. And that kind of goes against the whole China that just defuses everything because China loves AI sort of narrative.Grace Shao (56:16) Yeah, interesting. Thank you again, Kevin. Really appreciate your time and your insights.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• EVs taking on AI OS and the Delivery War. The Chinese Tech Winners Beyond BAT with Alan Zhang 06.01.2026 49min

  In this episode, I sit down with Alan Zhang (Principal & Portfolio Manager at Ox Capital Management) to map China’s tech landscape through an investor’s lens. We break down how Alibaba, Tencent, and ByteDance are approaching AI, and why the “AI OS” is the real endgame. Finally, we analyze what’s changing in China’s consumer internet, EV ecosystem, and embodied AI pipeline. We also unpack China’s delivery wars (Alibaba vs Meituan vs JD), why quick commerce is structurally different from traditional e-commerce, and how markets price geopolitical risk into China tech valuations.Alan Zhang is a Principal and Portfolio Manager at Ox Capital Management, a boutique investment firm focused on emerging market equities that he co-founded in 2021. At OxCap, Alan leads investments across Asia; previously, he spent years as an investment analyst on the Asia team at Platinum Asset Management.He studied Actuarial Science and Commerce at the University of New South Wales, and he’s even taught advanced econometrics. So if you like the intersection of fundamentals, market structure, and Asia platform businesses, well then, this one’s for you.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Chapters 01:34 Alan’s background: quant → Asia equities03:11 US vs China AI: frontier vs “two-legged” approach05:25 “Uninvestable” China and what changed07:31 Beyond BAT: Xiaomi, Meituan, Mindray, MicroPort09:24 BAT AI strategies and the AI OS thesis13:45 Tencent: tools, data, distribution, and model strategy16:33 AI-native phones: ByteDance × ZTE and what’s next26:51 China EV landscape: BYD, Huawei, Xiaomi, Zeekr31:28 Why phone OEMs can compete in EVs34:16 Embodied AI: robotics parts, redundancy, and Unitree39:38 Valuation + geopolitics: why Asia tech trades discounted41:53 China delivery wars: subsidies, quick commerce, Meituan’s edge50:27 12–18 month predictions + what investors miss (healthcare)AI-Generated TranscriptGrace Shao (00:00)In today’s world, there’s no shortage of information. Knowledge is abundant. Perspectives are everywhere. But true insight doesn’t come from access alone. It comes from differentiated understanding — the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend — someone who can help us see things differently.So today, joining me is Alan Zhang. And I’m Grace Shao. Alan, really excited to have you. I’m excited about today’s conversation because we’re going to get into the investor’s perspective on Asia tech and emerging markets — with a proper markets-and-math backbone.Alan Zhang is Principal and Portfolio Manager at Ox Capital Management, a boutique investment firm focused on emerging market equities that he co-founded in 2021. At OxCap, Alan leads investments across Asia. Before that, he spent years as an investment analyst on the Asia team at Platinum Asset Management. He studied actuarial science and commerce at the University of New South Wales, and he’s even taught advanced econometrics.So if you like the intersection of fundamentals, market structure, and Asia platform businesses, this episode is for you. Alan, welcome.Alan Zhang (01:31)Thank you, Grace. Pleasure to be here.Grace Shao (01:34)Alan, to start, why don’t you tell us about yourself — your background — and what it is that you cover now?Alan Zhang (01:40)I grew up partly in Hong Kong, mainland China — Shenzhen particularly — and in Australia. I spent close to a decade in Australia doing my schooling and education, and worked for a firm called Platinum Asset Management, then co-founded Ox Capital with Joseph Lai.I studied actuarial science, so I’ve had a lot of experience manipulating numbers, cleaning up data — and that helped me tremendously in public equities. Nowadays there’s no shortage of financial data, and the ability to understand them — and the intent behind them — is crucial to investing.Grace Shao (02:34)Yeah, yeah.Alan Zhang (02:46)At Ox Capital, we also built a tool called the Mode Model, which distills more than a million financial data points from various sources to help us understand our coverage region a lot more. In terms of my coverage, I build quant models, I look at equities, and I also help with portfolio positioning based on macroeconomics in Asia.Grace Shao (03:11)That’s interesting because you started off in quant, but now you’re looking at equities — the fundamentals, right? You’re covering a lot of ADRs, and a lot of China’s big tech.Let’s talk about that. What is the China big tech internet ecosystem looking like right now? How does it compare to the US?Alan Zhang (03:20)In the US, they are focusing more on frontier models, while Chinese companies are taking more of a two-legged approach — tackling AI with different approaches. The US has invested a lot of resources into advancing frontier models. On one hand, we see successful cases like Gemini, Anthropic, and OpenAI, while we also see a lot of AI subscriptions cutting their prices by more than 90% in the last few years.If you remember in 2023 and 2024, many subscriptions were priced at a few hundred — some over $1,000 a month — based on investment assumptions. Now they’re cutting prices to sub-$100 a month. Some may never make their money back based on those assumptions, but it’s not being discussed today because the benefit of AI far outweighs that blip, and large-cap companies are investing enough to offset the impact.If we look at China, they haven’t gone through this episode — and I don’t think they will. Anyone who looks at Asia understands Asian users will never assume people will pay over $1,000 a month for subscriptions. China is working on frontier models, applications, and infrastructure at the same time.In summary, China is still the runner-up, but they’re developing AI in a more balanced manner. And it’s also good to see the US pivoting — in the recent 12 months, we’re seeing more US companies investing in software and applications rather than just frontier models.Grace Shao (05:25)China was deemed uninvestable, especially for Western investors. Your fund is based in Australia and Hong Kong, and your LPs are non-Chinese. For public investors who want exposure to China’s AI upside — what are they looking at? What are they thinking?Alan Zhang (05:46)Usually the big tech. China went through the property adjustment and the antitrust campaign in the internet space. It was painful — people called it uninvestable because they couldn’t see new growth drivers. And if they could, they were too insignificant compared to the two most important industries at the time: internet tech and property, which were both recalibrating.But things are different now because investors can see new growth drivers scaling up. In hindsight, these adjustments also helped innovation: talent that dreamed of landing a job at Meituan, Tencent, Alibaba went to smaller firms or startups; capital that made easy money in real estate went to new areas.Economic transformation is still a work in progress, and investing in China becomes more attractive if we see AI, consumption, and advanced manufacturing play a bigger role. We’re still in that phase. But we’re glad to see some companies bottoming out and making progress under the current setup.Grace Shao (07:19)In a pragmatic way, does that mean we’re looking at BAT? What companies should we be looking at for exposure to Chinese AI and economic transformation?Alan Zhang (07:31)Besides Alibaba and Tencent, people should look at relatively smaller cap — but still large-cap — companies like Xiaomi and Meituan. And also industries outside the internet. For example, Mindray in healthcare, or MicroPort in surgical robotics — they can implement AI into their products and make their portfolio more attractive.Grace Shao (07:41)When we chatted offline, you said a lot of companies are overlooked. Beyond BAT — what are some “1.5 tier” or “second-tier” companies that are huge by market cap but not well known in the West?Alan Zhang (08:09)People will naturally see them more over time. Tencent and Alibaba were making active efforts overseas; now as the market matures, more companies are going global. If I’m on a roadshow, people ask about Keeta, which is a subsidiary of Meituan. Xiaomi is opening more stores in Europe — even Africa and South America. People will naturally see them more.If you come to China and compare what’s here to where you live, you’ll see a clear difference.Grace Shao (09:24)Let’s double click on BAT — Alibaba, Tencent, and ByteDance. At a high level, how do you compare their AI strategies? Are they playing the same game, or different playbooks?Alan Zhang (09:52)Same, but different. They’re all investing heavily in frontier models and infrastructure. Ultimately, they all want to build the AI OS people will use. The DoorDash–OpenAI collaboration was a good example of what AI and a commerce company can do. Whether it’s an app within an app or an app within a phone — that’s still an open question.Alibaba is e-commerce and cloud. They have to build a competitive model or their cloud becomes commoditized. Tencent is a platform — they build tools. In LLMs or AGI, late movers can have an advantage because users may be indifferent as long as security and usability are similar. ByteDance, as a private company with strong feed algorithms, has been AI-native for a long time — even back in 2018 they were investing heavily in AI and user intent.So they’re all trying to build an AI OS for users, just from different starting points.Grace Shao (12:29)I love that framing — I’ve been writing that 2026 is about the AI OS. Tencent has signaled they’ll double down on LLMs. It’ll be interesting to see whether late-mover advantage shows up — and whether they need to spend less on pure infra.How should we think about Tencent’s positioning? They’re late on LLMs, but AI is already integrated across touch points — WeChat, gaming, fintech, mini programs. Should they continue using open-source models like DeepSeek, or focus on proprietary models like Alibaba integrating Qwen?Alan Zhang (13:45)They’ll do both. With Yao Shunyu reporting to Martin Lau, they’ll try to build their own model like every tech giant. At the same time, Tencent’s bread and butter is building tools — AI tools to help merchants and users and improve the experience.Whether it’s an app within an app or an app on a physical phone — like the Doubao phone we saw — Tencent has the ingredients: ecosystem, quality data, and distribution.Grace Shao (14:37)When you say “building tools,” how is that different from Alibaba building tools for businesses? And how is that different from ByteDance’s “app factory” approach?Alan Zhang (15:10)One example: in WeChat’s input bar, if you long press, you can translate. People type in their own language and WeChat translates to the recipient.I also visited their AI showroom recently. They showed mapping genetic pools and building a genetic bank for seeds and animals — they have quality data. They can also build full simulators for flights and cockpits — one of only a few companies that can do that. They’re investing in spatial intelligence and data banks, and building tools inside WeChat.I think it’s only a matter of time before they move more properly into e-commerce and release something like what DoorDash and OpenAI shipped.Grace Shao (16:33)On hardware — can we talk about ByteDance and ZTE’s partnership? ByteDance worked with ZTE and launched an AI-native operating system on a ZTE phone. Instead of building their own phone, they partnered with OEMs. What do you make of that?Alan Zhang (17:11)As a user, I looked forward to it. A product like this may take longer to be widely available because it disrupts a lot of vested interests. But the trend is inevitable — AI OS will be valuable in ways we can’t even measure.This is what I envision for Xiaomi and Tencent too. Companies like these — and Apple — are planning for that day, but they’ll move when stakeholders are ready. OEMs have the protocols to make it happen. Tencent also has content and intent — ads revenue — plus distribution. Tencent and Xiaomi will try to tackle this new market.Grace Shao (18:13)Is ByteDance moving faster because it’s private? Xiaomi and Tencent are public companies — does that slow them down?Alan Zhang (18:29)Absolutely. ByteDance can try something new; if it fails, it doesn’t impact the core. If Tencent or Xiaomi do this, they can agitate business partners and users.Grace Shao (19:10)For an American audience, is there an apples-to-apples comparison to US peers?Alan Zhang (19:32)It’s difficult. These companies are mature and make decisions based on their own opportunity sets. In many spaces, Chinese companies are leading, while the US is still exploring new frontiers. Tencent has been relatively quiet until recently, and they work quietly with industries to understand how their AI stack helps.In 2015, Tencent founded a learning program called Tencent X — “X” stands for another Tencent. They work with business schools, bring entrepreneurs and business leaders to site visits and exchanges, and use the process to understand how to develop their stack to empower Chinese industries. A successful example was Pinduoduo — through this program, they found Colin Huang and supported the company through traffic. Tencent can find more companies like this in their own way.Grace Shao (20:46)[Connection drop]Grace Shao (21:04)Could you restart that sentence?Alan Zhang (21:08)[Repeats Tencent X explanation]Grace Shao (22:07)Looking at 2026 — what consumer AI applications might look different? Any sprouts inside super apps that people aren’t noticing yet?Alan Zhang (23:07)2026 will likely be an interpolation of 2025. I don’t expect a completely new form factor. Most Chinese companies are already super apps, boundaries are ambiguous, and they’re fighting for the same consumer pockets.But ads revenue will shift. Previously, ecosystems charged a lot for ads because of captive customers. With AI, people are reconsidering how they use apps — budgets will relocate to new apps.Grace Shao (24:20)On infrastructure: it feels like everyone is shipping models — not just BAT and the “four tigers,” but also Kuaishou, Meituan, Xiaomi, even EV players. Why?Alan Zhang (24:58)They have enough users, and AI improves experience and broadens reach. For example, older users didn’t use search much, but with AI they can adopt faster. AI makes products more interactive and easier to use.EV companies want more engaging products. Cars are becoming commoditized, so they invest in infotainment and ecosystems. That’s why every sizable Chinese company will try to build a model. And we’re still in the investment phase — nobody knows who wins, so everyone tries. It’s not as expensive as it sounds.Grace Shao (26:25)Isn’t it costly for EV companies?Alan Zhang (26:32)It’s costly, but a lot of money is spent on chips research and manufacturing. The LLM itself isn’t as expensive as people imagine.Grace Shao (26:51)Let’s double click on EVs. Who are the biggest players in China beyond BYD and Zeekr?Alan Zhang (26:55)BYD and Huawei. Emerging ones: Xiaomi and Zeekr.Grace Shao (27:15)How do you position them?Alan Zhang (27:21)Xiaomi’s selling point is ecosystem. You can call “Xiao Ai Tong Xue” — the voice assistant — to operate devices through the ecosystem, especially with HyperOS 3.BYD’s advantage is manufacturing — they can build similar-quality cars cheaper through supply chain management.Huawei has HarmonyOS and strong brand equity — customers pay up, so they can stack a more luxurious experience into the car.Grace Shao (28:15)How does that compare to “luxury EVs” like Nio — are they still relevant?Alan Zhang (28:24)They’re still relevant. Li Auto is more family-oriented than luxury. Nio targets younger consumers who want the driving experience. Huawei’s models skew more toward corporate executives and founders — generally 40 and above.Grace Shao (29:08)So there’s a shift — five years ago it was BYD, Nio, Xpeng, Li Auto; now Xiaomi and Huawei are making strides because of AI operating systems. Is that right?Alan Zhang (29:28)Yes. China’s auto market has many brands and licenses, no shortage of production capacity — and there’s overcapacity. The “anti-involution” campaign has targeted autos. The industry is commoditized, so companies need differentiated advantage. Xiaomi and Huawei have ecosystems; BYD differentiates through cost and can scale domestically and overseas.Grace Shao (30:41)Why are Xiaomi and Huawei able to lead? Does that mean EV-first companies become less competitive?Alan Zhang (31:28)EVs have fewer parts than ICE cars. Historically you needed over 10,000 parts; now EVs might have a few hundred to just over a thousand. You can break it into powertrain, battery, chassis, and battery management — and the rest is non-core. Many parts are commoditized except the battery and system.Xiaomi and Huawei can repurpose capabilities from phones: chips, screens, packaging. Xiaomi can repackage Qualcomm chips and repurpose them to be auto-grade; Huawei can do similar. Cars also have bigger screens than phones — manufacturing capability transfers.EV-first companies like Nio, Xpeng, and Li Auto spend on manufacturing and also on chips, because their bigger vision is robotics. They’ve said chips for EVs alone wouldn’t pay back — the bigger scheme is robotics.Grace Shao (34:16)So in embodied AI: you have Unitree, “Galabots,” UBTECH; you have EVs; you have Xiaomi/Huawei tech stacks. Who wins? Is it just cost and price?Alan Zhang (34:54)Cost, price, and redundancy for physical movement. Even traditional automation companies like Inovance are building robots. A robot shares parts with EVs — optics, gears, batteries — but also has new parts like PLC controllers where you need redundancy. On these fronts, many are on a level playing field.Grace Shao (36:12)Do Chinese EV firms have an edge in spatial intelligence, or is it mainly cost?Alan Zhang (36:21)China is still runner-up in spatial intelligence and will spend time to catch up. But China has a short feedback loop: optical components and supply chain are local; ideas can turn into products quickly and iterate fast. Not an advantage yet, but not far behind.On who wins: too early to say. Unitree is the one that can make a more agile robot and do more stunts than other players.Grace Shao (37:42)Where does AI show up in embodied systems — is it just visible “smart” functions, or more invisible?Alan Zhang (38:19)Besides user experience, AI processes many parameters in the background. With enough computing, embodied AI can make simultaneous decisions — what to move and what not to move. Humans blink, walk, and raise hands at once; without AI it’s harder for robots to act like that. With AI, robots can handle more parameters and make simultaneous moves.Grace Shao (39:38)How do you price geopolitical risk into valuation positioning? Export controls, trade wars, domestic regulation — how should investors look at China?Alan Zhang (40:17)The market is already pricing a discount. Asia tech trades at a discount to US peers — Samsung and SK Hynix versus Micron; BAT versus the Magnificent Seven. Tools may be less available, which can slow advancement, but it’s also encouraging to see alternate solutions like DeepSeek. Over time companies can become more technologically independent.For large caps, investors may feel safer sizing up. For small caps, we start small and see how it plays out. Entrepreneurs are agile and prepare for change.Grace Shao (41:53)A reader question: China’s delivery wars. Alibaba vs Meituan — subsidies, vouchers — why is this happening now?Alan Zhang (43:43)Meituan has led quick commerce — 30-minute delivery — and it surprised me Baba took so long to react, because quick commerce will take share from traditional e-commerce. A few years ago Meituan delivered iPhones at launches — a wake-up call for JD. The new delivery war kicked off with JD’s initiative around April; JD spent heavily to buy consumers, and Baba joined a month or two later.Money could be better spent elsewhere, but I understand Baba — if they lose relevance in e-commerce, other businesses stop making sense. E-commerce is the core.Despite growing daily volume from 30–40 million to 80 million — sometimes 90 — it’s discouraging Baba hasn’t improved delivery efficiency much. Meituan was already profitable at around 40 million drops a day by carrying multiple deliveries per trip and improving dispatching. It’s sad for investors that many platforms are still loss-making due to subsidies, but Meituan’s underlying efficiency advantage remains. As a consumer, the subsidies are great.Grace Shao (46:29)Why does Meituan have such an advantage in dispatching and logistics compared to Alibaba, which has massive logistics and warehouse footprint?Alan Zhang (47:39)It comes down to the core. Meituan built it through local business development — ditui — integrating merchants into inventory and payment systems. Inventory is kept locally, so Meituan focuses on dispatching and rider movement. Their algorithm can even predict demand and move riders toward hotspots ahead of time.JD invested heavily in centralized logistics hubs and infrastructure — that makes them slower to pivot. Baba used an asset-light model early, working with ZTO, and is more centralized — mostly Hangzhou. Meituan is more decentralized and localized. In quick commerce, doing well in one city doesn’t guarantee another — but once dominant, you can use profit from one pocket to subsidize another. Traditional e-commerce is more centralized.Grace Shao (50:03)That’s a fascinating lens — culture and management style mapping to business model outcomes.Alan Zhang (50:04)And risk. In food delivery, you can’t hold inventory. Meituan works on the assumption you don’t hold inventory. Baba and JD have more of a culture of holding inventory and keeping products in storage longer.Grace Shao (50:27)Closing: biggest prediction for China tech in the next 12–18 months?Alan Zhang (50:51)One for EVs, one for internet. In EVs, OEMs with a pure domestic focus and without an ecosystem will lose relevance in 12–18 months — consumers are making up their minds. In internet, with Tencent hiring Chief AI Scientist Yao Shunyu, we’ll see more AI functionality built into Tencent’s ecosystem.Grace Shao (51:54)What’s one company or subsector global investors are sleeping on?Alan Zhang (51:56)Healthcare. It can be resilient regardless of overall spending. The market is focused on frontier-model spending and ROI, but healthcare companies aren’t budgeting for “latest and greatest” models — they’re looking at applications that improve products and ecosystems. Even if we stopped advancing frontier models for four months, there’s tremendous value to extract from current models.I see a mindset shift among healthcare executives to build AI into products and sell superiority — historically, tech adoption was cost-driven; now it’s revenue-generative. Mindray in Shenzhen, or MicroPort in the Yangtze Delta — great companies. Surgical robots and medical devices are not far behind other systems.Grace Shao (54:15)Final question: what’s one differentiated view you have that’s non-consensus?Alan Zhang (54:37)Instead of focusing only on AGI timelines or capex or cloud consumption, we should think about daily businesses and smaller-scale businesses extracting real value from AI — even financial companies. I’m excited to see new form factors and more AI functions in consumer products.Grace Shao (55:24)So focus on practicality and real use cases — not just headline spending.Alan Zhang (55:33)Absolutely. Look beyond the top three, top five — and don’t go too far down the risk spectrum.Grace Shao (55:39)All right. Thank you so much, Alan. Thanks for your time today.Alan Zhang (55:43)Thank you, Grace. Pleasure to be here.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Z.ai/ Zhipu: one of the first major LLM start-ups to go public. Competition with giants and aims for AGI 29.12.2025 51min

  In this episode, I sit down with Zixuan Li, who leads the chat API and global partnerships at Z.ai, one of China’s leading LLM labs (one of the four tigers) and now one of the first to head toward an IPO. Z.ai started as THUDM, a Tsinghua data-mining lab best known in open-source circles for GLM and CogVideo, and has since grown into a model-as-a-service platform powering millions of devices and thousands of enterprises in China and beyond.We talk about what it actually means to be an “independent” lab in a market dominated by platform giants like Alibaba, ByteDance, and Tencent, why Z.ai pivoted from SOE-heavy infrastructure projects to a product-led GLM stack, and how they landed on a different business model, and the creation of the GLM Coding Plan, instead of charging by tokens. Zixuan is very candid about pricing (“If Anthropic charges $200, we charge 200 yuan”), the realities of on-prem-first China vs cloud-first West, and what it’s like to race against Minimax and Moonshot with fewer GPUs and less cash.We also zoom out and look at China’s AI talent pipeline (and the meme that the AI race is “Chinese in China vs Chinese in the US”), how he thinks about AGI as self-learning agents that live on your phone, why he’s comfortable being a white-label backbone in the Global South, and where he sees China’s AI landscape in the next 6–12 months. If you want a ground-level view of how a Tsinghua spinout is trying to survive, and maybe win, in the LLM wars, this one’s for you.Newly launched (Dec. 22) GLM 4.7: In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.01:20 – From THUDM to Z.ai: rebrand, Tsinghua roots, and model-as-a-service03:30 – Quiet period & IPO: pride, pressure, and the business challenge of LLMs06:33 – Pivoting from SOEs: infra projects, agentic models, and why strategy followed capability07:25 – Competing with Minimax, Moonshot & DeepSeek: focus, compute, and capital constraints08:34 – Chasing benchmarks vs real-world IQ: math, humanities, and alignment trade-offs11:05 – On-prem vs cloud: why Chinese SOEs still won’t touch APIs13:43 – Zero-retention and trust: can China’s culture around data ever shift?14:07 – Inventing the GLM Coding Plan: subscriptions, stickiness, and “pay by value, not tokens”16:00 – “If Anthropic charges $200, we charge 200 yuan”: pricing strategy and margins and GLM’s open-source flywheel19:41 – Who really pays: sticky indie devs, big tech customers, and bargaining power23:32 – GLM Coding Plan vs Cursor/Qwen/Claude: plans, agents, and avoiding lock-in25:57 – Z.ai’s AGI ladder: AutoGLM, self-learning, and personalized weights27:03 – Independent labs vs platforms in China: speed, resources, and “dirty work”29:34 – Moonshot vs Z.ai: chasing the moon vs being “down to earth”30:53 – Will China’s LLM market consolidate?: 5–10 players, Doubao, and video-generation winners31:44 – Doubao phone & Honor partnership: bargaining power with OEMs34:11 – Beyond China–US: Global South strategy and being a white-label backbone35:29 – Being comfortable as infrastructure: letting others own the brand38:05 – Who joins Z.ai and AI talent: thriving with scarce resources40:07 – Culture, 007 hours, and survival: what it takes to be infrastructure42:33 – Social welfare, AI safety, and cheap tools in India & Indonesia44:38 – How China actually talks about AI safety (or doesn’t)47:29 – Differentiated view: why Zixuan believes you should “enjoy lacking resources”AI-Generated TranscriptGrace Shao:Hey Zixuan, thank you so much for joining us today. Really excited to have you on. Walk us through your journey and what led you to Z.ai to start off with.Zixuan Li:Yeah, so currently I’m the head of Zhipu AI’s chat API services and also head of global partnerships. I collaborate with LMSys Chatbot Arena, OpenRouter, Vercel, these large companies, and ship our products through their platforms.The reason why I joined Zhipu is it’s one of the leading AI labs in China and I can do overseas businesses, because I have a background at MIT’s Schwarzman College of Computing. So that brings my knowledge into real-world practice.Grace:I see. Was there any incentive for you to move back to China versus stay in the US?Zixuan:I think it’s more personal, because my wife’s based in China and she’s used to her work, so there’s no way she can move to the US.Grace:Fair enough.So let’s talk about the company’s mission and origins, because I think it does seem a bit mysterious, especially to people outside of China. From the outside, people know Zhipu, Z.ai as one of the leading Chinese LLMs. But that doesn’t really capture everything you guys do, right?In your recent prospectus, you describe yourself as a MaaS — model-as-a-service — company first. So tell us about that.Zixuan:Okay, so before Zhipu AI, we were called Zhipu or THUDM, because we named ourselves by the AI lab’s name. We originated from Tsinghua University’s data mining group — THUDM. But I think it’s hard to pronounce, and also “Zhipu” is also very hard to pronounce. So this year we bought the Z.ai domain and finally changed our name to Z.ai.When we were called THUDM, we were very famous inside the open-source community because we had a lot of repos, a lot of models under the THUDM name. And we open-sourced not only text models, also CogVideo, CogView, these models. I think they were sold at that time.But with the launch of VEO, Hailuo, and also a lot of current top models, we began to be more focused — basically more focused on text models, visual understanding, and so on. So I think that’s the origination of the lab.But as you said, there’s this terminology called model-as-a-service. From our side, when we compete with large companies like Alibaba and ByteDance, we need to be more focused. They have their inference level, they have their cloud services, but we don’t. So we try to let the model itself provide the service — like the API, or technologies like visual understanding — and try to use the model itself to be the selling point.Grace:I definitely want to double-click on how you position yourself compared to peers — a few of them you just mentioned, whether it’s Minimax and Moonshot, and then you also mentioned the BATs.But to start off with, you’re currently in your quiet period as your prospectus just hit the public. And if successful, you will become one of the first major LLM startups globally to be listed on a stock exchange. How does that feel?Zixuan:I think we are proud of it, but things are very challenging, because it’s really hard to do LLM inference. Both OpenAI and Anthropic have very high revenue, but a lot of loss on their income statements. So we have to figure out how to make money from large language models and also provide cheaper service to the customer.So I think it’s only a starting point for us.Grace:Definitely. I think right now only the big tech companies in many ways are essentially seeing ROI, and the model companies and the model labs themselves are really finding it hard to make a profit.I want to ask you about the branding. You did say you guys changed your company’s name to Z.ai this year, partially because Zhipu is just hard to pronounce. But was that also related to the fact that you guys seem to have made a pivot into really focusing on going global? Z.ai seems to be a lot more non-Chinese-native-speaker friendly, right? So is that the push right now?Zixuan:I think that played an important role, because we have observed the success of DeepSeek, Qwen — they got famous globally and Chinese people will think that they are the “SOTA” in the domain and their models are the best. They are recognized by NVIDIA and other large company CEOs. So I think that’s one factor.But the other factor is when we changed the name to Z.ai, the dot also plays an important role. We want people to enter that URL into their browser and try to visit our website. Yeah, two factors.Grace:And tell me about your origin story, actually. You mentioned earlier you started off from the Tsinghua data mining group. Maybe provide some context to people outside of China. What does Tsinghua represent? I mean, it’s an institution, it’s a university, but why are so many of these LLM companies or even deep tech companies coming out of Tsinghua right now?Zixuan:I think it’s kind of a combination of Stanford and MIT. So talents are everywhere and there’s a lot of funding from internally and also externally. And also people are chasing the highest IQ there. So it will be very natural to pursue AI in Tsinghua University.Grace:So I have a question on that, because a lot of tech companies, even the previous generation internet companies that came out of Tsinghua, had some kind of connection with Beijing city. And my understanding is Zhipu’s original business model was also very focused on SOEs and local government work, both in China and even across Southeast Asia.Before the more recent pivot leaning into tools and APIs, what were the reasons for the pivot from the heavy AI infrastructure focus and SOE projects to a much more product-led tools and API strategy?Zixuan:I think it depends on the capabilities of the model, because nowadays the model can perform agentic tasks, use tools, use coding to perform tasks. But before that, we could only do customer service, data processing — these “dirty work.” I think it’s better for SOEs or other scenarios.But with the change of Cloud Code, GLM 4.5, these agentic stuff, people can really use the model in other areas like Manna, Gainsburg, Lobe. So I think it’s not only our strategy, but also the capabilities of the model have changed a lot.Grace:Yeah, and I think to put you on the spot, where do you see yourself compared to your peers — like the DeepSeeks and the Minimaxes and the Moonshots of the world?Zixuan:I think compared to Minimax and Moonshot, we are close competitors. We are startups, but DeepSeek is like another kind of enterprise because they have Qwen. So I think they’re very unique, and also ByteDance, Alibaba, they’re sitting at the same table. So they’re from large enterprises.We are all chasing somehow the same direction, but we lack compute, we lack money compared to these giant enterprises. So we need to stay very focused.Like Moonshot, they focus on the Kimi K2 series. They only release Kimi K2, another K2 and K2 Thinking this year. And also Minimax — they’ve become more focused and they kind of shift away from multimodal to text models. I think it will be very fierce. The competition will be very fierce in the coming months.Grace:And for yourself, when you say you’re chasing the same direction, what does that direction look like in layman’s language?Zixuan:In layman’s language, I think… more practical. Because the reason why we do coding and agentic is that we see people using it. We see people using Codex, Manna, Claude Code. So that represents high token usage.And also we are chasing AGI, or the IQ. So we want the model to solve very hard math problems, to memorize a lot of hard stuff. As you can see from a lot of benchmarks like ARC-AGI, HLE, we’re also chasing in that way.So we balance the two: figure out how to balance the performance on benchmarks and in real-world development.Grace:I actually have a question on that that’s a bit off-track from the business strategy side of things, but I wonder how you view this.So you’re saying you’re chasing benchmarks on math problems, IQ, advanced physics, etc. But what about the humanity side of things? I think people are still questioning whether AI can be used to replace humans in a more humanities-focused industry or sector.Zixuan:So that’s a very big issue. But for now, I think it’s still not there yet, because we see hallucinations happen inside the model and instruction following is not very good.So we test the status of that harm and try to assess what we can do with this model and try to synthesize a lot of data to make it more aligned to human judgment or other things. I studied alignment at MIT. I know a lot of stuff, but when I came back to China, I thought we were not there yet.So capabilities, I think, are still more important than alignment at this stage. But we need to focus on the future and try to prevent something really bad happening. I’ve learned a lot of news like suicide or emotional feelings, depression. But somehow I think it’s still not that harmful yet.We try to incorporate as much human judgment or human alignment into the model as we can. But as I said, it’s kind of a balance between different aspects.Grace:Yeah, it’s always a balance between setting up the guardrails and actually still allowing the technology and innovation to continue, right?I want to reshift the focus back on business model, pricing, deployment. Reading the prospectus, what stood out to me was how much you support both on-prem and cloud.What are the main product lines today of Z.ai or Zhipu, and how do you map those onto on-prem versus cloud deployments in terms of how customers actually adopt GLM? Because I do believe I read that in China there’s a very different preference. In China, it seems like more people prefer on-prem, right? Whereas in the US it’s more cloud — or did I understand that incorrectly? Please explain.Zixuan:Yeah, I think you have a very good understanding of the current status, because large SOEs, large enterprises in China, prefer on-premise or more private deployment. So it’s hard to do API services with them.But currently a lot of tech companies accept API services. So we collaborate with nine out of ten of the largest Chinese tech firms or social media firms with our API services. So it depends on their needs.We try to sell API, but actually some people have privacy concerns. They have policies not accepting API services. They don’t want any data to go away from their servers. So basically it depends on the users’ needs.Grace:This is actually kind of a reflection of what happened during the SaaS era too, right? Chinese SOEs and big companies would rather build their own app — maybe not even be as good — but they just don’t want to give their data out to anyone and have that potential security risk, right?So do you think that will change in terms of company culture as we see AI continue to develop, or do you think that will continue to be the trend in China — that this would be the differentiating point between the Chinese market and maybe the Western markets?Zixuan:I think it will continue to be the trend. As you said, we had that pattern in the era of SaaS. And when we go to the AI era, nothing changed.But somehow, we can figure out a way to balance, because there is more “private host on cloud” service. And we’re trying to store user data in a more secure way, with a zero-data-retention policy. That will mitigate the risk and the issues and try to let them feel more comfortable with it.Grace:I see.A lot of Western developer tool companies now go pure usage-based, but you guys also have a GLM Coding Plan — basically for developers with very low entry points. Why did you choose a subscription approach versus going with other pricing models? I guess this part, I just want to understand how you guys are making money right now, especially as you’ve just had your prospectus go public.Zixuan:Yes, I think we are the company that invented this coding-plan business model, because we found out that API users are not sticky. One day they use Claude, they can switch to Gemini or GPT another day. It’s the same with Chinese models.So we remembered: why do we pay for subscriptions — Spotify or YouTube service? Maybe we just listen once during the whole month, but we don’t regret it, right? So we don’t want our users to pay by tokens. We want them to pay by value or by the product itself.So if they just use it once or twice within a month, I think it’s totally fine. If they want to subscribe the other month, we’ll try to provide better service. We have GLM 4.5, GLM 4.6, GLM 4.7, trying to ship better models. But if they decide to quit, I think it’s still good for us because they paid for one month, not just several tokens.So the users may be very sticky here, and we have our branding — not only the model, not only GLM, but also the subscription, GLM Coding Plan. So when we use Cursor…Grace:But if they were to use it a lot, would it be loss-making for you guys then?Zixuan:I think it’s still an issue for Claude Code and also Codex. You have to balance the rate limit and also the service level. So for us, we are very generous, but we’re trying to operate globally, because that will make our traffic more stable — not receiving very high demand at one time and no demand during the nighttime.Grace:I see.I know that you’ve been quite active in a lot of podcasts recently. You were on ChinaTalk, you were on Steven Hsu’s. And one of the lines you said, I think it was on ChinaTalk, you said, “If Anthropic charges $200, we charge you 100 yuan.” I thought it was quite funny. It was very memorable.So how did you make that kind of decision, and how does it work in practice? Does that mean you have a long-term structural advantage, or does that mean you charge less and therefore have smaller margins?Zixuan:I think we serve different customer needs. For example, someone sells Rolls-Royce to people, but we sell Benz to people. Both are good cars, but Rolls-Royce charges way more than Benz. The performance, I think, is very close.But like I said, Anthropic deserves that premium. But by selling Benz, we can still earn a lot of money. Maybe the profit margin is very thin currently, but we can lower the inference cost. We can change our infrastructure to make it more profitable. So it’s a long-term strategy, not focused on the current cost structure.We’re trying to make people more sticky to the brand, more sticky to the service. I think it’s essential at this time.Grace:Essentially, you’re saying the utility purpose of having a car — getting from point A to point B — is the same, but maybe you’re selling a Toyota then or a Honda, right? Not even Mercedes, which still charges a pretty premium margin.I remember in the same interview, you were kind of challenged, saying: look, you only really take up about 5–6% market share in China for general-purpose models. But you said, “Wait, 5% is enough.” What exactly are you thinking when you say 5% is enough?You serve — I think from public disclosures — 123 large enterprise clients on-premise deployments, plus around 5,500 customers using cloud services. How does that actually stack up to your peers? Because it doesn’t look like huge numbers, to be honest. And that already is 5–6% of China’s market share.Zixuan:I believe that the 5% refers to the percentage of all the GLM services.Grace:Yes, sorry, GLM.Zixuan:GLM services, because we open-source our models. And it’s hard to get revenue when you open-source your model because you have to compete on speed and stability.But I think our model is good enough. Maybe it’s not like Toyota — it’s kind of a Benz. And we let more people adopt GLM, like what Qwen did in the past. They open-sourced their reflection models and more people tried out Qwen. They got famous, so people believed they got better service from Alibaba.It’s the same underlying methodology from our side. So if GLM gets really famous, even 5% is enough for us. But if it’s not famous, 5% is totally not okay. We’re trying to make our model more influential, like DeepSeek, like Qwen.Grace:I see. I do want to go into GLM and your tools later as well. But one last question on the business side of things. We kind of touched on this: you said a lot of your customers are the big tech companies, but in the beginning, they were the SOEs, right?So right now, is there any pattern you’re seeing in terms of who becomes the most valuable users and who becomes the most sticky users and who are actually willing to pay the big bucks for your product or for your service?Zixuan:So from my department, I think two types of customers. One is individual developers, because we have the GLM Coding Plan. Someone bought a yearly max plan. A lot of users bought yearly plans. They are very sticky.And the other type is large tech companies, because we are still leading the open-source models. So we have bargaining power. Maybe they want to shift away from our model and choose other models, but we keep evolving from 4.5 to 4.6 and 4.7. Every time they try to change the model, they find that we can ship better models.So these customers are very sticky. And they care more about performance because we are leading in performance. They care less about cost or relationship.Grace:Nice.Let’s actually double-click on GLM. You mentioned GLM 4.5 and 4.6. They’ve been positioned as highly competitive on coding and reasoning, and you’ve often been the highest-ranked Chinese model on public leaderboards.When you compare the GLM series to US and Chinese peers, what dimensions matter most to you beyond the leaderboard scores right now? And where do you think GLM actually genuinely stands out compared to other peers, whether it’s American ones or Chinese peers?Zixuan:I think real-world development, real-world practices, and general chat — these real practices — are more important than benchmarks. And in terms of real-world experience, we are tier two, because I believe Anthropic, DeepMind, and OpenAI have better user experience compared to us.But I think we are enough compared to other open-source models, because we understand user needs. We have better quality in data — pre-training data and post-training data — and we’ve figured out ways to synthesize agentic tool-use trajectories and very hard problems. That makes us stand out in solving these really tough problems.Because when you look at the benchmarks, they are not for real-world practices. Some are very tough, but it doesn’t mean they stand for human practices. Because we have a lot of customers… yeah.Grace:Yeah, I’m going to challenge you on that actually. What about the Alibabas of the world? Because when I speak to Alibaba or Tencent, they also say their biggest differentiating point is real use-case data. And frankly, they have all the existing touch points with their users, whether it’s getting data through helping with businesses, enabling businesses, or consumer use. They probably have the best data, right? So how do you compete with that?Zixuan:I think that’s their advantage in 2024, but not 2025. Because in 2025, most of the high-quality data we need, we have never met in real-world use cases.When you want to create a slide, you first do search and then come back and do another round of thinking, and then choose a design tool or something like that. Nobody interacts with Alibaba’s product like that. So you have to fully understand Cursor, Claude Code, Manna — how these tools interact with people.So ByteDance and Alibaba’s customer data cannot play a role in today’s agentic era. We have understanding of maybe Claude Code or Codex — we try to understand how a top-performing agent manipulates tools and how our model can be integrated in that system.Grace:I was actually going to ask you about the GLM Coding Plan. So for context for listeners, it’s essentially their tool, like a Cursor tool.So how does the GLM Coding Plan actually compare with Cursor or Alibaba’s coder, as you mentioned, or Claude, in terms of coding experience? For a developer who already knows these tools, how would you explain the distinction — or, you can be frank, is it mainly a pricing advantage here?Zixuan:Okay, so I want to compare Cursor with GLM Coding Plan, not the model. Within Cursor, you have one coding agent and you can switch between different models. But with GLM Coding Plan, you first select the model and then you can switch between different tools.You can integrate GLM into Claude Code, Kimi Code. You can even use GLM in Cursor with GLM Coding Plan. That made our product or model widely accepted or widely integrated into these systems — not just for Claude Code, but also it can be integrated into Cursor or Kimi Code.We understand different coding agents and try to synthesize data that best fits these coding agents’ needs. And there’s no lock-in for our users.Grace:So your GLM Coding Plan is not only your proprietary model, right? You actually are open to multimodal?Zixuan:Yes, it’s a model. GLM Coding Plan is called a plan, not an agent, not something like Claude Code. You subscribe to an API, you’re not subscribing to a product. You use that API maybe in Claude Code, maybe in Kimi Code. So you can choose the mode.Grace:Yeah. Okay.Okay, thanks for explaining that to me. That’s helpful, I was getting a bit confused there.Now I wanted to ask: in your prospectus, you laid out five stages of progression into AGI. We talked about your vision of AGI earlier. You said it’s about real-life implications, real-life practicality, usage of AI.When you look at where you guys are at right now, what does crossing the next stage look like in terms of concrete capabilities or products or tools? Or maybe a more straightforward way of asking this is: what should we be expecting from you guys in 2026 to help you progress on your so-called AGI pursuit?Zixuan:Maybe self-learning. Because currently when we do reinforcement learning, we synthesize all the data, we prepare the data beforehand, but the weights of the model won’t change during the interaction.For example, we have this AutoGLM. It’s a model that can be deployed on your phone and can manipulate different apps for you. It can order food or order an Uber for you, but it’s the same model for everyone.To chase AGI, we might have AutoGLM for everyone. When you interact with the model, the weights of the model may change. Currently, we have a memory engineering package that’s more on the engineering side — handling this memory stuff.But for AGI, it needs to be very personalized. Every model needs to be personalized. The model learns from the environment, from the interaction. We also call it on-policy reinforcement learning.Grace:I see.Let’s take a step back and look at China’s overall LLM landscape and competition. You kind of alluded to this earlier — you guys are in the same pool as the Minimaxes and Moonshots of the world. Then there are the big techs like Alibaba, ByteDance, Baidu, Tencent, even Huawei these days, right? There’s so many. Everyone’s producing their own LLMs now.From inside the ecosystem, what do you see as the structural differences between independent labs versus the big tech platforms — in terms of commercialization of their models as well as their incentives and objectives in the coming year or two?Zixuan:Strategy and objectives. Because we lack resources, we need to be very focused. And when we are very focused, we need to move very fast.For talents, our team is very small. I lead a team…Grace:It’s not that small — a couple hundred, right? You guys have like 800 people now?Zixuan:But for every team, there are just a bunch of people. We have sales, we have product solution, but for the product team, product solution, or training team, sometimes you need to be very lean. You don’t have to hire a lot of people, because they chase different directions.Sometimes you have to hire people that can do “dirty work.” Maybe one person is enough to do all the training on this side, and you have a bunch of people preparing data or understanding customer needs for you.Like I said, you have to understand Claude Code, you have to understand these coding agents. So there will be people studying all the products, looking inside these products to see why they are performing so well.But for large enterprises, they can hire a lot of researchers. They have enough resources to do a lot of experiments. They have compute, so they worry less. Maybe they can find some scientific breakthrough from those experiments.But in terms of model performance, I think our competitive advantage is we are closer to users and customers, because we move faster together with our users.Grace:So that’s how you position the startups versus incumbents. But what about just within the startups yourselves? How do you differentiate yourselves between one another?Zixuan:I think compared to Moonshot — because we both originated from Tsinghua University, we know each other pretty well — I think we are more down to earth. We are the ones that care more about real-world usage or practices.Moonshot is kind of… they have this “AGI plan,” chasing the moon or landing on the moon, and they have more imagination on the surface. We’re also chasing AGI, but when we train the model, we care more about real-world practice and usage.Grace:You’re taking a more pragmatic approach. And they’re definitely, I think, a very eccentric bunch, right? Even the name — how it came about — was quite interesting.So do you think eventually in the Chinese LLM space it’s going to be winner-takes-most? Maybe not winner-takes-all, but winner-takes-most? Or is it going to be able to support multiple strong players?Because there’s been rumors about consolidation for a while. There are quite a few players for how big the market is, and like you said, it’s extremely capital intensive. Not everyone has this much money to keep burning through it. So where do you see the direction of this fragmented landscape right now?Zixuan:I think the market is enough to include 5 to 10 players. I think it’s enough. And like I said, the large enterprises only accept on-premise deployments, so there’s no way a winner can take it all, because there are thousands of large enterprises. You don’t have the team to deploy models for every single enterprise.But in terms of applications, maybe Doubao will take more than half of the consumer side. And also for video generation, there will be a winner. But I think the market is still very large to have all these players, and they will compete for a long time. I can guarantee that they will compete for a long time.Grace:What do you think of the Doubao phone situation? This is completely random. This is not relevant to our LLM conversation, but I’m quite curious to hear your thoughts on it, because I think it’s making a lot of noise outside of China. People are quite curious to see where that will lead to.Zixuan:So we are the first company to launch this phone use agent. But I think the issue is bargaining power. We also collaborate with a phone company, and instead of using something like a “GLM phone,” we finally used their name. Their phone, powered by our model.Grace:Which phone is this?Zixuan:Rongyao.Grace:Okay — Honor. I think it’s called Honor, yes.Interesting. You know what? I really haven’t heard about it, but I should look into it. Is it actually already available to the mass market or no?Zixuan:I think the phone was launched last year, not this year.Grace:Okay, super interesting. I’ll look into it.Zixuan:Yeah, so a lot of phones at that time were powered by AutoGLM’s capabilities. But we don’t have the same bargaining power as ByteDance, so we cannot name the phone by our name. We just power their scenarios.So it’s about bargaining power, I think. Because like there’s the ByteDance vs Tencent issue, also with WeChat — it really depends on how you split the revenue, the value, how you make sure that you won’t influence other people’s business.So finally, you have a line: maybe this app will collaborate with you, and that app rejects your endpoint.Grace:Yeah. For context for you listeners, WeChat rejected Doubao phone’s direct access, and there was a huge headline war on this like two weeks ago.Okay, I want to pivot a little bit. Right now there’s a lot of focus on the China–US lens. And you yourself spent time in China and the US as well.But I did notice in the beginning days of Zhipu you guys were actually really focused on the so-called Global South — for lack of better words — Southeast Asia, Latin America, maybe even Africa. Is that still a strategy you guys are pursuing? Looking to sell or actually embrace markets that go beyond just China and the US?Zixuan:Yes, definitely. Because I think in GLM 2, GLM 3, we only had Chinese and English capabilities, but now we have more than 100 languages. So that can support us going beyond English-speaking countries. Maybe in Brazil, maybe in Malaysia, we have opportunities to showcase our model or showcase our product solutions to people and finally compete with those large enterprises.But I think things are really different in those countries, because they also want their data as private as possible. They accept on-premise, and maybe they want white label — they fine-tune the model and they want to ship it to their citizens under their name, not GLM or Zhipu’s name. So we have to meet their needs and see what we can offer.Doing business in the US, I think it’s much simpler because you have this API, you have products, you can do a coding agent, you can earn money. But when you do business in other countries, you have to go really deep, twist a lot of things, and try to make it happen.Grace:It’s also interesting — I think you touched on something. You’re quite comfortable being that white-label provider, versus I think a lot of other companies, whether it’s ego or belief, are not as comfortable. They definitely want their name on it.So it seems like you guys are actually the backbone supporting a lot of technology or clients without really having your name attached to it.I want to ask you about talent. This is a question we touched on in the beginning — you said yourself you came back to China for personal reasons, because your wife is in China. But I assume that’s not the case for everyone.There’s this interesting and funny joke going around saying right now in the AI war or AI race, it’s really between the Chinese in China and the Chinese in the US. It’s just funny — there does seem to be a high percentage of ethnic Chinese or Chinese nationals or Chinese-naturalized Americans or ABCs. If we’re being non-PC, people who look Chinese in the field.Why is that? I don’t understand. Did Chinese people just get a tip-off saying AI is gonna be really big early on and they went into this field earlier, or what happened?Zixuan:I think I cannot explain it, because doing math problems is simple for us. I’m not sure why other people won’t pursue this business.Because when I did internships and research at MIT, I saw a lot of talented people beyond Chinese — they’re still talented. They finally went to Anthropic, OpenAI. But somehow people only care about Chinese because they are co-launching products with Sam Altman or Elon Musk.I think people overrated the influence of Chinese people in the large language model area, because still there are a lot of enterprises not relying on Chinese.Grace:It’s quite funny — it’s kind of like the last generation, where every Chinese student in the US is either studying to be a lawyer or a banker, and now everyone switched over.Actually on a more serious note, how does the talent competition play out then? Do you see yourself at Zhipu having to really convince people to join you compared to a US peer?Or do you think there’s certain tendencies for certain researchers that would prefer to work for a Chinese lab or return to China? How do you see that play out?Zixuan:I think we finally choose the people that best match our environment. Like I said, we lack resources, but some people really enjoy the lack of resources — like me. Because I think it’s good to have a small team competing with a very large team, and you have better enjoyment when you conquer a puzzle or problem, or you finally win at the end.So people who enjoy this feeling, we try to hire them. And like I said, we want to move really fast. We want people — both the product team and the training team — to understand the user scenarios, to understand the data itself, not just theory or the algorithm.So we try to find those people, and they will finally choose us because they don’t care about compute or resources, or they find it too toxic competing with other teams doing the same experiments and the same thing. Because that happens a lot in large enterprises — a lot of teams doing the same thing.Grace:Yeah, for sure. I think even when I speak to the BATs in China, there’s so much internal competition that drives people crazy. It’s internal politics that drives people crazy. But that also becomes an incentive for people to really push.On that note, you guys are about what, 800 to 1,000 people altogether, roughly around 100 to 200 in R&D — something around that rough figure. It’s essentially not really a startup company anymore — it’s just small compared to how big the big tech incumbents are.So at this size, and as you guys head into becoming a publicly listed company, do you see the culture changing? And what are the ways you keep your researchers, scientists, and engineers motivated? Are we seeing crazy salary numbers as well, like the ones coming out of Meta? How do you keep people motivated?Zixuan:I think we are more lean, more entrepreneurial. Especially in our team, because I only slept 50 minutes for the past 24 hours. So we want to move really fast, faster than everyone else. Yeah… beyond that.Grace:You’re going beyond 996. This is not 996, this is 007.Zixuan:Because the competition is really fierce. Moonshot, Minimax — they’re doing an excellent job. And we also have DeepSeek, Qwen — not to mention the frontier AI labs in the United States. So we have to keep pushing. I think there’s no other choice.Because when we try to do AI, we want to survive. Frankly speaking, survival is a very high standard for the tech industry. When we look at operating systems: Windows, macOS, Linux — I think that’s enough. And when we look at phones — only Android and iOS.So the competition must be fierce when you want to be the infrastructure for the industry.Grace:Yeah, I agree on that. Okay, well, I hope you get some rest soon after this call. I really appreciate you jumping on the call after 50 minutes of sleep today.Looking at your long-term vision and where you guys are headed now, especially with an imminent IPO: in your public materials, you talk about AGI integration with the physical world and social welfare as a long-term vision. I think this is something not many AI companies frankly are really thinking about.Even within our conversation, you’ve talked about the balance between tech acceleration and actually putting up safety guardrails, essentially to prevent more sad, tragic happenings caused by AI psychosis, etc.When we look at this, how do you personally reconcile the social-welfare North Star with the commercial realities and the pressure you just talked about? Where are the areas where you guys are frankly more okay to let go a little bit for business gains? What areas are definitely your red lines that you cannot cross, where you really want to hold people accountable and ensure there are no AI-caused tragedies?Zixuan:Yeah, I want to answer this by giving an example. We have this GLM Coding Plan — it’s very cheap, three dollars a month. A lot of people in India, Indonesia, or even in the United States use GLM Coding Plan to do their side projects or even their startup.I just talked to a person today. He’s doing a startup that uses GLM Coding Plan to write a program that can collect recyclable bottles. They scan the bottle and recognize it, try to differentiate trash from recyclable products, and make it a real business. So we truly use AI to empower these businesses. You can see there is a lot of social welfare behind this.People just use the coding, but you can use coding to do a lot of stuff. We provide the service, but we let people decide whether they try to contribute more or only care for themselves. So I think it’s a starting point for us.With more powerful products — maybe next year — we can empower larger scenarios. Maybe we can empower robots.Grace:And in terms of this topic, I think in the US there’s a very dominant voice and discussion about the potential risk of AI or the negative impact it might have on society. AI Proem and Differentiated Understanding, frankly, very much focus on the business strategy of technology. So a lot of my guests and myself, we focus on capital deployment and feasible business models.But I do want to ask: you’re plugged in, you are in China, in the LLM space, in the AI space. Is there a discussion about AI safety, or are people really just quite focused on acceleration and pragmatic deployment and diffusion?Zixuan:I think compared to the United States, not that much. It’s more pragmatic. But that’s still on people’s minds, because AI safety is still an issue for us.We can see the ceiling, the threshold of all the current capabilities, and understand what’s the top priority for our model or our scenario, and try to fix those before going to the next step. But we always keep the security and safety issue in our head. And when that day finally comes, we can be fully prepared.I’m engaged in a lot of these conversations in the US and I’m also part of Concordia AI. It’s an organization focused on AI safety. I’m part of it in Beijing. But when I left that company, I saw them — and for anyone talking about this — it’s not because people don’t care. We can train a model with better capabilities and also a safer system.So there’s no trade-off at the current stage because we don’t have to balance performance with safety concerns. We can improve them at the same time.Grace:I see, it’s more like taking a mindful approach.I want to end on two quick questions. One is: usually people ask, “Where do you see yourself in the next five years?” Right. But I think for AI we can’t really ask that right now — no one will know what five years looks like.But for our listeners: where do you think China’s AI space will look like in, let’s say, six to twelve months? Where do you think the focus will be, or the potential breakthrough?Zixuan:Potential breakthrough may be integration with the physical world. When we see a lot of robotics companies and we see a lot of smart glasses, people are shifting focus from AI companies to these, we call it broader intelligence companies. So that might be a shift.And also DeepMind — I think they’re doing the same path. When you look at Gemini, it’s not just a large language model but also it can perform world knowledge or integrate with real-world use cases.When we look at Gemini 3 Pro use cases, someone is controlling the camera or trying to integrate with the computer. So there are a lot of things we can do with large language models.Grace:Okay, I think the last question I have for you is a question I ask every single guest, which is: what is one differentiated view you have — a non-consensus view? It could be about anything: about the industry, about how you see the world.Zixuan:I think for me, I’ll just share my thought: you should enjoy lacking resources — lacking people, lacking everything. In the AI world, that pushes you to the boundary. That pushes DeepSeek to change their architecture, to really do something innovative.For me, I don’t train models, but I build products and do marketing. I had this GLM Coding Plan thought because we don’t have very loyal customers. When they use API, they try to shift from GLM to someone else and then come back one day. So I noticed these difficulties. That’s what we aim for: to try to solve these really tough difficulties.Grace:Yeah, I think to your point — when you lack resources, it also means you have the agility and the flexibility to change things, because there’s no bureaucracies, there’s no chain of command, and it’s much faster.I appreciate that in itself too. I was talking to a friend about that recently as well. Since I left big tech and left traditional media to do this myself, you have so much more flexibility, and sometimes you’re upset that you don’t have the access or the resources you used to have. But it does help you build faster and connect with your community faster.Thank you so much anyway, Zixuan. I really, really appreciate your time. Please get some sleep after this.Zixuan:Thank you too. Yeah, I truly agree that you have your competitive advantage, because those large media companies — their journalists won’t reach out to me. So it’s my honor being here, but also a good opportunity for you to understand the Chinese market.Grace:Yeah, for sure. And I really appreciate you giving me your insights during this time. And for all the other Chinese AI labs out there, if you’re listening to this, please reach out. I would love to have a conversation. Thanks again.Zixuan:Yeah, thanks.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• What the U.S. Misreads About China’s Tech Rise with Kyle Chan 23.12.2025 52min

  In this episode, I sit down with Kyle Chan (Brookings Institution) to unpack the thinking behind his provocative New York Times op-ed, “In the Future, China Will Be Dominant, the U.S. Will Be Irrelevant.” We start with the DeepSeek moment and why it surprised the West, why it didn’t surprise many China-watchers, and why Kyle sees it as only “the tip of the iceberg.”From there, we zoom out into the bigger story: China’s rise isn’t just one breakthrough model or one champion company. It’s a system of interlocking capabilities: EVs, batteries, renewables, industrial automation, robotics, and AI, advancing in parallel and reinforcing each other through spillovers, supply chains, and fast-moving “Swiss Army Knife companies” like Xiaomi and Huawei.We also dig into what people often get wrong about China’s state role: not pure top-down command, but a mix of industrial policy + private-sector experimentation, including practical mechanisms like compute vouchers and local-government support. Finally, we cover India’s trajectory, geopolitical constraints, and Kyle’s “hedges”—scenarios in which today’s narratives (in both China and the U.S.) could still break in unexpected directions.Relevant links: https://www.brookings.edu/people/kyle-chan/In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.00:00 — The NYT op-ed + the DeepSeek catalyst: why Kyle wrote the piece, what he wanted to correct, and why DeepSeek was a wake-up call (“tip of the iceberg”).06:53 — Kyle’s origin story: infrastructure obsession (high-speed rail) → the path into tech & industrial policy.12:31 — China’s “electric tech stack” + spillovers: EVs, batteries, renewables, robotics, AI moving in parallel—and why “Swiss Army Knife” firms (Xiaomi/Huawei) can leap across categories.19:12 — Why autonomy pairs with EVs: the technical and architectural reasons autonomous systems “almost always” sit on EV platforms.24:01 — China AI ecosystem in practice: startups + hyperscalers + policy “tailwinds” (compute vouchers, industrial parks, local government support) and how that differs from the U.S. model.29:46 — China’s development playbook vs others + the India comparison: proactive bottleneck-solving (“ground game”), plus India’s tailwinds and constraints over the next decade.41:00 — The hedges + the wrap: what could derail or reshape the trajectory (trade backlash, geopolitics, bubble risk, robotics paths), and Kyle’s non-consensus take on policy intervention.AI-generated transcriptGrace Shao (00:00)Hi, Kyle. Thank you so much for joining us today. Kyle, I want to start with your recent New York Times op-ed, which had a pretty provocative headline. It’s called, In the Future, China Will Be Dominant, the U.S. Will Be Irrelevant. When I saw that, I was like, whoa, this guy, someone’s going to go get blood now. How did that piece actually come about? What was the main, I guess, objective or goal out of that piece?Kyle Chan (00:16)Yeah, yeah. Thanks for asking about that piece. Yeah, that piece, it got quite a reaction. I was surprised. And there’s been a number of pieces I feel like now—sort of, it’s almost become a genre of like all the things that China’s doing, all the things that the US is doing, the sort of divergent trajectories of the two countries, especially on technology. And for me, one thing I really wanted to focus on was—So we had the big DeepSeek moment earlier in the year, and that really got people to wake up and take notice of what’s happening in China and China’s tech development in a way that really, I mean, I can’t remember the last time that something like that happened. And so that was quite a big wake up call. But as someone following a number of different sectors in China for a while, I was like, this is just the tip of the iceberg.I mean, first off, within AI itself, DeepSeek—I think it was kind of funny—was surprising for a lot of people who follow actually China’s AI industry quite closely, because I think we might have been expecting some of the bigger tech companies to have made a bigger splash, but DeepSeek seemed a little bit out of left field. But within AI in China in general, there is so much talent, so much engineering talent, a vibrant developer community, top-notch researchers.So like when you look at, say, who is accepted, whose papers are accepted for NeurIPS, one of the top AI conferences, right? It’s many, many, many names from Peking University or Tsinghua University or Zhejiang University. And so if you follow that space for a while, DeepSeek was not so surprising. Maybe it was surprising that it was DeepSeek itself, but that China could produce a world-class AI model on par or nearly on par with some of the best in the US—that was maybe not so surprising.And then it’s not just AI. The other thing is like when you follow EVs or when you follow batteries, or if you follow anything related to clean tech—solar, wind, hydrogen fuel cells. If you follow robotics, anything related to industrial automation, industrial robotics, also self-driving cars, smart driving systems. I mean, the list goes on and on for all these different areas.And then it gets even down to sort of like the basics, right? So like some of these traditional industries where you just see this like classic China chart. Call it like the classic China chart where it’s like the share of global manufacturing for shipbuilding, say, or steel—is like, at first you see like China is growing and then soon it’s like eclipsing the rest of world combined.So to me, it was this bigger story that I really want to highlight: not just DeepSeek and not just AI China, but more broadly speaking, what is this bigger trend and why should we care? How is this going to shape not just Chinese society, but the rest of world?Grace Shao (03:11)I think to your point, DeepSeek was very secretive, yet it wasn’t like it’s within the AI industry in China—people were already noticing it and people were talking about, I think maybe six months before even they came out with their first R1 and then V1. But I think to your point, yeah, it was a shot to the West because it was like, wow, we always knew that China had strong industrial capacities, right? Like you said, like we had the manufacturing capabilities, the factories and whatnot, the hardware capabilities.They didn’t expect something like a software to come out of China that was almost on par with what they could produce in the West. I guess my question for you next is then to highlight that—what was your goal really? What was your real message that brought you public? Why did you publish an op-ed on the New York Times?Kyle Chan (03:55)Yeah, so part of it was to kind of point to the underlying drivers for what was happening because I also wanted to kind of correct this image of China, not only in terms of like China’s tech development, but also what really was responsible for some of that.So like the image I want to correct was basically this very old notion of China making, you know, low value added commodities like household goods, basic consumer electronics maybe—stuff that maybe is good for economic growth, but isn’t so impressive technologically and doesn’t really challenge, say, the US or Europe or other industry incumbents in these areas.I want to first point out that, yeah, this is different China now. And this process has been unfolding for a long time, actually. So I was trying to highlight some of the efforts that the government was trying to do to help accelerate not just industrialization, but innovation itself.This idea that it’s still so deeply controversial in the US—the idea that the government might have a positive role to play in supporting private sector development, supporting cutting edge technology—I think that that is still something that’s debated very hotly in the United States. And I wanted to point out how China has been able to use—not successfully every time, and there’s definitely issues along the way, but overall, quite effectively—it has been able to use industrial policy to really move the needle and support its industries and its private sector.And so this combination too of like, it’s not just one or the other. It wasn’t just sort of all top-down state driven and it wasn’t just all sort of bottom-up private markets. It was this interesting combination that has produced, I think, these sort of like world beating industries.And I think the lesson—a big part of the piece was about the US side of it and what lessons we might take away and how the US might need to step up its game. I don’t know if this competitive framing is the right one, but in general, a realization that, okay, there’s a lot happening. This assumption that China would always be the center of low-cost manufacturing and the United States would be the center of high-tech R&D, innovation, Silicon Valley—that the picture was much, much blurrier than that. So that was sort of like my overarching goal.Grace Shao (06:27)I definitely want to double click on the part where you talk about how the state and the private sector actually work together. And we can talk about that later. But I want to get a sense on what kind of feedback or pushback or even maybe criticism did you get from that piece?Then furthermore, I want to get understanding: how did you get involved in all of this? You’re in the US, right? You’re in New York, right? How did you get into studying China’s industrial policy? Tell us about your background.Kyle Chan (06:53)Yeah, yeah. Yeah, I mean, I just recently joined Brookings based in DC, which is a DC-based think tank that has a really great China center that does outstanding research on policy issues related to China, US-China relations. And my focus now is on China’s tech and industrial policy.But getting to this point has been like an interesting journey. So originally actually—I mean to go all the way back, I don’t know how far you want to go back—but like my family is actually from Hong Kong originally. And so I was born in California.And my parents—it was sort of that generation where my parents really wanted me to learn Mandarin. They’re like, that’s gonna be the useful language. And it turned out to be very useful.But also growing in California, I took cars everywhere. It was like a very, very much like a private transportation kind of city. And it was like a revelation to me—I mean, this sounds so ridiculous to anyone who’s grown up in a city with good public transportation—but it was like a revelation to me to later live in places like Chicago or even San Francisco and then later on Beijing and Delhi and Berlin, to be in places with like functioning subway systems and functioning public transportation.So I got really interested in infrastructure, actually, not necessarily industrial policy. That kind of came later, but infrastructure. And here there is really sort of like a very strong role for the government to play in coordinating, if not actually building and maintaining infrastructure, whether you’re talking about roads, highways, bridges, railways, subways, electricity grids.And I just found it really interesting then later on traveling to China—how this seemed to be like completely different there. I mean, and I remember at the time really being amazed by the high speed rail system there.And China didn’t have a bullet train system for most of its, most of the existence of its railway system until basically starting in the 2000s. They started to take seriously this idea of like, okay, maybe we can like really roll out a nationwide bullet train system. And they did a lot of R&D.And I became really fascinated by how they did this, how they built what ended up being, you know, within a decade, the world’s largest high-speed rail system, how they acquire the technology and how they built on top of that to create this sort of like truly like made in China kind of transportation system.And then how they did that repeatedly, not just for high-speed rail, but like for regular highways, expressways, for, you know, any kind of infrastructure.Okay, so that was sort of my foray into infrastructure and that was actually the focus of my dissertation. So I did field work actually for a number of years in China and also in India. I was based in Beijing and Delhi.And really it was like an enormous privilege to be able to travel around often by train across those two countries trying to understand their systems. And the railways were really useful for like understanding not just the railways and transportation, but understanding deeper political economy questions, like the structure of the governments, how their bureaucracy works, what are the main issues with building a mega project of that scale.And yeah, and so that for a long time was my focus. And then ironically, for an American audience, it was tougher to convince people that railways was interesting. I think most people were like, well, China can build high-speed rail because it’s the top down society and they just decide where to build and they build. I was like, no, no, it’s much more complicated than that. But it was hard to get traction.But then I realized like some of the same tools and the same patterns, some of the same institutions in China were also involved in boosting and accelerating development in key industries. So I mentioned clean tech, electric vehicles and batteries and solar, but also traditional sectors.And so I was really fascinated by this pattern that was, again, it kind of goes back to the New York Times piece. It wasn’t just one industry. It wasn’t just one company. It wasn’t just one state-owned enterprise, but this whole, like across the board effort to, you know, accelerate development overall.So that to me was so interesting—this process of industrial upgrading, which many, many countries are interested in doing. And that actually got more traction in terms of like, you know, I joke that the U.S.—every country in a sense is a developing country, right? There are areas where we’re trying to improve and areas where we’re trying to catch up.And I think now, you know, the question is like, what role can the right policies play in helping, say, the United States in a similar process? So that’s a long way of saying that it was—it was a long journey. But yeah, it’s an exciting time to study these topics because so much is changing. Every day it feels like.Grace Shao (11:30)That’s really good context and I think, you know, for me how I found your work was exactly because you had such a high level kind of view, a bird’s eye view of everything that was happening and how you were piecing it together.So I believe I came across one of your pieces on High Capacity, which is your newsletter on Substack for audiences who don’t know. You had this Venn diagram where you’re like, this is what China’s good at here, here, here, here, here. And this is what’s happening right now. And this is how it like actually relates to the current EV build out, the renewables, the AI, the robotics. It’s a very big ecosystem.And in some ways you argue that, you know, all these different sectors operate in parallel, whether is, you know, a top down direction or a directive, or it was organic, you know, growth, but they did grow in parallel. So therefore they’re now able to kind of find synergy and leverage each other’s strength, right? With LiDAR sensors, drones, robotics, all coming together.Could you tell us a bit about what that diagram really means? I’ll try to pull it up as well in the video. I think just help us explain that in a high level.Kyle Chan (12:31)Yeah, yeah. So what I try to capture with that diagram was this idea that it wasn’t just one sector or wasn’t just one area, one technology that China had been able to grow and foster maybe through industrial policy or some kind of state support, but it was this combination of these interlocking technologies.Now there’s a new term that’s coming into vogue, like the electric tech stack, or the tech industrial stack, or the electric industrial stack. And it’s really interesting because I think that really captures sort of this new paradigm that we’re entering.Those technologies—you mentioned a number of them—electric vehicles, autonomous vehicles, which for various reasons we can get into are built on electric vehicles. There’s strong reasons why those technologies go together. Lithium batteries, which also feed into drones, autonomous delivery systems.We can think about just regular consumer electronics, smartphones, but also more sophisticated robotics. I mentioned industrial automation as well. There’s a lot of overlap there.And then the big circle overlapping all of these is AI—different models, software platforms that might intersect with say autonomous driving and, I don’t know, even sort of the humanoid robotics world.So I just find it so interesting that China was making progress in a number of these different sectors at the same time, and progress in one sector would support greater development in another sector. So EVs and EV batteries grew up together in China.So the development of, say, lithium iron phosphate batteries that were increasingly inexpensive, that were increasingly energy dense on a, say, kilowatt hour per kilogram basis, that were safer—those developments within the battery world made Chinese EVs more competitive and more attractive overall.And then developments in China’s EV sector fed back into the battery world and also fed into other related sectors.And then the other big thing is that I really wanted to highlight companies that lay at the intersection of these different areas. So I think probably right now, one of the hottest companies is Xiaomi, right? When I was younger, Xiaomi to me was inexpensive smartphones and relatively like affordable, like household products, like air purifiers and things like that.And I think they had built up a brand and a very strong sort of customer base around this general idea. And what was so fascinating was seeing Xiaomi jump into electric vehicles and having such success with the SU7 and now the YU7 SUV, which both of which were like sold out for a long time and are very much in demand.And they have like incredible features, they have incredible performance, and they have smart driving capabilities. And so it just sort of like showed like, wow, this company that was originally kind of like a smartphone company could make this shift over into the EV world.And I would argue that it wasn’t just Xiaomi and Ledron’s entrepreneurship. I mean, a lot of credit goes to them for sure, but it was also because of this broader foundation that existed in China that allowed for this common supply chain ecosystem that would feed into these different worlds that would allow a company like Xiaomi to make that pivot.And you see it again and again. I mean, now you see—and this is where I came up with this term like the sort of Swiss Army Knife companies—now you see these companies like XPeng get into like a whole range of industries, right? So not just EVs, but also humanoid robots, drones, flying cars even.Huawei is probably the ultimate example of the Swiss Army Knife company, originally starting in telecom equipment, but then branching out into everything from sort of every aspect of consumer electronics—smartphones, tablets—into now AI chips. They’re a major player in semiconductors. For a while undersea cables. I mean, the list kind of goes on and on—EVs as well and smart driving, autonomous driving.There was just this like burst of companies coming out of China that could do all these different things and branch out into new areas very, very rapidly. And it was like shocking to me how big some of these bets were, like Xiaomi making a multi-billion bet on a new already highly competitive industry, the EV industry.But again, I think it all comes back to, in part, this foundation that was there in China—this like what I call these overlapping tech industrial ecosystems.Grace Shao (17:17)I have so many thoughts I want to throw out. One is, I think to your point on Chinese tech companies going into EV, it’s really fascinating because to your point, there’s obviously this price war and this crazy competition in China.But what I’ve heard from a lot of people who actually do purchase the Xiaomi cars and the Huawei cars and the Xpengs—they say that the technology itself is incredible. You have all the lights, the voice control, you have amazing AI-empowered functions. But that said, they’re not actually as good of a drive, like they’re not as smooth.So like if you pick a traditional OEM like a Mercedes or BMW, their voice control usually apps the crap, to be honest. Like they go off—like you see the reviews online because we were looking at family cars and it was like the reviews were horrible. Like they just get triggered by really random sounds, they can’t pick up like accents, you know, they’re not really good with other languages.And on the Xiaomi/Huawei/Xpeng/Zeekr side, they’re really, really good at this. But they’re not as good for driving yet. So it’s interesting that, like you said, what they’re good at in terms of day as the Chinese companies are the kind of technology that is quite recent and quite modern, but they’ve not really actually honed in on the craftsmanship or the, I guess, capability of building a really, really smooth driving car as Germans have—as they’ve honed the skill for over the last like four or five decades, right?So it is interesting where they’re good at. And then in terms of EV, I had a question you mentioned earlier. Most autonomous vehicles are now—sorry—most EVs are now being tried for autonomous driving. Why is that? What’s the synergy there? Why can’t old OEMs actually have strong autonomous driving functionalities?Grace Shao (19:00)Kyle, one thing we were talking about is the synergy between EVs and autonomous driving. Why is it that it’s better to actually build in autonomous driving functionality within EVs compared to like maybe traditional OEMs?Kyle Chan (19:12)Yeah, so there it’s basically because you get a lot more control and precision. And you can have things like steer by wire where rather than sort of mechanical steering, you can have basically a signal be sent directly to the transmission or directly to the engine or directly to the brakes. So it’s sort of all.And then on top of that, it’s helpful to have large battery capacity to handle sort of all of the different computing demands that would—including all the sensors that might be feeding data into the whole system.So yeah, you basically—the two almost always go together: having autonomous robotaxis built on top of EVs.Grace Shao (19:51)That makes a lot of sense actually. Okay, I never thought of it that way. You described China as building systems of capabilities and you kind of touched on this earlier with your Venn diagram. You talk about how China is not really just picking one winner or one winning sector, right?So how does actually EVs, batteries, renewables end up supporting each other? And then how does that actually extend out and spill over into the AI era with robotics, physical AI, or even the consumer AI products we’re seeing out there today?Kyle Chan (20:18)Yeah. So at one level, there’s an underlying driver here that’s almost not even specific to China per se. It’s something more about these technologies themselves and this broader convergence across them.So, I mean, I pointed out Swiss Army knife tech companies in China, but to be fair, right in the U.S., you have companies like Tesla that are going into many different areas, or even Google with probably one of the world’s best autonomous driving companies.So I think there’s something deeper happening here where there is this convergence of what we might have thought of sort of like lower end, like smartphone technologies or consumer electronics. Again, the lowly battery, right? Something so simple. This innovation in lithium batteries, making them cheaper, more reliable, and being able to scale up production—like that alone unlocked so much in terms of basically any kind of electronic device.And so I think that’s why we are kind of seeing this emergence of this like whole new technology cluster.And then for China, I think there is an awareness of the sort of synergies across different industries. And you can even go back further to China’s earlier industrial policy efforts, right? Take Made in China 2025, which came out in 2015.And some of the target industries there were chosen not just because they might in and of themselves be useful or important, but also because they had broader spillover effects. You think about things like telecom equipment, IT infrastructure, anything related to energy, or anything related to communication in general.Also CNC machines, right? So these are sort of like—they may not be like general purpose technologies in the way that electricity itself or computers are, but they might be sort of like multi-purpose technologies with broad applications in a range of different areas.And so by making that bet on those types of technologies, you know, whether or not you succeed in becoming a global leader in that area, it will help feed into everything else that builds on that kind of tech stack.So that’s what I see happen again and again, where for China’s approach to technology, where it’s not just about a single bet on a single technology, but trying to find these parts of the value chain that have large spillover effects and trying to support those—even if they in and of themselves might not be totally economically viable or the best businesses to invest in from a pure return on investment perspective, but they have these broader economic and technological spillovers.Grace Shao (22:57)Yeah, and I think to your point on like a lot of the planning from top down, it’s not like they were just more strategic in like picking the right track.But when I spoke to David Fishman a couple of weeks ago, he was saying China’s strategic planning on building electricity capacity is actually not because they foresaw like AI, the AI boom and data centers. You know, electricity is just simply urbanization actually was driving increase of energy demand as well.They knew that in the future, a lot of things had to kind of move from traditional coal to renewable to kind of actually even have the capability to power what is needed of the future. So it was a grander vision versus just like, I know AI is gonna come in 10 years. I’m gonna build up renewable energy and the renewable energy is gonna help power data centers. So it’s not like there are profits or anything. So that’s interesting to hear.I think I wanna understand how has that shift in ambition with like—Kyle Chan (23:42)Totally.Grace Shao (23:49)really China’s desire to move away from low wage, low margin, that trap into like higher value services and really like how has that shaped and driven the AI innovation that we’re seeing right now coming out of China.Kyle Chan (24:01)Yeah. So I think what’s really interesting is, on the one hand, you have like a lot happening within the AI sector itself. You have obviously this like very vibrant ecosystem of startups, of big tech companies, even down to servers and data center construction firms, and even the major state-owned telecom operators involved in data center construction.You have all these applications and developers trying to build on top of this whole set of foundation models, for example.And that’s sort of just within the AI industry, right? And then on top of that, you have the fact that in all these other sectors, whether you’re talking about manufacturing, you’re talking about biotech, healthcare, there’s a lot of investment and progress in trying to move up the ladder in each one of those industries.And so then you see people finding ways—either from those industries or from the AI side—trying to find ways to incorporate, to integrate AI.And actually there’s something that I really love about your work where you’re highlighting those areas where it’s not just about like the latest benchmarks on the latest models and this sort of like endless race, but about like, how is AI like actually being deployed? How’s it being integrated into existing services and platforms in a way that would really boost, say, drug discovery or would really, I don’t know, improve tutoring and education services for students.So I think that’s what’s so interesting here. And yeah, some of it might be supported by policy efforts. I think of some things like compute vouchers, for example, where startups—AI startups—might have trouble getting access to compute.So we’re not talking about like the Alibabas and Tencents of the world. We’re talking about the little guys who may not be able to afford to build a giant data center dedicated just for them. And then local governments might offer compute vouchers—subsidized compute—basically access to public infrastructure, essentially, to help them sort of get off the ground and have that little bit to deploy on and develop with.And so that’s an example of an area where you do have some government intervention stepping in and not trying to do it in a heavy handed way, but just trying to offer kind of like a tailwind of support.And ultimately, and this is one thing that I think is sort of special about software and the AI industry in general is, I mean, a lot of this is sort of like, you know, this creative explosion of different ideas from the private sector—from all these entrepreneurs—like trying to look in areas that are related maybe to their own areas of expertise or just kind of like scouring: where can we plug in AI? Where can we make improvements, even very small ones into existing industries?Grace Shao (26:45)That’s really interesting on compute bit, where I just met someone at Google in Singapore a couple of weeks ago, and he was saying that in some ways, the big tech in the West are actually operating in that capacity. Instead of incubating them and just taking another part of equity, and instead of just giving them capital, they’re giving them compute, essentially vouchers for these startups.So I guess my question is, how do you see the relationship between startups and big tech in China versus startups and big tech in the West? And in what way, I guess, in what way do the state actually play a positive role or negative role in all of this in the whole ecosystem?Kyle Chan (27:17)Yeah, that’s a question. I mean, in some ways it is a similar story, right? You have China’s own hyperscalers providing AI cloud computing services to a whole range of different players, you know, be they AI startups or, you know, large corporations or other, you know, maybe hospitals or other state-owned enterprises, for example. So that part might not be so different.But I think what might be a little bit different is the sort of like extra on the margin support that the government in China—or especially local governments in particular—might offer.Yeah, compute vouchers is one. Also these industrial parks where—and this is going back to like an almost an older model applied to like the age of AI—where, you know, AI startups, they still need office space. They still need help setting up a business. They might need help figuring out how to network with new customers.And those are other areas too, where local governments in China might play a more active role in troubleshooting, trying to bring startups up to speed, trying to connect entrepreneurs.And that’s something that I think is quite different than in the US system, where yes, you might have the large hyperscalers like Google or Microsoft providing that underlying infrastructure, that service for access to compute, but you won’t really see that kind of intervention or stepping in from the local government side, at least not so proactively by any measure.Grace Shao (28:46) Yeah, I think definitely the West what you hear more about is like applying for fellowships or acceleration programs within whether it’s VC funds or like you said the big hyperscalers. Whereas even in Hong Kong here, like the Hong Kong government offers startups like staff support, back office admin sharing, teams to share, then like even offices in like Cyberport out like, you know, in Pok Fu Lam.So like definitely the state plays a more active role and I think it’s kind of sometimes misunderstood by the West what that role means. It’s really many times it’s just like an incubator, a parent to someone, or even a mentor.So we talked about you living in many, many different cities across the world and you study industrial policies across different developed economies. You mentioned that you lived in India, then you studied obviously China. So from that perspective, that international global perspective, what do you think China has actually done differently compared to maybe other developing nations that started in similar circumstances maybe say three, four decades ago?Kyle Chan (29:46)Yeah, that’s a great question. So the thing that China has done that really has stood out to me, that really makes it so different than I would argue most other developing countries, is a very deliberate effort to basically build up industrial capacity and to move up the value chain.It wasn’t like, you know, if we invest in education, if we invest in sort of these general factors that go into development, that over time, you know, you would eventually sort of get there. It was like: how can we bring in foreign companies to form joint ventures with our own domestic firms and share that kind of knowledge? How can we build up world-class research programs and build interesting scientific collaborations with the Europeans or the Japanese or the Americans even?It was like: how can we try to like find those bottlenecks in the process?And I think this kind of goes back to your point where it wasn’t like, this is the one direction we’re going to make this huge bet and that’s what turned out to be correct. It was more of these sort of— you know, to use like a sports analogy—like kind of like the ground game, right? It was like oftentimes kind of more tactical things trying to backfill areas.Like let’s say for batteries, right? You need to have access to lithium. And so building a global network of lithium processing facilities and supply chains was really crucial to feed into EV batteries and then the EV industry itself.And yeah, so I think that’s something that I see other countries do like a little bit, but really for China, it’s sort of at a very deep level: this effort to not just sort of hope that you got most of the pieces right and then let the story unfold, but to proactively find ways to support industrial development and technological development.Grace Shao (31:44)What was actually, more personal, what was the most memorable thing living in Beijing and then commuting in New Delhi and then moving around the world so much over the last few years?Kyle Chan (31:54)Yeah, I mean, there’s so many things. Yeah, I mean, I can tell you from a research standpoint, I interviewed government officials in both countries. And I can tell you that it’s much easier to get access to government officials, at least at the central government level, in India.And so some of my sharpest memories are of long conversations over chai with railway officials in India, where I almost was kind of like a therapist for them in some ways, because they would have these complaints about the bureaucracy, about, frankly, their colleagues, about many thoughts about their country, about China, that they were just like very generous in sharing with me.And it was like really fascinating to kind of like see the world from their perspective. I think in particular in India, there is a bureaucratic elite, there is a civil service elite who are highly educated. They often have to go through extremely competitive national exams to sort of like get to their positions.And I think in some ways, I found this group of people to be both very proud, but also oftentimes very frustrated by some of the bureaucratic hurdles that they faced.On the Chinese side, yeah, like when I did get access to government officials, sometimes even getting access, right, like it might take a while to get to like a genuine conversation where people can kind of open up more.And it was so interesting because like I have my own theories about why say China was able to build high-speed rail so quickly, but it was always interesting to hear people’s own perspectives. To see like, what did they think was like really, you know, the thing that moved the needle.And I heard like sort of everything from, you know, that China is just a much more sort of like coordinated and aligned system across the board to cultural explanations to, I mean, you name it. But it was always fascinating to hear like from people within the system.So like to the extent that I was able to get that, and then I got to visit—I got to actually visit like the construction sites for like ongoing railway projects. And that was really cool.So yeah, I mean, there, there’s so many. Like the two countries are so fascinating and you could spend a lifetime in either one and feel like it’s not enough, in terms of exploring and getting to see different parts.Also, I did get to travel a lot. And like India, I don’t know if many people know this, like India has like the whole Northeast region, which is sort of like very distinctive culturally in terms of geography. It feels very different from the rest of the country.And, you know, like you could explore that whole area, or you could travel to the South and there’s like a big sort of North-South cultural divide that anyone from the North or South will tell you about.So yeah, it’s just like—I mean, these are sort of like continental-size countries. And I think they have that kind of continental-size complexity to them.Grace Shao (34:51)Definitely, I would love to visit India one day.Do you think what we’ll see India kind of become the next China, if you must put it that way? Because for many years people said, oh, Vietnam’s going to be the next China, right? And obviously all the talent, all the capital, all the interest in that right now moving to India, as well as even a lot of the supply chain, right? For a lot of the big companies in Europe as well as the US, what should we expect of India in the next maybe decade or so?Kyle Chan (35:14)Yeah, that is like the big question. And I will highlight some factors that are very much in India’s favor, and then I can point out a few challenges.So in general, the India story is like really fascinating because if it weren’t for China’s high growth story—like if you kind of remove that from the equation—India would have been the envy of the world.I think in many ways they have already proven to be able to have high rates of sustained growth over multiple decades now. So like in that sense, they’re already getting there in some ways.The other things that are really working in their favor are: there is a big focus on manufacturing, on trying to build up industrial capacity in a way that is very reminiscent of China. There’s a huge push in industrial policy targeting areas like consumer electronics, the automotive slash EV industry. There are big ambitions for the semiconductor industry in India.And the third thing I would point out is that there, up until recently, was a bit of a geopolitical moment for India with different companies trying to move away or diversify from China to some extent, diversify their supply chains.And this was really quite an opening for India where you saw, like for example, Apple was able to shift a fairly large share—I think up to like a quarter—of its iPhone assembly was shifted to India. And with that also began a process where some of the suppliers would start to follow. And so I think that was actually honestly, to me, surprising how quickly that happened.So those are some factors in their favor. But on the other hand, there are some very deep structural constraints. And one is that the bureaucracy, especially around anything related to labor, is very, very tricky to navigate.And I think for companies who want stability, who want sort of like a certain stable business environment, it’s harder to operate, and there’s a lot of regional variations. So some of the southern states like Tamil Nadu, Karnataka tend to be seen as more business friendly, open to foreign investment, open investment of many different kinds. And that is where you see the electronics industry really growing quickly.But overall, there are some bureaucratic issues. And then at a more fundamental level, something that China has been able to do is often reform its own internal organizational structure—sometimes pretty quickly, sometimes shockingly fast.And sometimes it takes a while, like the railways did take a while, the railway ministry. But for India, I think there are some problems that everyone knows exist in a policy sense, but nobody knows how to address and break, say, like a political deadlock. That happens a lot. It probably also happens—I mean, I know it does happen in China as well. But I do see it more clearly and especially in the areas that India is trying to target for growth.Grace Shao (38:10)And do you see actually a lot of the Chinese companies we talked about earlier, like the Xiaomis, BYDs and Huaweis of the world—are they exporting to India or even exporting their know-how and their supply chain and manufacturing? Because I know a lot of these Chinese companies have been doing that, like moving their talent, their know-how and also selling to consumers in Southeast Asia. But what is kind of the South Asia market looking like for these Chinese companies?Kyle Chan (38:35)Yeah, I think from the Chinese companies perspective, they would be eager to reach one of the largest markets in the world, right? And then South Asia more generally.And I think you see other areas like Pakistan and Nepal—Chinese, say, EVs or other smartphones or other consumer goods—like really taking off.India and China specifically, though, have a bit of rocky geopolitical relationship. And so that definitely cast a chill over cross-border investment and business flows for a while.And more recently, it seems like things are warming up, maybe gradually, and I think this is taking years. But I think there’s a recognition, certainly from the Indian side, that it helps to bring in that expertise and know-how—to have Chinese engineers, technicians, and managers participate in India’s own industrial development process.I think, I don’t know if we’ll see those days again when you had, say, Alibaba invest like, you know, hundreds of millions in PayTM, which was India’s sort of equivalent of Alipay. I don’t know if we’ll see those days again, but maybe there is a sort of warming up of relations between the two countries and we might see greater sort of business flows following that.Grace Shao (39:48)Yeah, because it seems like India has their own ecosystem of like FinTech as well as their own ecosystem of social media. So it’s not like they really need China’s software export.But I think on the hardware side, it was really interesting reading Patrick McGee’s book and he was saying how back then it was the Taiwanese and the Americans that actually had to—well, the Americans first trained the Taiwanese and the Taiwanese came to train the Chinese.And now it seems like if the supply chain is moving over to India, I’m sure the business will be business and, you know, it would make sense for the business to actually have the Chinese now train the Indian factory workers to really take over that new labor demand, right? Yeah, but geopolitics aside, that is.Let’s talk about AI quickly. You know, in your conversation with Heath Yap—also I love his show and a good friend of mine—really grateful for Keith connecting us actually.In your conversation with Keith, you said that you’re always careful to hedge. I kind of chuckled when I heard that. You’re talking about China’s rise and China’s rise in AI. You’ve obviously been quite vocal talking about how China has done some things right. But yet, obviously, as an American, you are saying it from a perspective of what we can learn as Americans.And what are the scenarios where your own projections—China’s technological rise—don’t pan out or that could derail this story, right?Kyle Chan (41:00)Yeah, that’s a really good question. So there are some ways that this story can turn out differently.I do think that China has been able to benefit for a long time from especially international partnerships. And I think the extent to which there might be skepticism or even suspicion about those sort of international partnerships—that could affect, say, research collaborations or business partnerships across border.At the same time, though, I guess at this point, I can point to areas outside of AI where there is a lot of interest, say, from German or European automakers in partnering with Chinese ones, trying to learn some of that know-how and technology. But that’s a big wild card.And then another big one is also the extent to which there is a bit of a trade backlash, right, to Chinese exports and to what extent other countries might push back and maybe put up more tariffs or maybe even just demand greater investment and localize manufacturing in their own countries. So that could change things, not necessarily in a worse way, perhaps, but it could change things in some ways.And then I do think that the relationship with the US is really crucial. And I think for China’s development, the instability in the relationship—I mean, it doesn’t help anyone, I think. So that’s sort of another wild card in all this.And then from the US side, I mean, it’s sort of interesting to think about American AI development right now, because I literally like every day my feeds are like talking about bubbles and talking about like, you know, circular investment deals and things like that.And at the same time, like there’s still so much enthusiasm and the data center build out is like seems to be still going full blast. So here’s like where I will offer like a hedge perspective where it could go in many different directions.It could be a bubble and we might see a pullback and it might turn out that these investment deals just don’t pencil out. The economics just don’t work out.But it could also be that we’re not—that this talk about AGI, I don’t know if like AGI per se will be achieved—but that broad-based artificial intelligence like computers that can do a huge range of tasks, not just coding, which is already impressive, and not just writing and things like that, but really sort of a broad range of tasks—that could take off.And the United States huge investment in compute capacity could give it an edge in the long run.So those are different ways things could pan out. And then the other wild card is like embodied AI and robotics and how that will pan out.Because I think there’s like different approaches being taken by different companies, by different countries. I think some companies are looking to a version of AGI for robotics, kind of a universal foundation model for robotics that can operate across any different sort of hardware platform.And other companies—I think about like Unitree—are focused on just deploying fast, like building tons and tons of like quadruped and humanoid robots now that have become kind of like the new hardware platform for many robotics developers, including in the United States.So there’s many different ways things can go.And then like more recently, there’s like, I have this tweet about like highlighting like street cleaning robots and just like very practical, like immediate real world use, right? So that could be another area where, you know, maybe we won’t get the AGI robots, but maybe we will have many, many different types of robots doing more specialized tasks.And that could be transformative for either country. These are all areas that I’ve followed closely. Yeah, at this point, if I had a crystal ball, maybe I would be in a different job.Grace Shao (44:57)I think he definitely touched on something I feel like it resonates with me. Like is when I speak to people in the US—not really just the West, but really just in Silicon Valley—I feel like exactly to your point, there’s such a polarized kind of take on where AI is going.It’s either it’s such a bubble or it’s going to burst. It means nothing. This is complete fad, right? Or it’s—or even, you know, people are like, this is going to ruin our lives. We must stop it right now.Or it’s a very polarized view, which is like, OK, this is like the best thing that’s ever happened to us, like, you know, AGI or nothing.And in China, when people ask me about what’s happening on the ground, it does feel like—I’m not sure pragmatic is actually the right word anymore. I don’t like using that word anymore because I feel like it’s now overdone. But I do feel like it’s a bit more, I would say centered.As in everyone kind of feels like it’s just another wave of technological advancement. Whether how it might pan out—to your point—it might not be AGI, it might not be this crazy intelligent being that will take over our thinking capabilities, but it might be something that will transform how we humans interact with each other, how we work, how we actually increase productivity.And I think it might have some—it might lean on the fact that frankly China really did see this kind of technological advancement driving productivity gains only just recently, within the last generation. Whereas in the US, you haven’t really seen this kind of crazy gain for the mass in a while, right? Since the Industrial Revolution. So people don’t feel it as much and there’s more fear around it.Now I have another question on this. Just to kind of wrap everything up: you look at a lot of the industrial policies and you look at how the state works with the private side in the AI sector right now in China. How much are we seeing that is actual genuine market pull versus state-driven push?I think you kind of alluded to the fact that like Chinese government does work with helping talent and raising capital. But right now with the big names—like the Unitrees, the UP techs, the deep CXC, Galabots and the whatnots, right?—that we’re seeing, are the state behind this or like the West often seems to think that way. How do we understand that?Kyle Chan (46:56)Yeah, so I think it’s a complex relationship. So I don’t think it’s sort of like a top-down, the government is sort of like dictating what direction these companies need to take.And I think there are probably national priorities where I think Beijing or local governments would want to see companies focus on more.And ironically, like there’s a very strong convergence right now on wanting to make progress on AI more generally from the public and the private sector.And then I do think that some of it comes down to problem solving on the ground. So this is something we talked about earlier: trying to figure out what issues the private sector is running into, what issues certain firms are running into, and then trying to troubleshoot and support them in some of those areas.And then, yeah, overall, sort of like offering this broader roadmap or this broader set of like this broader package of sort of like a policy support, whether it’s sort of like a robotics-specific national strategy plan.And part of it is just about even signaling that there is state support for these sectors, that these are areas where if you venture in as a private business, like you will not have all your problems solved, but you could at least have ways to have some of your sort of like more day-to-day issues addressed.Grace Shao (48:17)Kyle, if you look at the AI sector today in China, how much do you think we’re seeing genuine market pull versus state-driven push? And especially now we’re also seeing the AI plus policy being pushed out. Is that diffusion driven by the state or is it driven by the actual entrepreneurs, especially amongst the Unitrees, the UP techs, the DeepSeeks and the MiniMax, Moonshots of the world?Kyle Chan (48:39)Yeah, so that’s a great question. And I think what’s interesting about this moment is there’s a strong alignment between the public sector and the private sector, where clearly Beijing at a national level, and then many of the local governments want to see a booming AI industry in China.They want to see China as a whole make progress in applying AI to all different parts of life—support economic growth and improve social services and improve education and all these good things.And at the same time, you have so many different private sector firms from the big tech companies to the smaller startups really jumping in and also eager to innovate.And I think one thing that’s really sort of exciting about all this is you can really like drill down into like subsectors or different aspects of China’s AI industry and just see how vibrant the startup ecosystem is now in terms of all the different, say, new coding tools that are emerging or different efforts to try to expand overseas or interesting ways to integrate AI into e-commerce or social media or EVs or advanced manufacturing—you name it.And so right now, I just think that there’s just like this huge sort of creative explosion of different ideas and an eagerness, especially in this sort of post DeepSeek moment, to try new things and take that risk.Grace Shao (50:04)Super interesting. Kyle, I really appreciate your time today. And for listeners, we actually had a few technical glitches today. That’s why we’re going to have to chop up the conversation a little bit. It might sound a little jumpy than my usual podcast.My last question for you, which is a question I ask everyone: what is one differentiated view or non-consensus view you hold? I guess your New York Times piece was a pretty strong opinion and you really publicly put it out there. But is there anything else you think that, you know, you think differently or you have a different kind of view on that compared to your peers maybe?Kyle Chan (50:35)That’s a question. I do think in general that my take—maybe it’s becoming more mainstream—that there is a role for very creative and very thoughtful policy intervention in supporting strategic sectors, in boosting technological innovation.I think that’s something maybe perhaps widely accepted in other countries, but at least in United States, still something that a lot of Americans are not quite comfortable with.And there are certainly risks along the way, but I think also to keep in mind, there are risks of not doing anything and of not trying to support R&D, scientific development and all these different things and to leave it purely to the market. That’s sort of my slightly left field take.Grace Shao (51:21)Thank you so much, really appreciate your time and your insights again. Thank you, Kyle.Kyle Chan (51:24) My pleasure. Get full access to AI Proem at aiproem.substack.com/subscribe
• AI Governance From Brussels to Beijing: George Chen on APAC’s Different Path 16.12.2025 52min

  Most AI policy conversations still orbit around Washington and Brussels, but Asia-Pacific is already writing a very different rulebook. In this episode, I talk with George Chen, Digital Partner at The Asia Group and former Meta policy executive, about how AI is actually being governed, built, and deployed across APAC, China, and the global south.George traces his own path from journalism to big tech to advisory work, and uses that vantage point to explain why APAC is not “one market”—and why the EU analogy breaks down almost immediately. Countries like Japan, Korea, Singapore, and China are leaning into AI as a tool for economic recovery and industrial upgrading, often taking a much more pro-innovation, pro-growth stance than the EU’s more precautionary approach. At the same time, Southeast Asia is becoming the physical backbone of the AI build-out: Singapore as HQ and regulatory hub, with Malaysia, Indonesia, Thailand, and the Philippines hosting the data centers, power, and connectivity—along with all the local tensions that come with that.We also get into what “responsible AI” actually looks like inside a company. Beyond the buzzwords, George breaks it down to three pillars—security, safety, and privacy—and talks through how mature players like Microsoft or Meta build these into product design from day one, versus the reality for startups trying to ship fast with one lawyer and a single policy person supporting multiple markets. He also makes the case that fragmented regulation and the lack of international standards are becoming a real tax on innovation, especially outside the US and EU.Another big thread is the emerging US–China competition over AI governance itself. It’s no longer just about who has the best models or chips; it’s also about who exports their rules, norms, and defaults to the rest of the world. The US is pushing an “America-first” innovation and safety model to allies, while China is pitching AI as a kind of public good to the global south—combined with a more cost-efficient, top-down deployment model and very strict cyber and real-name rules at home. George argues this divergence is already shaping how content, deepfakes, and AI-generated media are treated in different jurisdictions.We talk about the local edge of Chinese models—why in places like Beijing, models such as DeepSeek can be more useful than ChatGPT or Gemini for everyday queries because they’re trained on more localized, timely data. From there, we zoom out into the new AI talent map: countries like Indonesia, Vietnam, Kazakhstan, and Uzbekistan trying to position themselves as low-cost AI talent hubs and “back offices” for global AI companies as coding gives way to prompting and applied ML.We close on a more philosophical note: should AI be built as a subordinate assistant or a true partner? George shares his uncertainty here, and we talk about what happens when we give AI more agency, emotional intelligence, and continuous workloads. At some point, the conversation shifts from safety checklists to ethics, culture, and even “digital colonialism”: whose values, whose norms, and whose worldview are encoded into the systems that end up mediating how we see the world.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.AI-generated transcript.Grace Shao (00:00)Hey George, thank you so much for joining us today. I’ve been really excited and waiting for this chat. You know, you are a very busy man. You’re constantly traveling. I can barely reach you in Hong Kong. So really appreciate your time today. Sit down with me and share your insights with my followers and some of our listeners. To start with, you’ve worn many, many hats. A journalist, tech executive, policy advisor, and now a partner at the Asia Group where you advise a lot of force, you’re probably helping companies on, I believe, geopolitical positioning, right?George Chen (00:29)Thank you. First of all, thanks for the invite. It’s quite an honor to join a growing cohort of guests for your program. Really happy to have a discussion about tech and policy issues because I think you’re right. My first 10 years in media, similar to your background, and most recent decade, I work very much on the intersection between technology and policy.My biggest takeaway from my last job at Meta, one of the platform operators in the world, is sometimes we very much focus on technology development, like the breakthrough, while the resources for policy support are actually quite limited, especially in the Asia-Pacific region compared with the US. think for all the...Big tech in the US, given the politics domestically, they have to do a lot on political and policy part. But for Asia Pacific, the policy work, compared with other investments, like in data center, technology, hiring of engineers, it’s still very, very, very understaffed, under-resourced, and sometimes under-appreciated. This is why we need to...address some concerns about policy issues as we advance the technological part. Because I always tell my students, tell my friends, tell my partners that the key challenge, even you have CharGBT 5.0 or 6.0, the key challenge is how to get the government to understand new technologies and also get the users to have more trust in those new technologies. Otherwise, nobody use it, nobody trust those things. And that makes them.Grace Shao (02:15)I think that’s super helpful. A lot of times when we think about policy or safety issues, we think about it as like a siloed part of the ecosystem. But really like exactly to your point, like, you know, we need the developers to understand the concerns of the users. We need the users to understand the safety risks of the products. We need the regulators to understand what it means to implement these like technology throughout our economy, right? So there’s it’s like, it’s actually all interrelated.I think today to start off with, let’s like go into big tech, just give in your background with Metta, working with a lot of these big tech companies. You’re based in Hong Kong for the listeners, but actually work predominantly for American big tech companies. What is like the, I guess, the fundamental feel right now as we see the evolution be from a social media company for AI to AI of focused company as this is now the forefront of their strategy.George Chen (03:11)Right, so for the Asia-Pacific region, it’s big. I always try to explain to my clients and friends, when people talk about Asia-Pacific, the first gross perception, perhaps from Western perspective, is, okay, treat Asia-Pacific like the EU, right? But EU is a single market. They have very much shared the language, English, also one currency and they have the European Parliament to pass legislation for EU member countries. Asia-Pacific is far diverse, far different, and much bigger. So it’s hard to just copy whatever works in EU and then let’s also do it in APAC. Using AI regulations as a clear and classic example, you know, you is the first You know government, you know to have the world’s first AI act, right? But the so-called the Brussels effect didn’t really happen this time in Asia Pacific countries You didn’t see like all the countries, you know, like Singapore or you know Japan to quickly follow up on You know to have a similar like a risk-based approach or penalty focused approach to AI, right? Instead, you know if you look at Japan. They are very much welcoming. Japan declared to be, they want to be the most friendly open country for AI developments. The first data exception for AI testing was actually in Japan. And then Singapore followed, and Hong Kong’s also not considering, right? So APAC took a very different regulatory approach to AI versus EU. I think this is something all the American tech companies have to realize. It’s not like America leads technology and then EU matters because of the special relationship between US and EU. So as I mentioned at the beginning, the resources for public policy work are very limited in AIPAC, but EU still enjoy a lot of resources, this English-speaking market that has lot of political connections. And then Asia-Pacific, when it comes to policy enforcement, like policy support it feels more like a third country, overall speaking Asia-Pacific as a whole. So there’s still a lot of educational process, the learning curve for big tech, largely from the US to understand what are the challenges, what are the opportunities in the Asia-Pacific market. However, I also need to highlight for many big platforms, Asia-Pacific is actually not just the largest market by internet users for American tech companies, for almost for all of them, right? You know, in terms of user base. It is also a very important revenue source, know, the source of revenue for those American companies. So now you see the imbalance, right? You you make a lot of money from Asia-Pacific, but the support you give to Asia-Pacific is quite limited, know, compared to in the US ⁓ and EU. So the learning curve is there.American tech companies want to have a more sustainable development and want to have a more constructive relationship, sort of a more constructive partnership with Asian governments. I think there’s still a lot of work to do.Grace Shao (06:31)I think that’s really helpful to help listeners understand because sometimes people also approach me, they’re like, what’s APAC? I’m like, APAC is gazillion different markets and it’s actually so fragmented, right? And I think people sometimes misunderstand it kind of similar to like what you said. They think it’s like a EU. It’s not like actually there’s no consistency in currency. There’s no consistency language or no consistency actually even income or anything. So it’s quite scattered. that sense, I actually want to ask you, you mentioned something just now. George Chen (06:39)That’s right.Grace Shao (06:58)Japan and Korea this time is taking a more proactive actually approach as the countries themselves are taking more proactive approaches to really embracing AI and you know actually compared to EU’s more wait and see or more protective measures right which is not very yeah not not not what they usually would do what do think the trade-offs are actually in that sense do you actually think that means we are seeing more innovation or more technological breakthroughs or even economic diffusion of the technology right now in Japan and Korea.George Chen (07:30)Yeah, yeah, let me put it this way. So AI technology, you know, we believe, you know, still in the very early stage, right? Even you talking about, you know, trying to redefine Polisero, you know, but, you know, if you put that in the overall development for AGI, you know, we are still very much under the, in the early stage of the curve. So for Asia Pacific region, yes, it’s diverse, you know, but we can still see some sort of patterns, similarities in terms of different AI strategies. At the Asia group, my firm, we did a research paper on the different regulatory approaches to AI governance in the vast Asia-Pacific region, from Australia to even in Mongolian. Long story short, you are right. Some countries in Asia-Pacific take ⁓ a more economic benefit focused approach, right? Take a more innovation focused approach. Countries like Japan, Korea, Singapore, they want to see how AI can help them to drive economic impact, right? It doesn’t mean like they don’t care about the safety, the security issues, but they want to have certain flexibility, to encourage more startups to succeed, right?in to a certain degree, actually maybe too many surprising because China is very well known as one of the strictest internet market in the world. Basically, none of the American, very few, I will say, like very few American tech companies can really succeed in China. The only two exception in my mind are like Tesla and Apple. But they are more like consumer related if you touch on content.We talk about Google and Meta, that’s a completely different story. But even so, China at this time is also taking a more pro-innovation, pro-economy approach to AI development because this is a very top-down approach because President Xi saw the success of DeepSeek and he basically wanted more success stories like DeepSeek. Japan and Korea are in more or less the same category, like pro-innovation, pro-economic recovery. For Japan,I talked with my friends and colleagues in Japan. The sentiment in Japan is like, we’ve lost 30 years, guess, three decades in terms of economic recovery. This is like our last chart. And Japan has been quite strong in robotics, those fundamental technology development. So that’s the sentiment in Japan. We have to grab the AI opportunity. In EU, have to say, part of the reason why EU is so keen to develop regulations, legislation in recent like five to 10 years. In my view, some may argue and disagree. I think the EU does come with a sense of protectionism, right? Because if you look at all the market leaders, you name it, OpenAI, Google, Microsoft, AWS, all of them are big tech from America, right?I remember there was a chart to list the top 10 most advanced AI models. There’s only one model from EU, actually from France. The rest are from the US and China. So that tells a lot. If you are the EU regulators, look at from a competition perspective, you will more or less have a sense of anxiety. And then you will look at all those big tags like, no, we need to do something, like a country that pays in the name of safety and security. I’m not blaming EU regulators for doing it. But in the meantime, we also hear more and more concerns, even from the state heads, like French President Macron. He’s concerned that tough regulation in EU on AI will harm innovation in the EU rather than help European startups.Grace Shao (11:14)I think we can double click on China later. It’s going to have its own special segment for sure. China is just such a big story. But for some context for lot of listeners, Meta and Google, the likes of these companies actually do exist in mainland, but they mostly only have their ad services there. So basically they help enterprises with their ad sales to the West. But to Georgia’s point, they’re not really operating at the full capacity that you would see them elsewhere in the world.George Chen (11:33)That’s right.Grace Shao (11:39)Now I do want to kind of finish up on the APAC kind of narrative and then the APAC focus right now, which is for ASEAN right now. Let’s set apart like South Korea and Japan and China, just the Northeast Asian countries are frankly economically much more, you know, like developed as well as more economically focused, right? For ASEAN right now, especially since I just went to Singapore last week, it’s really interesting. Like we basically have the players, like you said, OpenAI, Google, Meta, all of these. Well, APAC headquarters based in Singapore, even the 10 cents and the bite dance of the world, right? However, Singapore is tiny, like just in terms of size and its resources. So what we’re seeing is they’re extracting essentially all the compute energy data centers, connectivity, any of the infrastructure you need to think of actually to Malaysia, in Malaysia, in Indonesia, in Thailand, even they’re building them out over there. How do we actually understand this right now? Is this a net benefit for these economies? Or is it actually really hurting the local economies and, you know, in some ways exploiting them and really just only serving the companies based out in Singapore? How do we understand that?George Chen (12:45)That’s right. So let’s talk about Southeast Asia. It’s complicated. When we’re talking about APEC, actually the most complicated part, I think it’s like Southeast Asia. Because when we talk about Korea, Japan, China, even China is a socialist country, but in terms of economic models, there’s a lot of elements related to capitalism. So those are the most economic economics in the Northeast Asia. Southeast Asia is very diverse, very different from each other.Singapore is like the exception, the most advanced economy in South East Asia. But they come in terms of population, the user base is pretty small, like 4 million, 5 million population, even smaller than Hong Kong. You’re right, a lot of the tech companies, even before AI become a trend, they talk about like Meta, Google, Apple, they all had their headquarters in Singapore. It has really become the hub for big tech over the past 10 decades. Unfortunately, Hong Kong, thank God that we still have big banks like JP Morgan, Goldman Sachs in Hong Kong, we remain as a financial center. But in the aspect of tech innovation, you have to give some respect to Singapore. They did very well to attract those tech headquarters. So this also became, you are right, sort of a point ofI don’t know how to describe it. Some of the neighboring countries are jealous, certainly jealous of the success in Singapore, right? And then countries like Indonesia or Malaysia also wondering like how to get the benefits from the fact that all the big tech have their headquarters, regional headquarters in Singapore, right? But if they only care about the relationship with Singapore or in government, because they have headquarters in Singapore and their neighboring countries will not get any benefits, Malaysia actually founded their own ways in the regional AI race. And their offer is data center because of the stable supplies of electricity, relatively much cheaper labor costs and land costs and overall cost for data center operations. So this is why Malaysia got a lot of attention from Big Tech too, like AWS, Microsoft, they all made huge investments in Malaysia. Not AI, R &D, maybe yet, but first our data center. In the AI industry, we have a popular saying that AI is like electricity. Sam Altman said that. Basically, this is like the new kind of utilities for everyone’s life, right? But to develop AI you also need electricity. You need a lot of investments in infrastructure. This is why Malaysian already stand out and Philippines too in a way, as sort of the cheap, reliable alternative to data center investments in addition to Singapore. Everybody complains about Singapore in terms of living costs, even like how difficult it is to get work committed in Singapore these days. Even you have a call like qualify the job, but it doesn’t mean like you will get work permit immediately. Actually, in comparison, Hong Kong is doing quite well to attract the talents more easily these days in the tech and financial space. Back to the AI governance issue, yes, Southeast Asia also took a very different approach in comparison with Korea and Japan. think Singapore is an exception. Otherwise, if you look at the countries like Indonesia, if you look at countries like Vietnam, they still take a very more security-focused ⁓ approach, especially Vietnam, given their political system, right? So, Meta used to, and I think Meta still have a lot of problems in Vietnam. One of the key issues is about content and moderation, right? There’s a lot of human rights and similar struggles, Thailand too. So those countries, I feel like the sort of the older problems from the social media era was not really solved yet.And those problems will be brought into the AGI era. And when the government look at the AI, their first question is, okay, so how can I prevent people from using AI to cause any unnecessary trouble, which means like a social instability, right? So that will be the same older problems facing big tech companies. And that tells you a lot when those countries look at AI, they still come from very much a security focused on mindset.Grace Shao (17:09)That’s really fascinating because actually when we were just talking about the infrastructure build out on my end, really just I’ve done some research and writing on, you know, the Johor build out and the over capacity with data centers right now. And the unfortunate cause that’s just like the local infrastructure is not able to actually support the the rampant build out. It’s actually affecting the livelihood of people. Right. But your point is really interesting. I didn’t really think about it that way. It’s actually for the from the perspective of these big tech actually.It’s to prevent bad actors using their technology to actually propel even further, like, you know, bad, intentional, harmful content, right? And then essentially, like you said, cause social unrest that would really be very troublesome for the local government. So I guess from the policy perspective, from the social media era. But what would be something different? What would be something that, you know, big tech will have to start thinking about that they didn’t even have to worry about before.George Chen (18:03)Right. Well, you know, as I said, know, lot of the older problems from social media era will remain in the AI era, such as misinformation, know, skin, political speech, you know, especially for countries like Vietnam and Thailand, the real content. It’s always, you know, when I worked at the Meta, you know, those South Eastern countries are always considered as like a high risk countries, you know, when it comes to content policy risks, right?On the other side, Vietnam, Thailand, Indonesia are much bigger. They are also smart. They also consider AI as opportunity. So they are thinking, how can I use AI to train the next generation of talents, digital talents? Those countries also have relatively younger demographics. So there’s a lot of smart kids who can get on AI and then to learn. So I think that also posed the opportunity for partnership for those American tech companies. Can we do some training program for the purpose to grow the next generation of AI talents in those countries? I think those governments will be very much welcome those initiatives. And this is not just happening in Southeast Asia. You may know somehow I also have my exchange of career experience in Central Asia. I can tell you even countries like Kazakhstan and Uzbekistan trying to focus on talented developments. Because they believe, if you think about learning how to code 10 years ago, this is actually quite an expensive experiment. You need to get professional tutors. You need to get long hours to learn one language. When I grew up, I learned like, I don’t know if you know, we started with Microsoft, the DOS system, and then C12, no one talking about it.So it took like a year to just get a basic sense of those languages, right? But like AI, you don’t need to learn ⁓ the code. It’s more important for you to understand how to write a proper talk. So those countries like Uzbekistan and Kazakhstan also catch up trying to be the back office for big tech to train, to grow the basic, like the junior engineers. So hopefully they can get some basic work done in those countries for cheap labor cost reasons rather than you need to hire all those engineers in Silicon Valley. And I think that posed the same sort of opportunities for Indonesia, Vietnam, Thailand and other Asian countries.Grace Shao (20:30)That’s really interesting. So there’s like a reshuffling of talent and then also like just the talent strategy is actually changing from the social media era or just like the big tech era. I want to kind of look at responsible AI. So we hear the phrase a lot, right? Responsible AI, AI safety from your experience right now.What does responsible AI actually look like inside of a company and what changes in org charts, KPIs, or decision making when we’re talking about responsible AI? What are the metrics we must track?George Chen (20:59)That’s right. Okay, so you mentioned that I wear a lot of hats. You I don’t want to speak like a professor, but I do teach a course at the University of Hong Kong and the Tsinghua University. My course is about digital society and governance. One of the lectures is actually about AI governance for corporates. So responsible AI is a term, you know, very popular, not just in the tech industry, but you now hear more and more just in business in general, right?It’s, in my view, responsible AI is something like the privacy statement, right? You know, for different companies, you know, when you go to a website now, like the privacy statement already become like the very normal thing, right? You know, when you use a service, you know, have to get, you know, they have to get the user consent first, and they need to tell you, you know, what kind of data they’re collecting for what purpose. That’s the privacy statement. Every website, you will find, you know, a privacy statement. Responsible AI is similar.So the government is doing their job ⁓ from regulatory perspective, from self regulatory perspective, the government work with NGOs and associations to have an industry code. But for corporates, responsible AI is kind of like the business led principles. I want to use Microsoft as a perfect example. I think Microsoft is leading the way how business can take a more responsible, sustainable approach to AI. Microsoft is responsible for AI. They call it the trustworthy AI, but it’s just the name change, more or less the same. Microsoft very much focused on three pillars, and I believe many other AI type companies focus more or less the same. First is security. You have to have a very secure AI system. That’s the basis. That’s also where the user tries to come from. Second is about safety you talk about online safety, particularly for those more vulnerable groups like children and women, how to address those issues. Again, the same social media problems like harassment, online safety, even suicide prevention, exists, if not get worse. The last one, at least, is privacy. That’s easy to understand. So, safety and security privacy. The three pillars are the key foundations for responsible or trustworthiness or other names. When we talk about the process for big tech or just traditional business like Starbucks, when they want to implement AI in their business, we have a massive means called the privacy by design in the social media era, which means privacy should be the first thing to consider when you develop a product. This is like a rule. ABC like a 101 for any product manager, right? You know, when I worked at the Meta, we always got a reminder like, you know, the engineers, right? It’s not like you have a great product idea and you talk to everyone and finally you think about, okay, I should talk to my privacy lawyer. All right, you should do it the other way around. The first of all, you should talk to is the privacy legal, the privacy team, right? Responsible AI poses a very similar approach. The first thing, when you develop an upgrade or a new service backed by AI, you should think about whether you can tick the three boxes, security, safety, and privacy for the AI services and product you’re going to launch. Microsoft has set a very good example when they’re developing the co-pilot. That’s their AI platform. for you users. So I hope that can give you a very rough sense of what responsible AI is about.Grace Shao (24:39)I think what you mentioned just now that stood out to me is that a lot of these big tech companies like Microsoft or Sell, they have very mature, legal, and safety teams in place, right? So it’s much easier for the developers to actually tap into their know-how and their knowledge. And obviously, like you said, an extension of how they use their regular content as well as not just content moderation, but also just product safety. But for startups, I don’t know if you work with them at all or not, but like,I just the proliferation of AI tools right now, right? It’s like, it’s very, it’s very crazy right now. Basically, like you also kind of hinted at this where like, you know, developing a new product is so much easier than it was say 30 years ago. It’s not only that, like, you know, the language of coding has made it easier, but now we have a jented coding tools, right? So you can have vibe coding, whatnot. How do we actually understand product safety and like responsible AI when we start talking about new products within these startups. And also my question is on a broader like picture, how do we understand responsible AI in a big market like China where a lot of products are consumer facing AI versus maybe the US where it’s a lot more enterprise facing. Can you kind of give us some color on that?George Chen (25:55)Right. So first of all, startup, yes, you we do have some startup clients. I’m very glad that the startup clients we work with in the tech sector are very much, you know, either backed by some leading figures in the Central Valley or by global VCs. So I think that they do have, you know, like a stronger internal compliance control. Right. And over the years, I think all the big tech, you from know, met up to Microsoft, you know, to other companies. I think all the classic incidents, know, the lessons, remember, you know, when Mark, when Mark Zuckerberg had to apologize, you know, you know, the Cambridge Analytical incident, right? It feels like a not too far away, you for people who had short memory. I think that those incidents that did, you know, ⁓ serve as very good lessons, you know, for those, you know, I will say like a more US-funded back than startups. I think their goal is clear. If you want to get listed on Nasdaq someday, you’d better do things very right from the beginning. There are some naughty boys, cases from China. You probably noticed there are some AI-cub startups from China.Like grab the content from Disney, know, Parliament, know, Sony, right, you know, to make those funny, like the AI, you know, effects. But in fact, it was like a serious violation of, you know, IP prototype content, you know, but those start up like, I don’t care. I just like to have fun. Like, let’s see how it goes. then suddenly, you know, they got like a 1 million to 2 million, you know, and then to 10 million users within a week. So, but they’re not going to go far away, right? There’ll be like a long series of this and that. So this is not the right approach. I do think that startups need to be very clear about the boundaries. It’s not like, okay, you are a startup, so you can lower your compliance requirement to do whatever you want. And the end of the day, you need to be responsible, not just responsible AI for the users, you also need to be responsible for your investors, right? So that’s on the on the startup part. In terms of compliance, think the startups in general do pay a lot of attention to compliance with different, especially now AI, as we mentioned, right? If you look at the APAC, there’s no unified approach to AI governance, right? It’s not like EU has AI efforts. So the compliance cost is indeed very high startups. This is the luxury Big Tech have. We just discussed Microsoft as a case study. So Microsoft has pretty decent size of legal team, security team, enforcement team, to support those three pillars, like safety, security, and privacy. But for a startup company, you can imagine they probably only have one legal, one policy manager for everything.That is it. That is a challenge. And then this is also why a lot of companies complain about very tight regulatory environment in EU because as a startup you don’t want to spend all your money, not even like half of the money as your compliance cost, right? So I always joke with my friends like if you hire more lawyers than engineers for a tech company, I don’t think that’s right. So this is a constant challenge for startup, how to comply, but in the meantime, also keep innovating.Grace Shao (29:28)I think that’s really interesting because essentially, like you said, whether it’s a startup or enterprise, in many ways, it’s faced, they’re facing the same issue. But my issue right now with kind of the AI space is actually there’s lack of international standardization, right? So for example, like globally, wherever you go, you can’t really go stab someone. The rules around drugs or even other issues like driving and other safety issues may vary, but there is like a standardized base normalization or what we believe as humans that should not be done, which is essentially don’t kill people. Homicide is illegal anywhere you go, right?So now with AI, regulation, is that like right now we’re not seeing countries come together and say, this is a sanitized belief that we should just not have. Maybe like you mentioned, child pornography and child safety is something very high on the radar, but even that can be quite subjective from culture to culture. So how do we make of that when we’re going to have AI proliferation across the economy and different touch points in our daily lives?George Chen (30:33)That’s a very important, interesting question and point you make. You you reminded me, you know, when I teach my students in the classroom, one of the examples I give them is, you know, I travel a lot, right? So different countries, to different countries, the first question I ask myself, like, which socket do they use for plug, right? And then even in EU, you know, like, well, in the UK, it’s no longer taught on EU.But even you cross the border sometimes, know, from country to country, you need to, that’s why we always bring a travel adapter, right? So when it comes to AI governance, it’s actually the same problem. You absolutely right in the very spot, EU has the EU AI Act. If you are a startup, you want to expand into EU, no argument, no negotiation, you have to comply with EU AI Act. Plus, several other regulations like the Digital Market Act, the Digital Service Act, then plus GDPR. So to expand into EU is not easy. The compliance cost will be very high. But same thing in APAC. You go to different markets. Indonesia is going to have their own AI regulation versus Singapore, versus other markets. Ideally, the UN should take up a bigger role, a more powerful role to sort of you know, have, you know, control or supervision over like how AI should be used. Right. You know, think about the same question about telephone, you know, when telephone was invented, right? Why, you know, Hong Kong’s, you know, country code is A52, right? Why China is like A6, why US is 001? Because someone made the standard and for telephone code,That was ITU, the International Telecommunications Union. So some people say we also need someone like ITU. Maybe the UN has an AI panel, but I don’t know how powerful the UN AI panel is. I mean, not to mention that the US government is not really a big fan of UN these days. So I agree, we should have international standards on AI, especially on AI safety as a key part of AI governance. We should have some principles.So I think this is something all the countries are looking to. We will very soon have the new annual AI summit in India in February in 2026. I think that India also wants to use the AI summit as opportunity to discuss those standard issues. And also to a point, I don’t know how many already realize, actually US and China are not just competing.In the aspect of AI technologies, like official recognition, deepfake, and other issues, but also compete with each other on AI governance. Basically, the and China are competing, like who’s going to write the rules for AI usage for the next generations. So this is also another flashpoint between the US and China when it comes to technology innovation, not to mention the two countries who are continuing to compete in the aspect of AI technologies, you who’s going to have more faster motors, whether it’s Gemini or DeepSeek in winning the battle. So that would be also a story we watch very closely in terms of competition and struggles.Grace Shao (33:57)Yeah, I think it’s also just because the technology is moving so fast right now. It’s really hard for regulators to keep up even domestically in each country at this point. So yeah, I do agree. I think we need some kind of international standardization. I met with Quaishou’s representative a week ago and it was very interesting to hear. They’re very ⁓ focused on the text to image and text to video kind of space. And basically they said in China,To your point cyber security laws are one of the strictest in the world actually in terms of AI Content AI GC is also one of the strictest in the world She said that actually if you remove the watermark that is actually a criminal offense or like literally you will be like, you know Yeah, it’s quite interesting. And I mean on one hand you think it’s very extreme on the other hand I think it’s very needed right like to make sure that deep fake or the mouth practice or you know, Fabricated content does not spreadGeorge Chen (34:37)That’s right, yeah.Grace Shao (34:52)And kind of lead to the social arrest you mentioned earlier or company disturbance, etc. Or even human to a Harm. Anyway on that note, I want to talk about China. You are interviewed a lot by the media on China US Whether you want to frame as tensions or competition or you know or the race? you know, whatever we want to frame it there is going to be right now two camps essentially, right? ⁓ How do we actually understand the two ecosystems at a high level? Where are the real fault lines? Are they chips, cloud, data, or like you mentioned, regulatory rules? Help us understand the two ecosystems.George Chen (35:21)Right. So let’s talk about China. So US and China don’t just compete in technology. US and China also get more more clashes on AI governance in how the way AI should be regulated. US published the AI action plan under the Trump administration. The AI action plan published by the Trump administration is actually already a shift from the AI policy approach taken by President Biden when he was in office. When Biden was in office, it was more like about protection. Biden focused very much on the online safety and this and that. They even set up the US AI Safety Institute. When ⁓ Trump took over, things changed quite differently. Now, Trump is taking American first approach for know, like America’s version of AI innovation, right? Which means like how we can keep American competitive in the aspect of AI technologies. In the meantime, think the Trump administration also want to export the US governance model on AI to the rest of the world, to many of its allies, especially in Asia, you have like Japan, Taiwan, Korea. While China is also trying to influence perhaps mostly global South countries and Bayer Road countries to be more aligned with China’s AI governance ⁓ model. So the two countries are not just competing in technology, but also in the way how AI should be governed.Grace Shao (37:00)Think on that note, if you’re a developing country right now, whether you’re in Asia, Africa, Middle East, and you’re listening to pitches from both Washington and Beijing, like you said, essentially they want to capture the rest of the world, what questions should you be asking to avoid being locked into one ecosystem?George Chen (37:17)Right, I’m always asked by my friends from global service countries, which side should I take? And my answer is no, you shouldn’t take any side. You should take whatever that fits you to have a sort of combination of the best that you can take from both the US model and also the China model. In some ways, China was quite innovative to solve some unique challenges caused by like a... know, defake and this and that. But in the meantime, the people will say, oh, wow, you know, but you have to sacrifice a lot of, you know, all your privacy, right? You know, even the internet, you you to get on the internet in China, you you must be a real person. know, China has the real ID, you know, policy, right? In Hong Kong, not the same, you you can’t just have, you know, like a, a, a, like a, don’t need to, you know, you have, you know, even for the mobile phone, need to register a number with your real ID. But in the US, this is quite unthinkable. But the real ID approach in Hong Kong and China can certainly strengthen a lot of people’s policy concerns. Versus in the US, everybody can join the party, basically. That will also waste a lot of time. think in China, you will see very much led by companies ⁓ like the traditional BAT and DeepSeek and Huawei to enhance their AI governance through a more company-led plus state-supervised model on AI governance.Grace Shao (38:48)I think that’s really interesting. You just mentioned something that struck a chord with me because I was just in Singapore and I was reflecting on how I basically hated my experience there six years ago, seven years ago when I was single without kids because it feels very like, not control, but everything feels very watched and very sterile and you know, your point, everything is very top down. But this time going as a mother of very young children.George Chen (39:04)Right.Grace Shao (39:10)I loved it. was like, wow, it’s so clean. It’s so safe. I rather give them all my data so they can protect me. They know where to track like ad actors. And I think to your point is very interesting. The idea or the value of Liberty per se may be very different in different cultures and also might change as you go through different phases in your life. So it will be interesting to see how companies or countries choose which ecosystem to join, right? Based on their own.George Chen (39:17)Ha ha haGrace Shao (39:38)belief system or value system. I want to ask you, how are the extra controls right now actually affecting companies operating between China and US? Because I know a lot of your clients probably are operating between these two large economies. You sit in Hong Kong and most of your clients, would assume, are actually like MNCs and have some kind of a... You use Hong Kong as some kind of a gateway to intern exit, mainland China, right?George Chen (39:48)Mm-hmm.That’s right. think that the China model so far, the China model is basically a more multi-barad... the more parallelism, right? So like, you know, to work with different countries, more stakeholders approach. This is also what, know, Premier Li Chang, know, when he was in Shanghai, I think in July for the World AI Conference, you he also called AI as a public goods. You know, I find that that concept was quite interesting. Basically, said this is not just something you and I should exclusively hold, right? This is public goods. This is for, you know, like, almost like, you know, this is like for the fate of the whole mankind in the future. So we need to share the success, share in the growth, versus like the American approach is very clear. You know, this is America first, we need to take the lead. And America has always taken the in AI and technology ⁓ innovation. And again, you know, don’t get me wrong. I think both Li Qiang ⁓ as Premier for China and President Trump have their own very good reasons to manage AI in their own ways. One is to keep raising the American flag high and to make yourself a role model, right? The other is to have a more open model everybody can come and share. I think so far the Chinese model perhaps is more appealing to a lot of developing countries, given the-It’s more like a cost efficiency and a top-down approach also boosts highly efficiency rather than more like a button-up, democratic approach. You need to talk to 10 companies that get alignment, this and that.Grace Shao (41:39)Yeah, actually, you just touched on something I was going to ask you. China has pushed up the AI Plus initiative and they were, like you mentioned, Li Qiang and them are just kind of embracing this idea of exporting AI to global south. But beyond the branding, I was going to ask you, do you think it’s actually successful? But it sounds like they are, right? It sounds like the global south is adopting China’s AI ecosystem because it’s more cost efficient, deployable, scalable given that’s open source, open weight, right? I think I want to ask you one last question on this section, is, you comment on China’s AI ecosystem law in the media. What is something that we’re missing here? What are people kind of missing, maybe even in mainstream media that you think is very important for people to know?George Chen (42:06)Right. That’s an interesting question. think that the international cooperation part is something I’m quite concerned about. China has a lot of good engineers. Actually, we also saw a lot of engineers coming back to China from the US. However, both sides, the US and China, should talk to each other more to maximize the research capacity for the overall interest of the whole mind can. So far it’s not happening. And then the result is, I also think AI, the reason why AI is so special is I believe AI also touches on ideology, the way how people think about things. So the US right now is very ⁓ US-centric, just focused on their AI. And then China is very much oriental focused, trying to focus on their version. So when the two AI is a basically, you know, a developer like in the parallel approach, you know, you don’t talk to each other and that will result in a more divided world when it comes to content moderation, when it comes to understanding, you know, certain issues, you know, which policy approach you take, you know, to explain a historical event, you know, for example, this, if the two countries, US and China, they don’t talk to each other, it’s not going to be helpful.For the overall development in R &D. So again, when I teach my course about the YouTube governance, I use the world’s most popular apps as example. Can you believe, it may not be a surprise for you, actually seven out of 10 most popular apps are in English, originally from the US, very much from California. The other three apps are either from China or Singapore, it tells you something. I think when social media companies began to expand into the Asian region, a lot of countries were fearful of the impact that American social media could bring to their markets. They are also talking about the so-called digital colonialism. Which AI you use that will influence your thinking.So I think in a way, people also need to be mindful whether you are too much into the US model AI, and then that also begin to change the way how you think about things. I actually tested the Chinese AI, and I tried some new features of the AI models. Sorry, I’m trying to think about which model I use. But my point is, the Chinese models are very local, very efficient. For example, when I’m in...Beijing, right? I’m not going to use Chai Chi Bti, not because of BVPN, but I just find like the DeepSeek, you know, the database they have is more practical and efficient and timely than like say Gemini and Chai Chi Bti 4, right? So if I look for the best noodle restaurant, you know, the DeepSeek in Beijing actually, the answer from DeepSeek, you know, could be much better, more accurate, you know, than Gemini and OpenAI.Grace Shao (45:17)That’s really interesting because I think it reminded me of one of the Chinese LLM startups and they said that they’re actually working with local governments in the global south and exactly to your point is that they localize information, localize the culture or language. It goes beyond just the surface language, right? I think that’s really interesting. I wanted to ask you one last question, which is...What is one differentiated view or non-consensus view you hold? This could be about the AI sphere or it could be about something in life.George Chen (45:56)I’m still trying to understand how we should position AI. There is a debate in the industry whether we should position AI as your assistant or more as your partner. I don’t have a clear answer on that. In some cases, I want AI just to be my assistant, which means I tell AI to do what and then you do exactly what I want you to do, right? But...I also understand if you just position AI as your assistant, that that will also limit the potential of the development capacity of AI. But when you treat AI more as your partner, I’m thinking about one of my favorite movies. I don’t know if you remember, there was a movie called Her. There was an engineer talking to the computer. Scarlett Johansson played the sound part.Grace Shao (46:41)Scarlett Johansson, Yeah.George Chen (46:50)for the AIs and that was quite a romantic movie, but the ending was not very, it was not a happy ending. So I’m trying to think like if you produce AI as a partner, you can empower AI to do more things, but then, know, whether eventually we will also enter some dangerous territory, you know, to have AI to have, then we need to talk more like ethical issues, right? You treat the AI more as a partner. There were discussions, know, yes, know, AI doesn’t have feeling.But should we also ask the AI to work for like a nonstop? If you think about human rights, should, if human will work for like 10 hours, 12 hours, you should have to get a break. Why we shouldn’t just keep asking AI all these questions, keep them running the models to get the result. And even AI doesn’t have the touch feeling, does AI have emotional intelligence? I believe at some point that AI will have emotional intelligence. That is to say, if you use AI as sort of a slave.they will also be unhappy. So back to my point, I don’t have a clear answer, but I’m still wondering which sort of status, like a category, we should put AI into, more as AI assistant or as ⁓ a human car parking.Grace Shao (47:58)I think that conversation can like, warrants another conversation on its own, that topic, because I think to your point on the technology aspect, we are seeing a shift from AI just to consumer and just as a chatbot to agent AI, right? So to your point, know, ⁓ AI can actually start completing tasks for you. They can be more proactive, remind you to do things. They are more like a thought partner versus an assistant. But again, to like, you know, to even the conversation we had earlier, it’s like who...George Chen (48:01)Ha ha!Grace Shao (48:26)Who can play God? Who is to say, where is the line, right? And your 10 to 12 hour work ethic thing is very Chinese and American. Definitely in Europe, people are not working 12 hours a day. That is like a, that is a normal work day for Americans and Chinese. But again, yeah.George Chen (48:40)That’s right. You already see the cultural difference here, even for the real world and for the AI in different parts of the world.Grace Shao (48:46)Right? So who gets to set the standards? And I think it will become harder and harder. It’s because then it becomes more philosophical and ethical than just, you know, ⁓ practical, which is right now what we’re talking about AI safety is just like, okay, child pornography is fundamentally wrong. Like, homicide videos is not allowed. Don’t create fake videos of people doing fake things. That is very black and white. We can almost just all universally agree.But when it becomes a cultural, evidently because they’re culture norms, even language norms, societal norms, et cetera, right? Or even each person’s emotional capacity is even different. Then who gets to decide when AI needs to stop, right? That’s definitely like a very interesting topic. And I’ve been having this conversation with friends as well. It’s like, has technology hit a point of actually further development does not progress society as a whole anymore? Or are we still actually benefiting from technological advancements. So anyway, I really, really appreciate that. And I can go on forever. This is an interesting topic. Thank you so much for your time, George. I really, really appreciate your insights and all the expertise and experience you bring to us. AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Why American Investors Should Have a Pulse on China with James Wang 09.12.2025 1h 5min

  In this episode of Differentiated Understanding, I talk with James Wang, general partner at deep-tech fund Creative Ventures, author of What You Need to Know About AI: A Primer on Being Human in an Artificially Intelligent World, and writer of the newsletter Weighty Thoughts. James has sat on nearly every side of the table — Bridgewater investor, startup founder/CTO in healthcare, engineer at Google — and now backs “real-world AI” from semiconductors and interconnects to diagnostics and industrial systems.We start with how the AI investing landscape has evolved since 2016: why “AI” used to be a dirty word in pitch decks, how the post–ChatGPT boom funneled capital into a small set of model companies, and why so many AI startups shot up to tens of millions in ARR only to fall back as incumbents absorbed their features. James explains where he still sees real opportunity — especially in vertical AI built on hard-to-replicate proprietary data — and why moats in healthcare and industrial AI look very different from the “GPT wrapper” era.From there, we zoom out. We compare China vs. the US on AI pragmatism, industrial policy, and consumer vs. enterprise strengths; unpack the open-source vs. closed-source model debate; and talk about how agentic AI is already furthest along in developer tools. James also breaks down the energy reality of AI: why GPUs turn power into intelligence, how much additional load AI really adds to the grid, and what the Inflation Reduction Act and its partial rollback actually changed (and didn’t) for data centers and renewables.We close with James’s differentiated view: that over time, AI’s gains will be largely socialized — diffused into everyday life via cheap, ubiquitous models (often running at the edge) rather than captured as persistent monopoly profits by a tiny set of firms.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Topics we covered:* What “real-world AI” means: interconnects, power, semis, diagnostics, industrial systems* How AI investing has changed from “don’t say AI” to “everyone is an AI startup”* Why many high-flying AI startups lacked moats and saw revenue fall back to earth* The case for vertical AI built on scarce, proprietary data (e.g., medical imaging, acoustics)* China’s strength in industrial AI and consumer apps vs. the US edge in enterprise SaaS* Open-source vs. closed-source models, and what really matters for enterprise buyers* What “agentic AI” actually is, and why dev tools are still the most advanced real use case* AI’s power appetite, data centers going “behind the meter,” and the limits of US grid politics* Why James believes most of AI’s value will show up in broad productivity gains, not just in a few mega-capsAI-generated transciptGrace Shao (00:01)Hey everyone, welcome back to another episode of Differential Understanding. Today, joining me is James Wong. James is a general partner at Creative Ventures, spearheading investments in AI across the stack. He was previously the co-founder and CTO of Lioness Health, and before that he was on the core investment team at Bridgewater Associates. He founded a nonprofit consulting firm specializing in microfinance and had a short stint at Google. I’m very excited to actually have you on today, James.James Wang (00:33)Super excited to be here too. Thanks so much, Grace.Grace Shao (00:36)James, it’s really great to actually finally meet you, I guess, in person. We were just kind of laughing about this. We’ve talked on and off on Substack on WhatsApp, on email for quite a while now. I’ve been a really big fan of your writing and you are actually one of the first paid, I think, subscribers to my own newsletter, AI Proem as well. So for listeners, his newsletter is called Weighty Thoughts. He writes everything about the startup space, VC, AI, know, FinTech, I think, and know, bigger pictures as well, right? But ⁓ start with James, why don’t you tell us about your day job? What is it that you do when you say you invest in AI? What are you investing in? And what kind of businesses are you looking at these days?James Wang (01:19)Yeah, sounds great. yeah, glad to be one of the early subscribers to AI Proem because yeah, as a VC, you have to catch the good thing early. it’s a part of the job there. Yeah. So Creative Ventures is an early stage deep tech fund. So deep tech has gone through quite a few evolutions in terms of what it is or isn’t to people.For us, it is things with harder science and IP barriers. So that includes things like battery manufacturing platforms that cost a billion dollars a piece, AI diagnostics that are completely end to end with no clinicians in the loop, things like materials discovery platforms using AI. So a lot of these different areas that have gotten pretty exciting with AI as well. think one of the interesting things is deep tech, especially within say like some of the materials discovery spaces, the bio space, like a lot of these areas have accelerated quite a bit with AI’s ⁓ involvement at this point. And there’s a lot of exciting things coming up in those areas.Grace Shao (02:27)I think you know beyond a business background which a lot of investors have you actually have a technical background as well ⁓ What do you think that like does that make you? More understanding of the deep tech that you’re looking into or do you have any unique perspective on technology companies when you look at investing in them?James Wang (02:47)Yeah, totally. So for us, for our team, actually, I’m one of the few folks without a PhD. So a of the team does actually have that kind of background, which is needed within deep tech ⁓ in large part because it’s you do need to understand how the technology works in order to understand the market that it goes into. That being said, like like most technology startups, the ultimate challenge is finding the right market and scaling.But if you don’t understand the technology on a base level in terms of what it does, it’s really, really hard to actually figure out how to scale the thing. So a lot of that technical background, especially within these areas is quite critical. And I guess just my opinion as well, like a lot of different asset management areas undergo evolution. VC historically has been one that has allowed a lot of generalism within it just because of the nature of how a lot of the software boom went came up and went through and everything. But our opinion is actually a lot of the investors in this particular area will get more and more niche, especially with AI, which I think we might jump into as well. AI does actually involve and help enable a lot of vertically integrated industries ⁓ in interesting ways, which means that, you end up with investors who get more and more specialized in their areas.Grace Shao (04:00)Mm-hmm. How big are these ticket sizes that we’re talking about when you’re investing in and how early are you looking at?James Wang (04:13)Yeah, typically speaking for us, we are often the first institutional investor in that being said, some of our companies have five, 10, even like 20 million dollars in non dilutive government grant funding or research funding before we actually invest. So it’s kind of hard to say when you’re trying to pin that down. That being said, yeah, we’re among the first investors in. Usually we invest around a million in terms of initial check size and sort of ramp from there.Grace Shao (04:28)I see.James Wang (04:42)⁓ And then our companies obviously like as they get larger and larger later on They can have quite a bit of range in terms of like where they end upGrace Shao (04:52)I see. I definitely want to double click on the vertical AI space later. But to start with, another personal question I want to touch on is your book. You’ve just launched a new book. It came out in October, I believe, right? It’s called What You Need to Know About AI, A Primer on Being Human in an Artificially Intelligent World. Can you tell us a bit about the book, a preview of like, I mean, the gist of your book or why we should go read your book?James Wang (05:19)Yeah, totally. Well, think the most recent thing I can remember was someone told me the other day, think yesterday actually, that this is a great stocking stuffer for all the boomers in your life. I believe that was a compliment. So ultimate, I think so. So the book is actually a end to end. Here’s what you need to know to kind of get up to speed on AI.Grace Shao (05:33)It sounds like a compliment.James Wang (05:44)⁓ It goes through the history of it. It goes through some of the technical background. ⁓ Not too deep, but then again, like also doesn’t really pull too many punches in terms of like actually getting into the structure of it. And then finally it goes into how it’s being used today and some implications of it coming up. So it’s meant to take you through end to end. ⁓ You know, we have a lot of interesting endorsements of it. Reid Hoffman actually had read through it and gave a great endorsement of it as well and I’ve had both people within the AI business sector. So basically people trying to market AI and push it out into market as well as engineers, both tell me that they’ve learned something from it. So I think actually a lot of people can get something out of it. It’s just different people will find different parts of the book difficult or not. but it does attempt to like step you through it. so that was the aim that I had writing the book and, ⁓ hopefully I achieved it so far. sounds like it, if it’s a boomer, baby boomer stocking stuffer.Grace Shao (06:44)I think I need to get a book for myself. Does it overlap with what you write about on Weedy Thoughts or is it a bit different?James Wang (06:52)It does, but it ⁓ gets into more depth and basically takes you through A to Z a little bit more. Since I’m sure you know as well, it’s like your experience as well there. It’s like for a substack post, inevitably you leave a lot of things out. You kind of hit the high level, you hit the points, but yeah, ⁓ you can’t make the article 10,000 words long or something like that. On the other hand with a book, you do need to actually bring it from beginning to end.Grace Shao (07:21)What inspired you to write the book? Because I’m sure you have a busy day job already.James Wang (07:24)Yeah, I mean, the first part of it, ⁓ which is the real part of the the marketing story that I give is that which is actually true as well. But the marketing story I give is I sort of looked at the landscape and generally speaking, there’s a lot of great technical resources. There are actually a lot of great sub stack newsletters on AI. A lot of other good places to dive into to get a sense of what’s going on. The problem is in general with the book landscape, a lot of the stuff has been like productivity, get rich quick, et cetera, sort of things within AI, or somewhat more alarmist or very thesis specific driven. It’s like, ⁓ AI is going to do this thing. It’s going to kill us all. It’s going to take all our jobs. It’s going to revolutionize this. Like they’re pushing an agenda. I didn’t really see anything out in the landscape that takes you from A to Z for people who didn’t actually know what was going on and wanted to get up to speed. So I got kind of tired not being able to recommend a book for all the friends who are like, hey, you know what’s going on with the say hi stuff. I don’t know what’s going on with this. I stuff. Where can I go read a single book to find that out? So that was like a big impetus on why I decided to write the book. And in terms of like how I came to it, the publisher actually found me through my sub stack and kept bugging me to write a book. And eventually when I decided to go on this direction, they were like- Are you sure you don’t want to write another thing about how it’s going to kill us all or something? We think that might come off and fly off the shelves a little bit more or like draw people in a little bit more than a light textbook. But nonetheless, it has actually got number one in lot of categories on Amazon for a couple of weeks now. So I think it’s working so far, which I’m happy about.Grace Shao (09:13)I think that’s a really pragmatic way of approaching it. like, your point, I think I’ve also noticed in just the AI sphere in general, there’s like kind of the, you know, the investors are talking about money. What’s the return? What’s the return? It’s only about monetary return, right? The tech people, when you speak to them, sometimes frankly, they’re a bit ignorant to the societal implications or they’re not, most of them, let’s just say, people are not evil intrinsically. They just think, okay, tech acceleration is a fundamental goal for them.And I think sometimes to put the blinders on and they forget about the potential implications of society and the environment on, you know, shifts in like, you know, even dynamics between people and countries. And then like you said, unfortunately, I think the people who really try to bring the awareness and be mindful sometimes can go too extreme. And then what happens is like, let’s reject it. But the reality kind of sits somewhere in between where like you can’t really reject a technology once it’s out of the Pandora box, right?And then you can’t really only look at the value creation in terms of monetary terms. And you can’t really say, OK, let’s only focus on technology, but not think about all the other consequences. So I think to start our interview, let’s really talk about your philosophical view on this. I think your book really ⁓ resonates with me. What I write about also is trying to help people understand, OK, these are the business implications. This is where the return will be. This is what will be.This is how the technology change your interactions with each other. But it’s not like you can’t approach it mindfully, right? So I’ll throw it back to you. How do we understand the relationships between if you have to simplify the three kind of caps that we see right now?James Wang (10:54)Yeah, I mean, in terms of camps, mean, there’s definitely the I mean, it’s been interesting, right? Because for a while you had did have a group that basically said AI is just a fad. It’s not going to actually do anything. Ultimately, it doesn’t matter. Blah, blah. Not a big deal. You have another group that’s basically like, AI is the AI is going to either kill us or AI is going to take off in terms of singularity. And we’ll have a post. ⁓We’ll have post-scarcity utopia where everyone will have UBI and AI will do everything for us kind of thing. ⁓ So between those two extremes, I mean, you do also have people who are like, yeah, this is a great business opportunity. We’ll go after it. It has some aspect of all of these things. And like you’re saying, from a societal perspective, I think a lot of the people who boost AI quite a bit, ⁓Don’t take into account. Yes, there’s going to be disruption. mean, even if you look at the Internet boom, which I think at this point, people have ⁓ misremembered certain aspects of it because it came and went and a lot of our economy has been restructured around it. AI is the same way. It’s going to create disruptions. It’s going to create winners and losers, but it’s going to also help accelerate productivity and do a lot of good in the world, too, like most technologies have ⁓ in the entirety of history.So it’s a big part of just needing to understand what it can do, what it can’t do, and really where we will see the benefits come out. Because I think otherwise, if you’re just very much on the utopian or doomer perspective, you lose the reality of what ⁓ AI actually is, which is a tool, and what the capabilities of that tool is.Actually, in terms of this, I think China currently in terms of Chinese AI, which we’ll probably get into and Chinese AI companies have generally had a more pragmatic view of this. ⁓ I’ve had a lot of conversations in Silicon Valley, including with different folks at the model companies, where some of the goal and some of the ultimate aim was, hey, we’re going to get to AGI and then we win. And then it’s all it’s all done and that like, we don’t have to worry about anything else.I think a lot of that particular mindset viewpoint folly has sort of gone away in the past year or two. But even so, like it tells you something about like the way that a lot of people are looking at this, which is almost pseudo religious.Grace Shao (13:27)Yeah, I think definitely to your point, the the caps kind of become bit of like cults themselves as well. It’s even when you cover this space, it’s interesting to meet people who are like all or nothing, very much all or nothing, right? They’re either like, let’s go all in and ignore all the noise and all the issues or go like, let’s reject this. This is just ⁓ inherently evil, which to your point, like all technology disrupts what we know, butLike you can’t reject it, right? Okay. I think moving on from that, I want to talk about investing in AI. You’ve been investing AI since the early 2020s. I wouldn’t say it’s earliest, but you’ve been in space for longer than most people where, you know, they really jumped in after the chat GPT moment in 2022. So how has that relationship between AI companies and capital change? Like, we’re now hearing a lot of buzzwords. Is this a bubble? Is this like, it’s going to flop? Like, where are you seeing the market kind of likeWhere is it at right now? And has that relationship between the founders and the investors changed over the last few years?James Wang (14:35)Yeah, it’s an interesting question because yeah, when we had started out, and that was 2016, AI was actually kind of a dirty word just because whenever someone tried to throw AI at something, it was like, yeah, this is kind of scammy. It probably isn’t going to actually work. So you literally had different startup pitches pull AI out of their pitch and basically say, no, no, it’s just statistics. Maybe it’s machine learning, but it’s really just statistics and stuff like that.So, mean, the way and the evolution of it like changed quite a bit. I think around 2020 was when it started to get a little bit more accepted. And we started to see like certain pitches where it’s like, hey, look, we’re going to do AI for movie studios. And really what we want is to do motion capture for this or something like this. Actually, I think I saw two or three startups around that period trying to do this. And what you really want to do is like, you really want to fund us. We’re going to gather a ton of data.And then we’re going to train a huge model and then we’re going to win the market. So that was actually kind of a popular thing at that particular point in time, which also ended up becoming unpopular because it ended up not working. but it was post yeah, like some getting some towards the chat, GPT moment that suddenly like everyone knew about AI. took off. A lot of people started putting money into the LLM company, model companies, lot of the companies adjacent to it. And at this point now, I mean, it’s.Interesting, right? Because I think if you look at headlines, would think anything that has AI in its name instantly gets a ton of funding. But I can tell you just being on the ground, that’s actually not hugely the case. Private markets and VC have still been somewhat more sluggish ⁓ since 2022, since interest rates rose, and since there haven’t been super significant exits across the board, which means a lot of that capital market has been frozen.If you look at the stats, actually for all of the startup funding and AI funding, a huge proportion of it is actually just the giant model companies having mega round after mega round, or some of the second tier model companies who’ve also had a bunch of mega rounds. We haven’t actually seen like tons of AI company like Dragon, a ton of startup financing. ⁓ Even so, likeAI currently still is the hot thing. So if you’re trying to raise money, especially in Silicon Valley, you generally will probably try to put some sort of AI pitch into your thing, whether or not it actually makes sense or not.Grace Shao (17:07)Yeah, I was just gonna say it’s really funny when you said like it used to be a dirty word, whereas now you meet any company, like they could be selling chocolate bars and they’d be like, we’re AI, we’re AI company. They’re really trying to use AI as like the kind of buzzword to hook people in. And it’s interesting to hear from your perspective that actually AI startups are not getting a lot of funding, right? Because in China at least, what I cover in this part of the world,There’s already the jokes about the last round of AI startups dying out already, like phasing out recently. Yeah, so I don’t know. Is that happening in SF as well?James Wang (17:45)Yeah, so it’s interesting. maybe I’ll make the amendment that it’s not AI startups are not getting a lot of funding. AI startups are getting more funding than other types of startups. So one of our companies actually just recently was told that, hey, you’re actually an AI startup, not a health care startup. They’re definitely both. But that was their greatest compliment because that meant that the investor was actually interested in putting money into them. So it’s like it tells you.Grace Shao (18:12)That’s so funny.James Wang (18:14)Yeah, it tells you something about the landscape.Grace Shao (18:14)Yeah.James Wang (18:15)⁓ yeah, AI companies get more funding. But yeah, it’s definitely not as much of a bonanza ⁓ as you might think from the outside just by just the numbers thrown at the screen because a lot of the big companies are absorbing. But it is definitely the case that I’ve talked with a couple of other investors who’ve told me about some of the revenue numbers for some of their companies. As much as the revenue numbers shot up, ⁓Grace Shao (18:27)Yeah, yeah.Mm-hmm.James Wang (18:43)Let’s see, I’ll obfuscate what this specific company is, but there’s one company that I know of that basically was like shot up to 20, 30, 50 million in ARR in a very short period of time. And the investor who I was talking to was saying, yeah, they’re definitely going to get to 100, 120, whatever it is. The last time I checked in, I think they had dropped back down to 20 and maybe stabilizing down towards 10. So in terms of AI startups dying, the interesting thing is likeA lot of these startups don’t actually have moats. ⁓ Whether or not it’s like some random model company that likely isn’t ever going to get off the ground in terms of having enough compute resources or whatever, or a GPT wrapper, which has become a dirty word or had become a dirty, like derogatory, like phrase to say to some of these companies that just wrap their product around like a chat, GPT API. A lot of these companies don’t have any barriers to entry.So we’re already seeing them shoot very high up because their products are actually useful. And we can get into that. A lot of these are like programming developer tools, agentic ish tools within developer realm. They go up very quickly. They’re actually quite useful, but then everyone else can utilize, everyone else can build the similar kind of thing very quickly orSay as Codex came out from ChatGBT or as Cloud Code got better and better and added more capabilities, you have a lot of the big model companies themselves end up just incorporating the functionality that these developer tool AI companies tried to put in, but now they’re obsolete.Grace Shao (20:24)Yeah, I think that’s something like I’ve been writing about for the Chinese tech space as well, right? The incumbents end of day have a distribution and what they call the flywheel effect, right? I used to hate the word because I think it sounds really silly, but now I think it really makes sense, right? Like in this case of AI, it’s like if it’s not like we have a new device that we’re interacting with it, so whoever already had existing reach on these operating systems can easily basically just swallow another business, like a newcomer.and just create a function. And to your point on the coding, the agentic side, I know like Alibaba just came out with ⁓ Codar, who I interviewed a while back. ⁓ ByteDance has their similar tool. Obviously Cursor is still the leader globally right now in terms of capability. ultimately, if one day they all reach a similar, I guess, efficiency or user experience, then if you’re already using Alibaba Cloud and you’re already using their like blah, blah, blah service,you’re already buying your API there, then why wouldn’t you just use your tool, right? I’m sure it’s the same in the US, like the big players just kind of capture all at this point. ⁓ I want to talk about creative ventures specifically. You guys say you invest in real world AI, right? ⁓ That’s really much like you kind of even touched on healthcare, robotics, manufacturing. It’s maybe less about like the consumer side of things, right? What are exciting businesses you’re seeing and you think are being overlooked right now?James Wang (21:52)Yeah, I mean, think two areas. One is there are still a lot of interesting things within. For us, real world AI does actually include things like interconnect companies. It includes power management, storage. It includes like different like semiconductor based companies or semiconductor tool companies. So there’s actually a lot of interesting things going on in that realm. It has been a, for example, with like optical interconnects, optical switches, these other things. That’s been aplace where the semiconductor industry has been interested in going for years. And there’s been roadmaps and industry like things talking about like how we need to go that direction. But ultimately, no one actually ever moved because while we have an existing business, it costs time, it costs money to actually move into that. But the interesting thing with the AI boom has been, OK, all of a sudden there is an impetus to actually start adopting a lot of these technologies with some of the optical interconnects and whatnot.And there’s been actually some large exits within the space even recently. So that’s an interesting area to us still. And it’s an area that most investors have still shied away from because there’s still been a historical wariness, especially within VC towards hardware, which is ironic since that’s actually where Silicon Valley started in terms of VC. But on the other side, too, there’s a lot of stuff within health care that has been something that we’ve been pretty excited about, at least in the US. Medtech and health care has gone throughQuite a few years now actually of kind of a funding winter where a lot of well-known health care companies, digital health companies just didn’t do very well. A lot of them also tanked on the public market. So it’s just been a super unpopular area for investors. But some of the most interesting and exciting things that I’ve seen within AI have been within the healthcare sector. There’s some that are basically like healthcare productivity optimization.One of our companies, not to talk up our book too much, but one of our companies is currently the only ⁓ and first and only currently end to end AI diagnostic for them. They’re starting in lung disease, but essentially they’re a diagnostic tool that now you press a button. It sends off the scans. It comes back and gives a diagnostic and gets paid reimbursed by Medicare and all the private insurers. Like there’s no clinician, no technician, nothing in the loop, which actually doesn’t justlike increase margins to software like levels is actually insane in the healthcare sector because there’s just so much red tape. There’s so much red, so many regulatory barriers. There’s so many like pieces that can go wrong and thus you need to check and thus you need to have all these other layers that if a piece of software can actually take all that away and be FDA approved, that’s actually a massive productivity improvement in the space. Again, ours is currently the first and only, but I don’t expect it’s going to stay the only one there’s going to be a lot of really interesting things happening, especially within healthcare and biotech.Grace Shao (24:51)It’s interesting because I was just listening to a podcast with a 16 C’s podcast. They’re saying that actually a lot of biotech and med tech companies are routing their trials actually in China, just given that there’s less red tape around a lot of these issues. And I met with a company in Singapore just last week. They are one of the leading AI companies actually using AI to do clinical research and trial a clinical trials. And it’s really interesting. Like you said, AI is advanced enough in some ways that they can actually guarantee there are no issues in these kind of like more tedious work or knowledge work that it doesn’t really need that much human judgment and then the efficiency gain is crazy. So that’s an interesting space. think you’re right. And I think it’s going to blow up not just in the US, but also maybe in China space as well. ⁓ I want to ask you on China, on China AI, open source, closed source, that’s the big topic, right? It’s a whole China’s embracing open source.The US may be still leading on the closed source models. How do these choices really actually affect the products and the margins ⁓ when you’re looking at these companies? there’s a lot of conversation about their performance, about deployment, but what about when it comes to actual nitty-gritty implementation for the companies? What does it really mean?James Wang (26:11)Yeah, I mean, the way that the Chinese ecosystem and the US ecosystem, it’s partially just path dependent. They’ve evolved in very different ways. In the US, you still have a lot of API or subscription usage, like specifically, you know, open AI and anthropic and Google sticking Gemini in essentially every single productivity tool that they own. ⁓ So the way that that market works is you have a lot of paid usage go out there like they wrap, they do GPT wrappers and whatnot. And because China ⁓ was later to some of the, to some of this in terms of like big breakout, essentially some of the open weight, open source stuff helped ⁓ spur adoption. So for some of the people who do not want to simply pay for chat, GBT or Anthropics API and just wrap their stuff around it and want to control some of their own destiny.It’s great to have something like Quen that you can basically fine tune. can locally host. You can locally host DeepSeq. You might not choose to. Ultimately, you might go to a number of different providers who all allow you to have it available there. But nonetheless, you basically have an easier way of saying that you won’t have lock-in. You’ll be able to use this. You’ll be able to go out there and...Uh, go out there and integrate it. So it is a, again, a little bit of path dependency there. It’s a little bit catching up in, from that perspective as a whole, ultimately in the longterm, I do think like some of the open weight stuff probably does make sense for the same reason. Open source made sense within a lot of the software ecosystem. It does actually spur a lot of enterprise uptake. can pay for support and other things around it. And the bigger thing will ultimately be, can you sell it faster? One of the things about SaaS, so software as a service, has always been, there’s no real barrier to entry for you switching to someone else, except for the fact that I made, especially for enterprise SaaS, an enterprise decision, and I don’t want to switch away from your product now there’s nothing really at the end of the day that makes Salesforce versus some other CRM versus some other productivity tool that different from one another. And at the end of the day, for a lot of the LLMs, especially as we start hitting plateaus in noticeable performance improvement for people. They might become quite interchangeable in which case it becomes similar to a SaaS decision. Do I want to choose the closed source one whereI will have to pay for it forever and also potentially have it go down and only have a single source vendor. Or am I going to take the open weight, open source version where maybe I’ll still pay for it to be hosted, but I can always be comforted that I can always like take it, roll it, and put it in my own infrastructure as well.Grace Shao (29:15)Yeah, I think right now where we’re at as the models, their own performance are getting closer and closer and like the gap is not as wide anymore. I see your point. It becomes like an infrastructure. And then I guess for enterprises, biggest issue or the hurdle is really the switching cost of like the bureaucratic switching costs. It’s like going through the legal work internally, making sure all your departments are upgraded the same way that that becomes a switching cost. So thenI guess incumbents still have the advantage once they’re like chosen as the default provider, they get to kind of own that space, right? ⁓ I want to talk about agents. You kind of mentioned earlier, like enterprises like Google are essentially plugging in Gemini into every single productivity tool. And right now we’re hearing a lot about agentic AI, how that’s going to even increase the capabilities of these productivity tools even further. How do we understandJames Wang (29:53)Yeah.Grace Shao (30:13)what even is a gendered AI right now. I think a lot of people still think of AI as just chat GPT chatbots.James Wang (30:20)Yeah, I mean, agentic AI became a big buzzword. personally, my personal take, I’ll define it a second, but my personal take is eventually agentic as a term will go away. And we’re just going to say AI because all the big model companies are also going that direction in being agentic. Now, agentic, what does it mean? Well, definitions vary, but at the end of the day, it’s, it doesn’t just chat with you.It goes and does something. So instead of I plug into chat, GPT say, Hey, where would be a good place to go on vacation? I would tell the agent, okay, I want to go on vacation. helps me research, but then it also helps me book the tickets, the hotel and like give me like the roadmap and plan and stuff like that and do the things for me. So it’s that layer of being able to interact with the world, whether it’s like true real world or whether it’s like digital world in terms of booking stuff, it can actually do things for you. Now, why, why I say it’s eventually just all going to be agentic, in which case we’re stopped going to stop saying agentic. It’s because the direction of where the model companies have gone is this direction. Uh, as some of the performance differences have disappeared, they’ve implemented more and more agentic tools. would say the most advanced agentic area, even though we don’t usually term it that way, is a lot of the developer tools. Ultimately, at the end of the day, for all the developer tools, they will take your input and they will make changes on your machine, on your code, and commit it. So at the end of the day, it’s doing things.James Wang (32:19)So at the end of the day, in terms of these agentic tools, where developer tools have been the most advanced because they actually go out there and make changes on your machine, your code, commit it, ⁓ all the different model companies have been rolling out more and more tools that specifically are helping you do things in the real world. As you stop having as much difference in how well they chat with you, there’s going to need to be other differentiation. And the place where they’re going, where there’s lower hanging fruit, is being able to actually implement and do things for you. that’s why I think agentic is interesting. Agentic is actually going to be big, but it’s just not that interesting of a term because ultimately most all the AI companies will likely go into it and implement it with their models.Grace Shao (33:07)You’re right, because I think even just this week Deep announced a new like v3 too and they’re saying oh our capabilities are really focused on a genetic AI and every single model is kind of coming out say the same thing and then this is a bit random but it reminded me of this like funny thing was growing up my friend who’s German descent one day asked me she said grace in your household do you guys say eating Chinese food Chinese food do you call Chinese food Chinese food I said no we just say we’re gonna eat food tonight you know and because he normalized and that’s default thing then you wouldn’t really like actually add these prefaced ⁓ adjectives, right? So I get your point about that. ⁓ What kind of agentic products are actually good right now, actually on this point? While models are all becoming more agentic, what actually is it being used in right now? And what tools are actually proving themselves to be really capable and productivity and enforcing?James Wang (34:03)Yeah, I mean, I’ve seen various attempts to do things like ⁓ shopping aids, ⁓ different things with, yes, travel booking, concierge services, ⁓ email responses, bot, chat bot, like customer service ticket, chat bot kind of things. I would still say like among these different things, the most advanced is actually still the developer tools ⁓ ultimately.Grace Shao (34:07)Mm.James Wang (34:30)⁓ it’s close to the companies. It’s close to the people creating it. It’s right now the most advanced area that I see. ⁓ but there’s been a lot of experiments in many of these different areas and they are starting to get better and better and work better and better. So I do actually expect for a lot of these different things where, especially where I guess the framework that I would put on it is if the agentic application is low enough stakes from the perspective of there is a human in the loop that will check it at some point, like, hey, I’m going to book your vacation. Here’s all the pieces of booking your vacation. And you look at it and go, yes, you are correct. I am going to Athens, Greece, not Athens, Georgia, and something like that. If there’s a human in the loop and something where there’s a check, ⁓ these are actually applications that the AI can do quite well and is actually something that’s very implementable currently. So that’s kind of the layer that I put on it. But yeah, like some of these areas are getting quite advanced.Grace Shao (35:32)Yeah, like you said, you need to have that human verification. It reminded me of that new show. It was like a silly rom-com. It’s like they bring these women to Paris, but it’s actually like Paris, Texas, and everyone get off the flight and they’re like, my God, this is not the Paris I imagined. Okay, I wanna kinda go into China a little bit. Like I said, you were one of the first paid subscribers to my newsletter and I very much focus on the China space, although I do cover a bit of APAC and different companies as well. What piqued your interest in China AI? Because it’s not like you actually directly invest in China AI, right? So tell us a bit about that.James Wang (36:11)Yeah, we’re not able to for various ⁓ investment restrictions, some of our investors and whatnot. But at the end of the day, ⁓ it’s a global market. China is a huge market and China also has a lot of talent within it. It’s kind of funny because it’s like, why was I interested, for example, with your newsletter? ⁓ One, you write well, so there’s that. But also it’s just like having the perspective within the market is super important.So I actually still have a lot of conversations with Chinese companies. They know I can’t invest, but they’re always kind of interested in also learning about what’s happening in Silicon Valley, et cetera. So I keep a pulse on the Chinese market that way. And that helps inform my decisions, but also my understanding of like, what does the landscape look like in the U S it’s both compare and contrast, but it’s also thinking longer term. What’s going to happen as all of these companies go more and more global.whether or not they’re competing directly in China or the US, you’re going to encounter each other in the rest of the world, right? So there’s a question in terms of that. As for why, there’s also not very many good sources on China. I think you covered this in some of your articles about some of your own personal history, Grace, but it’s just like, even recently when I was trying to prepare a presentation for talking to some folks about some of the things happening in China where AI is being pushed.especially by the state into a lot of areas like insurance and whatnot. I was trying to Google like what’s going on in China with that, like with just like very simple terms. And really what came up for me was like New York Times articles about Chinese surveillance state is China like doing these things to the Uyghurs is like what it’s like all of these different things that were obviously had its own like bias, let’s say. ⁓ And at the end of the day, it’s like China.Grace Shao (37:36)Hmm.James Wang (38:02)especially for the West has been a little bit of a cipher. It’s either can’t innovate at all and only copies things, or it’s like the crazy, huge country that suddenly like will be able to overtake everyone and like whatever it is. It’s the country that who’s the state dictates everything and thus has like total control and everything. At the same time, it’s like, it’s like a super like, like, you know, lot of the private sector stuff does things, but also like the states can’t seem to like do anything right. And then there’s corruption and ⁓ rockets, rocket fuel being used in hot pot or whatever. And the reality is it’s like, it’s all of these things together. Right. But a lot of the way that the media portrays it is fairly biased in terms of that. So it’s always very useful. And I’d say critical for investors in any part of the world to have a pulse on definitely the two biggest economies in the world. Right.Grace Shao (38:58)Yeah, I think I can go on forever about the media coverage. I wouldn’t even say it’s biased. I would just say the media business model itself actually awards, you know, clicks and attention and in this time and age and what gets attention, the joke amongst a lot of like I expect journalists in the APAC region, it would say it’s big China, bad China, weird China. So I was like, my God, it’s so many people. It’s so big, weird kind of, my god, they eat dogs, which like honestly, majority of people don’t, but sure, I’m sure some certain small segments of people do have weird diets, right? And then it’s like, like bad, bad, bad, like it’s so bad. So I think that gets the clicks, but I do get your point, you know, not to kind of bring it back to myself too much, but it really is why I started writing about it. was like, there should be a nuance understanding what’s happening, especially in the business space when it’s sometimes not really relevant to what we just talked about, the big, bad and weird. It’s just innovation and business. So I want to bring it back to that. lot of people are talking about China being very, very strong and in industrial planning, right? I think this is something that’s all of a sudden for some reason, making headlines all over the U S and last like three months, whereas like industrial planning didn’t come out three months ago. They they’ve been around for the last 30 years, really. From your investor lens, where do you see China moving the fastest? it like only sectors within industrial policy support, like the EVs and the hardware and the robotics? Or do you think China actually has its own mayor and certain sectors are being overlooked? And in kind of comparing that to where you’re seeing the US in terms of the companies you’re investing in, what are things you can actually learn from? I think we can talk about that a bit.James Wang (40:44)Yeah, I mean, in terms of China, there’s definitely a strong advantage in some of those physical industrial areas, ⁓ including say industrial AI, because it doesn’t exist in the US. The US doesn’t really build that many things anymore. It’s really hard to actually get any sort of manufacturing up. I can tell stories in terms of lioness. I can tell stories in terms of some of the med tech companies have helped try to like figure out where to do manufacturing. Essentially, you can’t do it in the US.So all of those sectors and areas ultimately do end up being a very strong advantage for China because it exists there and it doesn’t exist here. In terms of like other things that are overlooked, mean, China and actually Asia in general has had an interesting brilliance within consumer, like the consumer sort of trends there, the apps, the like different ways that, yeah, sure. Like WeChat and other things like.Overall, like China has a much more interesting grasp of like some of the consumer landscape than the US has. I think the, the, I wouldn’t say it’s the, it is, it’s like the stereotype is essentially US companies are the ones that do enterprise SaaS. And the other side of it, which isn’t really spoken at least around here is actually China and actually a lot of Asia is really good at consumer.A larger consumer market, maybe you can argue it’s like some aspects of that, but maybe you can argue it’s some aspect of taste as well. That may be changing ⁓ over time, especially as like the two markets are more and more divorced from each other. ⁓ Ultimately, the U.S. will probably have its own like consumer ecosystem because it’s divorced away from some of the Asian companies in China in particular. China will get separated from like the enterprise sass in the U.S., in which case there has to be its own stack.So maybe that will change over time, but there definitely is like sort of a strong, I wouldn’t say internal cult. I wouldn’t say like cultural from a cultural perspective, but more like cultural from there have been entrepreneurs, successful tech companies and sort of playbooks on like, how do you do this? That are much more mature, say in like the Chinese ecosystem than in the U S ecosystem. Well, you know, the last big consumer app was Snapchat and before that, you know, Instagram and Facebook rewinding to the dinosaurs.Grace Shao (43:06)Yeah, I think that’s interesting. And I think I’ve been hearing more from founders in AIPAC, not just China, but they’re saying with AI, they actually think there’ll be more opportunities in the enterprise space. And the reasoning is because a lot of the reasons why China or South Korea, a lot of these countries at that time, I would say in the 90s, did not want to adopt a lot of SaaS from the US was actually because they frankly didn’t even have the infrastructure in place as in you know internet was not that like you know, you took What’s the word for it ubiquitous and then it was like they don’t have the ability to actually even Serve the you know the need and then on top of that It’s not just China that has a lot of SOEs actually a lot of Asian countries have a lot of state majority companies and I think people again in the West might not realize thatAnd because of that, there’s actually a lot of concern on data privacy issues. And they’d rather have maybe even a shittier quality product than actually jeopardize their data being ⁓ taken by someone else, a third party. So that also goes back to why a lot of companies now in Asia actually want to adopt open source AI versus sending their data across the world. Not quite literally, but giving that re-access to like say a chat jpg or a google gemini so i think there’s a lot of reasons why people might find more use cases of sas ai in asia now ⁓ and they might actually you might see more entrepreneurs in this space popping up ⁓ i think on that i want to ask a question on vertical ai i think we kind of touched on healthcare right now we’re talking about there’s a potential growth in enterprise softwares are ai empoweredHow do we understand vertical AI? Will they still basically have to rely on the incumbents ⁓ infrastructure to build their tools? should we actually kind of see them grow out? should they be compared to the Googles and the Microsofts of the world? Or should they have their own kind of racetrack themselves?James Wang (45:19)I think it’s more their own racetrack because it’s kind of a very different ecosystem based on how AI is evolving. So I think the first layer to talk about is just do any of the LLM model companies have a strong, durable advantage outside of, know, OpenAI and chat GBT has a big brand. Google has a lot of distribution. Alibaba has a lot of distribution. It’s like, is their durable advantage? All of their compute tends to be the same.You know, they’re using the same GPUs. They’re mostly still on CUDA. Maybe it’s moving a little bit, but they’re mostly still on Nvidia GPUs. The models are basically the same. They’re all transformer based models, essentially. And their data at the end of the day, scraped from the internet, is largely the same. Yeah, maybe it varies a little bit, but it’s largely the same. They’re ultimately going to be very, very similar. That just doesn’t give you a lot of differentiation across that. that goes towards like, well, ultimately it might be capped in terms of how different these things are.That goes into the difference between that and the internet, right? The internet was very much an aggregated landscape where it’s like everything’s on the internet. You can go find, like go through to Google, whatever, find whatever you’re looking for. You can go to centralized marketplaces, you know, how about like Amazon, whatever you can find your things. And ultimately like it very much like put people into the same place.What AI is doing, interestingly, is if you think about where there’s actual durable advantage, it’s probably still the computers largely going to be the same. The models in terms of deep learning models, whether they’re transformers or not, are probably also going to be fairly similar. The difference is going to be in the proprietary data. And once you actually get down to the proprietary data level, we’re not just talking about, within your individual corporation, you have your OK.Internal documents or whatever fine. That’s like one thing Maybe you can still use like chat GBT or like one or whatever like some sort of open source LLM But what if you have raw acoustic data from ultrasounds to be able to detect liver disease? You’re not gonna be able to put that into chat GBT at the same time like that data Right now with the current models and compute you don’t actually need that sophisticated of a modelor even that much compute to make it actually do something really interesting with where AI has advanced to at this point. The same has happened with a lot of the drug discovery area, material science area, industrial AI, which again, China has an advantage in terms of this with their actual like running operations and data gathering exercises there. But that’s where the vertical AI comes into play. If you essentially have the ability to have this proprietary data that’s very expensive and difficult to generate,Grace Shao (47:40)Yeah.James Wang (48:09)that you have a great proprietary source for, you can build a durable advantage with your specific models there. So that’s where we’re seeing a lot of split in this area. And the way that this will work is less like aggregation, like the horizontal aggregation of search engines or marketplaces. Instead, what you have is a lot of vertical productivity, where you can build.Like say for drug discovery, can very easily imagine where if you can make a huge amount of productivity increase in drug discovery, that is a massive market, even if you stay just there forever.Grace Shao (48:46)Yeah, and it’s not like these big tech incumbents are going to move into that space. Frankly, it’s not it’s not just like your point. It’s not even just having the money to buy these data. It’s also having the know-how and the like decades of, you know, data gathering that you had to collect. And you can’t really get that immediately right now, right? From the market, from just like the public market or anything. It’s very different from the data you’re scraping from the Internet. Yeah.James Wang (49:09)They also do have their own flywheels that prevent followership. So the flywheel is specifically this. You get the data when it’s cheap because no one thinks it’s valuable. We have one company that literally did this in terms of raw acoustic data. No one usually keeps raw ultrasound data. You usually just have the images, if that, that is kept. No one keeps the raw ultrasound. They bought a lot of it.Afterwards, in terms of other AI companies that come, they go, ⁓ actually, this stuff is super useful. We didn’t realize that they also approach hospitals and say, let’s buy this. The hospitals wise up, right? They’re like, ⁓ this is actually useful. We’re not going to sell you it for next to nothing. We’re going to sell it to you for a lot of money. In the meantime, the other AI company has been using this data to generate revenue, to do these things, to get to a certain level of quality, such that now it’s like, the gap is getting bigger and bigger. The data is getting more and more expensive because it’s becoming like clear that it’s valuable. At the same time, the level of productivity, the level of quality of your AI is likely going to lag the incumbent that already gathered a bunch of data and is generating data at the same time and continuing to improve the model. So they have their own flywheel effects with these as well.Grace Shao (50:24)I see, see. And I think, okay, I don’t know if this is too small of a use case, but I can already imagine, like I just had a baby, right? And then, you know, we went through the private clinic kind of system in Hong Kong where you basically get, they know you have corporate insurance, so they just like make you pay it a ton. And you go in every month, which is absolutely ridiculous. But at same time, when you are a mother, you want to check in and see if there’s anything that needs to be, you know, there’s any flaring issues. Whereas I know in the public sector, if you go to the public system,You don’t get like ultrasounds until you’re what, like three months into your pregnancy, 12 weeks in. You don’t get checked very regularly. It’s usually every trimester maybe once or twice max. ⁓ So yeah, like for something that’s not simple, but as predictable as like a healthy pregnancy, I can imagine if you have an AI telling the mom and you just pay like 20 bucks a month, like everything looks normal, everything looks fine.I can totally imagine that being quite useful for the general mass. if they flare up anything like glaring, that’s something that needs more attention, you can then go into a medical staff’s office for proper checkup.James Wang (51:27)Totally.Right. Well, for one, congratulations. But for two, there’s all there’s so think about it. So one of our companies actually has a product like this, not for the consumer. But ⁓ I’m sure after the baby was born, the doctor did a child hip dysplasia check. So basically checking whether or not the child has this specific condition where if you catch it early, all you need to do is it’sGrace Shao (51:41)Thank you.Mm-hmm.James Wang (52:07)not all you need to do. It’s kind of annoying. You do have to put braces on the child and everything, but it fixes through a few months the problem for the child’s entire life. So it’s very worth it versus having a condition the child’s entire life. One of our companies, ultrasound company has an AI. Basically their end to end product is to do a quick sweep with the ultrasound where it tells you hip dysplasia. Yes, no. So in terms of that,Grace Shao (52:29)Mm-hmm.Yeah, actually, my baby, we had to go through for that and you have to go to like you have to wait like two weeks for the specialist to check you and the specialist you have to wait hours in a clinic like it’s a long tedious process and it’s expensive. So I can imagine this just being lot more affordable and you can actually deploy it to the mass like across like mass market a lot faster.James Wang (52:46)Yes. Well, even with that, it’s just like, again, taking it from like a, I don’t know, evil commercial VC hat, but not really. It’s just like, think about it from a hospital’s perspective. That check is really easy to miss. Like you have checklists, you have all of these things, but you can forget quite easily. That is a huge impact if you forget and it isn’t caught, right? At the same time, it’s like, it’s a doctor that needs to go do the thing. It’s a very valuable resource that needs to stop get sectioned out to go do the thing. If you are a hospital and you are able to basically just have a nurse do a quick sweep and scan and tells you yes, no, and you’ve checked it off the box, you’re probably willing to actually pay a lot for that at the end of the day. And it actually is more beneficial for the consumer too because it is actually doing the thing, super important. It gets done. It’s very accurate.A lot of the vertical AI areas are this way. like, it’s not just productivity increases from the, it’s not just like cost decreases maybe is how I see it. It’s not just the cost goes down. It’s just that the quality of it, the productivity of it, the like accessibility that goes up. And if anything, a lot of cases, the hospitals that whatever the vertical AI case is perfectly happy to actually pay up to their cost, previous cost of the thing.because it’s just so much more reliable, easier, and takes away other workflow concerns. So that’s why a lot of this vertical AI stuff is interesting. Its individual price tends to be actually higher, which is not what you typically think with AI, but it’s just more accurate, easier, smoother, and better for the workflow.Grace Shao (54:25)Yeah. I can see that. Yeah, that’s really interesting. I want to talk a little bit about the big picture policies between China and the US right now. I know you invest across the stack, including some of the infra stuff. When we talked about earlier, you said, look at data centers as well, So help us understand this. Data centers aren’t new.Like, you know, AI needs a lot of energy. AI needs a lot of data centers. How do we understand the relationship between these moving pieces?James Wang (55:11)Yeah, mean, the big thing is AI data centers tend to take a lot more power ⁓ in general for lot of the internet services and other things. ⁓ Because if you’re using GPUs inference, these tend to be much more power hungry components. For internet services, part of the reason why SAS could basically make money with queries that are fractions, fractions of ascent.is because essentially it’s almost free. Like you can use a lot of the way that the internet worked is much more around uptime. So if you actually have, and this may get a little bit technical, if you say have like AWS cloud provider share resources, you’re able to surge up and down your capacity and share it across in terms of virtual instances. actual cost of service, a lot of websites, even massive ones is actually not super high. It’s only high from the perspective of like it may be millions of dollars.But then again, you’re making billions of dollars off of your service that you’re servicing it from. It’s actually not very high. For AI in general, ⁓ its inference costs have been dropping a lot. But even so, with larger models, with needing a ton of memory, with needing a ton of these different things, with GPUs that themselves are both power hungry, but also heat, generate a lot of heat, you basically need to spend the currency of AI is essentially power.Grace Shao (56:21)Mm-hmm.James Wang (56:38)You need to spend power to literally power the GPUs or whatever XPUs, like TPUs, whatever thing you’re doing to run the AI. And you also need to cool it, which also is generally active, which means it’s also power. So all of it boils down to, okay, we need to spend power to be able to do this thing. It’s the closest thing to it is actually like cryptocurrencies in terms of you actually think of the one-to-one translation between power and actually the thing, ⁓ like what the thing does. So.Because of that, the sheer density of power requirements means that usually some of these data centers that are trying to serve AI might exceed the power able to be provided from a local grid that was otherwise serving, just like city, resident, and like normal kind of activity. And you are seeing a lot of these data centers for that reason basically doing their own power purchase agreements.having their own power plants. So they’re not actually on the grid, but they’re basically connected to their own power plants or connected to some of these power systems that are not within like say residential grids or something like that. So that’s been a big part of like why AI has needed that.Grace Shao (57:38)Mm-hmm.As like a average user of AI, should that mean that we should just be more mindful and not use so much AI? Or does that mean that the future of energy consumption will drop as technology advances? Like, how do we understand that? Because like, you know, when we use the internet, it’s not like we think about, my God, how much power consuming, right?James Wang (58:11)Yeah. And the thing that I said before was if you take various stats, it’s somewhere between like 70 to 90 % decrease in inference cost each, like each year. So why haven’t like, you know, inference costs falling through the floor while we’re getting more advanced models, we get reasoning models, which actually use way more tokens or words in order to spit out like the same number of tokens that you see.Grace Shao (58:36)Yeah.James Wang (58:38)So we’re using more and more and more. And that’s why, even though the cost has been dropping so rapidly, we’ve basically kept pace or exceeded it in terms of power. That being said, there’s a question. Where will some of that power requirement ultimately go? How much will be needed? And yeah, will it be the case that we end up just needing exponentially more power? So there’s actually a piece on my sub stack that aa hedge fund buddy of mine, hedge fund friend of mine from Bridgewater wrote, he does a commodity hedge fund now. His point is actually, even if you take very aggressive estimates as for how much power needs will grow for AI, it’s around like a 3.5 % incremental. That 3.5 % is basically the growth rate that we had during the 1950s in terms of the US power grid growing.That can pretty easily be hit by renewables, which have intermittency problems. So you basically need battery storage, which is why we also invest in stationary batteries in that area. Or it can be hit by natural gas, or it can even be hit by just retiring coal plants slower. So actually, a lot of the power needs are not as insurmountable as you might think. And I personally suspect it will ultimately be the case that as we plateau in terms of, hey, this thing likeWe don’t need it to like give us like, has much reasoning anymore. We just needed to book us vacation tickets or something like that. That’ll ultimately level off while the requirements in terms of compute costs, in terms of power costs will keep falling too.Grace Shao (1:00:14)I see, I see. And I think it’s also interesting, so just spoke to David Fishman recently. He’s an energy expert on the China space. And he was saying that, like he kind of mentioned in passing the US side, which is like essentially the US energy kind of, I guess conundrum is more exacerbated becausethe center of living has not increased drastically. So people’s consumption of energy have not actually increased drastically. Whereas in China, over the last two decades, energy consumption has been increasing anyway because of urbanization, because of modernization of maybe your home, the economy as a whole. So there’s been more energy planning in China to actually support that kind of energy increasing demand.And when that AI is now part of the picture, it doesn’t feel like a sudden gap that needs to be filled because you have the renewables, you have small nuclear plants being built out, et cetera. So it’s interesting to hear your perspective that actually the increase in demand, the increasing energy demand is actually not that significant. I think, again, headlines of news articles often really highlight that and really showcase a different picture where sometimes it’s more about like, OK.People are experiencing higher utility bills. The grid cannot actually support local economies or local people’s livelihood anymore. It seems like it’s causing a big issue for the average citizen. ⁓ But yeah, thank you for putting that into perspective.James Wang (1:01:42)It’s totally, well, it’s a self-inflicted issue on the US side. Again, like the US has expanded faster than that at periods in its history. There’s a lot of different energy sources that you can actually use to go after that. It’s just the problem is political in part, like the US has a lot of bureaucracy red tape that’s hard to cut through, in which case it’s hard to build anything economically in the US, which is part of the problem. There’s no nuclear being built.Grace Shao (1:01:47)Mm-hmm.James Wang (1:02:09)So like you’re saying, China is actually building nuclear at a pretty rapid clip. The US is at best unretiring or maybe retiring slowly its existing nuclear capacity. It’s actively retiring its coal capacity, whereas China is what building a new coal. I think it’s one or two new coal plants every week or something like that in terms of the pace. like it’s just a very different kind of environment.But it’s also not because yeah, the U S has no technological ability to go after that. It’s yes. Like you’re saying it’s like we have plateaued and a lot of our energy use. There’s also been a big push towards green renewable energy sources, which especially with the U S grid, low power storage, ⁓ it has its own challenges and can’t actually do the base load for AI. So if we wanted to, the U S could actually pretty quickly solve its problem. The question is, is there the political will and is there the willingness to stomach some of the trade offs for sake?James Wang (1:03:09)higher carbon cost.Grace Shao (1:03:11)Yeah, and I think that’s something David talked about as well in that episode ⁓ where it’s like the trade-off in China is more like, okay, we need more energy so we build more coal, but it doesn’t mean that we stop our renewable. But just because we have renewable doesn’t mean that we stop our coal. The trade-off obviously can be criticized, know, environmental issues, pollution, et cetera. But again, it’s just state level, I guess, mandate or state level priorities a bit different. ⁓So we’re not a political show. We’re going to move on from that. I want to ask you about the Inflation Reduction Act. So this relates to what you just talked about, a lot of the push on renewable energy. then Trump kind of taking it 180 degree on this. So the IRA was introduced in 2022. It tried to make solar and wind more affordable on the grid. How did that actually work out?What does it mean now with the Trump’s one big beautiful bill? Give us a high level explanation what’s happening there.James Wang (1:04:09)⁓ let’s see, data center developers keep getting whiplash in terms of renewables being good and then bad and then maybe not so bad, but not good either. Something like that. I think that’s sort of the quick high level. I mean, so, ⁓ a lot of the incentives, ⁓ were definitely something, things that a lot of data centers, lot of other folks, like hyperscalers tried to take advantage of, ⁓ when the inflation reduction act was more the law of the land before, you know,Some of that got thrown out, big, beautiful bill, et cetera. ⁓ But I mean, the big challenge for the US, though, even just stepping back from that, is regardless of how much legislation you throw at it, it’s just like the CHIPS Act, right? You can throw as much legislation at the CHIPS Act to say, we’re suddenly going to build all our chips in the US now, or something like that. And it’s like, well, ⁓ you’re not spending enough money to do that.And also legislation doesn’t like magically change things unless it specifically hit some of the core problems, which is yeah, the US doesn’t have for chips is like the US doesn’t have enough like labor for this like expertise moved over for sure. It’s ever for the power side. The problem actually goes back to the same thing we just talked about. Transmission interconnects lines old, hard to do, lots of red tape, lots of bureaucracy. It’s hard to build much anywhere.unless you’re building in places that might not actually be super optimal for say like data centers. So, you know, some of the South in terms of Texas or Southwest has been more amenable to some of the data center and like power build out. It’s also hot there. It would really be nice to put it in a colder place. So you have less power needs to cool the thing too. ⁓ The bigger story, I think with all of this, there’s been a lot of legislation that the US keeps throwing out.Grace Shao (1:05:54)Yeah.James Wang (1:06:00)Maybe the bigger story I’d say is just the legislation has done some things around the margins. It has not made like a huge 80 20 change, at least from what I’ve seen. It’s like the same problems, the same ultimate macro problems that plague the U.S. and building stuff. And also it’s aging power grids and interconnect problems between different grids are still the same ones, like regardless of the legislative regime that we’re in.Grace Shao (1:06:24)Yeah.An agent issue with the grid is actually like also just a reflection of like, frankly, the US developed and modernized so much earlier than China. And the grid just by nature is older and therefore the capacity and capability is like weaker because technology advance. Right. I think sometimes people forget about that. Just the reality that China didn’t become China that we know of today until like this decade. And the US has been basically the US that we know of today. The last four decades. Right. ⁓Grace Shao (1:07:33)Then I have one last question for you. I have one last question for you and it’s a question I ask every single guest that comes on the show, which is what is one differentiated view you have? Our show is called Differentiated Understanding. It’s about how you piece together the information you have and how you form a differentiated view, right? So what is something that you think is a bit non-consensus or against what the majority might think?James Wang (1:07:35)Sounds good. Yeah, I mean, I probably would have said it was my view about the vertical AI thing before, because I was talking about that a lot earlier than a lot of other folks, when there was still the talk about foundational models, which still is somewhat talked about. People are really pushing that a little bit less, that foundational models will cover every single use case in existence. And I think there’s been a lot more consensus moved towards that. So maybeThat was a very non-consensus view I had. The consensus has moved more towards. Let’s see, is there any other big non-consensus view right now? ⁓ I think I have one, actually. So another one. So my personal take, because of the way that LLMs have developed and everything, and a lot of the different AI areas have developed, I actually think a lot of the valueof AI from a GDP economy, et cetera, perspective will ultimately be socialized. I don’t mean that as in the government will. Yeah, I don’t mean the government will take it and redistribute it. I don’t mean like something will happen from that perspective or socialism will suddenly take over the US or something like that. What I mean is in terms of economic theory and whatnot, you can either have excess profits be captured by specific corporations and companies.Grace Shao (1:09:03)What does that mean?James Wang (1:09:25)which is frankly as a VC what I’m trying to invest in and basically have essentially monopolistic power, whatever, and base essentially have a lot of rents from society gathered towards the corporation or the company, or you can have a go to labor or you can actually have that value be socialized. Meaning because of competition, because of diffusion of the technology, because it can’t be controlled as much, it just improves society’s lives.and isn’t actually excess captured by any single company. Even though like we have these huge model companies, they’re absorbing a lot of money, all these different things are happening. My personal take is like, they don’t actually have such strong barriers. Do I think OpenAI will go to zero? No, I think they have a pretty strong consumer brand. Do I think Google will go to zero? No, they have a lot of things to like distribute out. There’s a lot of uses for it. The companies will still survive, but they won’t become like essentially like world like consuming companies in the way that some people have talked about AI or talked about AI as in it’s a sector where a couple of large companies will suddenly take over everything. I actually think AI will diffuse within the economy quite a bit where we’ll use it in our everyday lives, but we won’t necessarily need to pay a company a huge amount to do it. For example, in the future, you might have edge models that just run on a very like a fairly powerful inference chip on your smartphone.And you don’t need to pay ChatGPT or anyone else for that. It’s just something that makes your life easier, better. And it’s just there. So that’s one of my takes. I actually think the majority of the value hard to measure as that is will probably be socialized.Grace Shao (1:11:07)That’s really interesting. think that reminds me of something I wrote about recently and I think we engaged online about this as well, which is ⁓the diffusion of AI will in some way look like the diffusion of internet, where it’s not like we just think of four companies as internet companies anymore, but even the tangible real world. Like, you you think about food delivery, you would have never imagined a food delivery company is an internet company. However, it is an internet company these days, whether it’s Food Panda or, you know, like Maytwan or, you know, Seamless in the US, that’s actually like...not a physical world business only, right? And like when you think of a ride hailing, when you think about even like, I don’t know, apartment hunting, whatnot, it’s not limited to just the physical world. Internet companies actually encompasses all these things that we do. It’s just become the infrastructure. So you’re saying AI essentially will just be part of everything we do and it’ll be empowering everything we do. And it won’t just be limited to like the five companies that we think about nowadays. Yeah. Cool.Thank you so much, James. Really, really, really helpful, really insightful conversation. And I really enjoyed talking to you.James Wang (1:12:16)Enjoy talking with you too, this was great, thanks so much, Grace.Grace Shao (1:12:19)Thank you.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Unlocking the Future of Startups and Super Individuals with Bei Zhang 26.11.2025 42min

  In this episode, I speak with Bei Zhang, VP of Growth at Tanka, about the company’s mission to empower AI-native founders. The conversation covers why persistent, organization-wide memory is the missing ingredient for truly proactive agents, how Tanka stitches together chat, email, calendars, and documents into a single “remembering” teammate, and what agentic work could look like over the next 12 to 18 months. We also take a closer look at the future of founding teams and how agent tools can enable a super-individual way of working without losing control, auditability, or taste.Tanka sits inside a three-layer stack incubated by Shanda Group. EverMind is the AI infrastructure arm that builds a long-form memory orchestration platform. MiroMind is the research lab, built on Qwen models, focused on long-term memory and reasoning. Tanka is the consumer-facing agentic workspace that applies those capabilities to help startup founders run their day-to-day.All three were incubated by the family office of Tianqiao Chen, the Chinese internet entrepreneur and investor behind Shanda.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Topics we covered:* Tanka’s Mission: to empower future AI-native founders to transform their ideas into successful businesses swiftly and efficiently. * The Problem Tanka Aims to Solve: Founders often struggle with information overload, with critical insights scattered across various platforms such as Slack, Google Drive, and numerous AI tools. * How Tanka Works: Tanka’s unique AI memory framework.* The Team: Tanka’s diverse team is rooted in the heart of Silicon Valley, comprising individuals with rich backgrounds in big tech and startups. * Competition is not with general agents—focused and niche market.* Connecting Founders with Investors: It actively seeks to connect founders to investors, creating a community and offering consulting services as well.* Risks of Using AI agents: Human quality control remains essential; a hybrid model is a sustainable, long-term work model.AI-generated transcriptGrace Shao (00:00)Hi, Bei thank you so much for joining us today. I understand you lead Tanka’s growth right now. It’s a very, very exciting startup. I’ve heard a lot about it. Why don’t we start with your role and just tell us about the company, Tanka, the founding mission, what problem you guys are trying to solve, and just a bit about the team.Bei (00:16)Sounds good. Sounds good. Hi, Grace. Thank you for having me. Hi, everyone. My name is Bei. ⁓ I lead the product and growth in Tanka. Before I joined Tanka, I had been in various roles in different AI and SaaS companies, mostly in the GTM function. So what Tanka is about? Tanka is on a mission to empower the future AI native founders to go from ideas to founded very fast, very efficiently. And the core technology we’re putting behind the Tanka is the long-term memory behind the agents. End of the day, we’re trying to create a proactive companion, or we say that AI co-founder, because compared to typical AI chatbots, we are putting more power behind our AI agents that can remember all the conversation, remember all the relationship, and eventually can be the proactive.AI partner to propel the founder to move as fast as possible. So that’s essentially our mission. And we’re hoping we believe the future is the world of super individuals and the lean teams. We’re trying to make the Tanka to be the powerful operating system for the future startups.Grace Shao (01:25)So I think there’s some fun and irony in that, right? what does it mean really when you say it’s an AI co-founder? Like for someone like myself, I’m a independent or I would say like a founder of a startup, I have a small team. What is Tanka really helping me do in a very practical sense?Bei (01:43)Yeah, yeah, great question. We are essentially the kind of startup, so we help ourselves, right? We’re trying to leverage the resources to help others too. what it would mean, maybe we’ll take a step back to get down to the problems we’re trying to solve. Essentially, we being the center of the Silicon Valley, we’ve been hanging out with a lot of founders, or a lot of individual, a lot of lean teams, a lot of them are just like you, Grace. ⁓ You are a super individual. We also have friends being just a three to five person teams. And the common problem we’re seeing they’re facing is the highly scattered information, overload of information, a bloat of different AI tools, and a very spread of key knowledge across different platforms. So even though we’re saying we’re putting the many of the platforms are wonderful. You got all the nice conversations on the Slack. You got all your documents in Notion and the Google Drive. And there are some offline chats. I’m sure there are valuable informations embedded into various GPT tools or AI chatbots. So the core challenge is not having the right tool. The core challenge is when founders are all of a sudden going from a single threat, trying to take on the world, trying to build a business, the tendency is that there is overloading of the information from all kinds of directions. Because for example, we’ve had a very good friend being a very technical researcher in Stanford. But the moment when he or she step into the founder role, he or she will have to handle not only the product, but also engineer the sales, the marketing, the product dev, legal and tax and BD, right? All kinds of stuff going on. So essentially having all those information scattered in different places create a few effects. Number one, it create a huge overload on the human brain, right? Nobody can process the information so effectively. Especially, we even come across multiple founders doing multitasking because they are trying different ideas, right? Which they will just multiply the pins. And separately, when the brain is overloaded, it instantly distracts the founder from the core duty, which is building the product. So that is causing many problems to be happening. It is causing key information getting lost. It is causing one part of the valuable information not necessarily getting fit into the other nice tools or very powerful AI agents, so the outcome isn’t as optimal. It’s far from it, right? The outcome is far from optimal when they’re trying to make a progress on the project. So that’s the mission. That’s what we’re trying to solve in Tanka. So in Tanka, here are a few things we’re trying to tackle the problem. Number one is the AI memory. Without putting the fancy word out here, just thinking...As of you have, let’s say, today, whichever, most of the AI tools are not really memorizing your conversations. Because when you open a window, it has a conversation with you. But the moment you close the session, it doesn’t really record anything. So the next conversation is new. So with the 10Cut AI memory framework, all the conversations and all the documents you put in the tool are automatically compressed, stored properly, and also stored with a high fidelity so that when you have a conversation once, the future conversation will always remember what you had before. So it put a piece of mind to founder’s head so that you know there is a trusted partner that never forgets anything. So every company is about moving forward, not to remember what happened in the past. So on top of that, we’re adding the connectors, making sure Tanka can digest information not only happening within Tanka, but also connected from other sources as a deep memory and context. And with the memory, we’re able to put in the right AI agents, whether to produce the business plan, whether to just do the deep thinking and a deep conversation, or whether to produce an investor-ready pitch deck.They are all based on the actual information in greater details, without you having to chase across all different things. So that’s what we say. That’s the actual specifics we’re putting in behind the tanker, because we’re not calling that just, we want to go beyond the typical AI assistant, because when we say AI assistant, meaning there is some, it’s a reactive, right? There is a AI sitting there and waiting for me to ask the questions or waiting for me to give the proper prompt. So we almost have to treat the typical, even for the very powerful AI chat bot, we have to carefully curate. We have to carefully protect the conversation, making sure it doesn’t generate anything wrong because garbage in, garbage out principle. But with Tanka, because the more you work with Tanka, the more Tanka knows about you, we almost can forget about prompting. It is an actually intelligent person sitting right next to you as a founder. So whenever the conversation happens, we just keep marching forward. And we’re even building more of a proactive AI functions because now that Tanka knows everything, what do we have happening in theory? You should know what I need to do next. even before, in theory, even before I ask,Tanka to do anything, there should be more proactive actions. For example, hey, I need to follow up with certain investors. I need to update the pitch deck, for instance. Some of them are already realized, and many are definitely on the road as we speak. But that’s what we mean by AI co-founder, because we want to essentially have an AI that can essentially propel you to go forward instead of just waiting there for you to comment the way I do things for you.Grace Shao (07:43)That’s super interesting. think to me, when I heard that, I was like, that’s going to be so helpful for me. Cause like you said, there’s so many to do things on the to do list every morning. And then if someone’s actually proactively reminding me or getting things done, that would be really helpful. I first want to talk about the team.First before we get into the product. Just like I understand you guys have a pretty diverse team. A lot of you guys, ⁓ including your founder, came from even ex big tech. How did your team come together? What’s the background? And I guess what is your edge right now making an agentic tool like this, especially with a lot of even the big AI labs are pushing out agentic tools. Like what is your niche and edge?Bei (08:05)Yeah, yeah, great question. So you’re right, we’re a very diverse team. We’re headquartered in Redwood City, California. We do have a global team across different parts of the world. But the core leadership and the product team are right here located in the center of the Silicon Valley because we are a company building for the founders. We want to be where our customers are to shine light on a few other things you covered. When Tanka was born, essentially it was born within a family office that has been actively curating multiple companies and also has been actively investing in hundreds of early stage startups. All the memory problems and all the context switching, all the information overload are very much experienced firsthand, both for the funding members within the family office and also being well observed by the company, right? The family office has been investing and curating in. So it’s a common problem that hasn’t found a solution yet. So that’s where I would say one of the edge is our deep understanding.We’re not an enterprise tool and we’re not so much to a pure consumer tool. We’re living in a breathing in the startup world because the people has been working in the company or surrounding the company has either been advisors, investors, ex-founders of this kind of startups. So we know the problem from a different angles. So that’s number one. And number two is you’re absolutely right, the CEO, Kisson.She came from a Meta, from TikTok. So definitely had a good discipline and a very structured approach from well-formed companies. She also co-founded another company that has a similar form of Tanka. So she brought in tremendous discipline in both the AI agent and from 0 to 1 and from 1 to 100 scale.And I personally come from Grammarly. I happened to have an experienced growing company at a scale and also helped establish the B2B function from the beginning. And other than that, we do have ⁓ members coming from various startups. So we have all been experiencing the problem, first hand, left hand, right? So that gives us a deep understanding on what we want to solve for ourselves.Grace Shao (10:53)But $29 a month is quite steep, let’s be honest, especially if founders are cost-conscious. I want to understand what was the thinking behind that. And again, how does it compare with peers, even more general AI tools like Manus coming out of Singapore right now, obviously, as well as the incumbents that have been integrating AI into their apps like Slack, Salesforce, Microsoft Teams, even Zoom AI companion, right? Like in some capacity, they’re all trying to become a more proactive, I guess, whether you can call it a co-founder or a colleague per se, they’re all trying to be there to be more present to help you actually get things done, right? How do you compete with such an array of competitors, essentially?Bei (11:36)Yeah, yeah, good question. So to your first question about pricing, we put out a pricing more to create ⁓ a sense of familiarity to begin with. So purely on the number, I think it’s a mid-tier. It’s not that high. It’s not that low either. But it’s something people can, our users can correlate to.And if you look at our free tier, we actually have a pretty generous free tier. We have daily bonuses. I think for lot of users to get a feeling, the free tier actually can get a lot done to truly feel the memory behind the agents. And also, separately, we’re paying much less attention on the pricing versus our attention on the value.Because at end of the day, what our users weigh in is how much benefits, how much value they are getting out of the tool. So we’re just putting the pricing as a stake in the ground. We’ve been doubling down on understanding what our users need. They need a collaboration, so we built the AI agents in the chat to empower the team.They need a generative function to turn the conversation into the actual shareable documents. So we did that. We made a very smooth process to go from the chats and the team conversation into the outcomes without you having to reprompt. The users are also looking for more help in the fundraisin, related features just so when they are ready for investor conversation, they can get it funded faster. So we have a whole pipeline of efforts to empower the founders to realize the benefits. in that, our goal is to make everyone feel like the price is a huge bargain. So that’s something we’ve been actively validating. And also separately, to your point, there are it’s an agentic world, right? Everyone, every company, whether the big ones or whether the startups are making various kind of AI agents. We do keep an eye on a lot of the big names, like you mentioned. I do have a lot of admirations to the great tools. But at this point, our belief is that in this age, the AI tool will come out in different formats and different forms.So I like to think of them as inspirations and role models, right? More so than the competitions. If they are doing something similar, right? We would say, how can we fill our own gaps, right? How can we do better than them? But often than not, actually have way more gaps. We think even this big names are not even addressing between on the path, right? Between the ideas to startups getting funded. So we’re hyper focusing on filling the gaps more so than worry about the competition. Because we believe the world is big. The world is big. In the future, everyone will be a builder. Everyone will be a founder. If a user don’t use us, it will not be because of a competition. It will be because we’re not delivering our promise and not creating the value for the users. So that’s where our minds are, mainly.Grace Shao (14:46)So instead of trying to compete on distribution reach right now, you’re really focused on serving a very niche kind of audience, right? And then really just delivering exactly what they need instead of a general mass audience.Bei (14:56)That’s correct. We’re not trying to build a tool for everyone. That’s the job for the big tech. That’s the job for Tech GPT and Cloud. We are in the center of the Silicon Valley. We are hanging out with all the founders who are using all the tools you’re mentioning, but are still struggling in pushing the ideas into tangible business plan. And even for serious entrepreneurs, they are very struggling in getting connecting to the right investors and getting funded very efficiently. So we’re just hyper-focusing on this persona. Because again, we deeply emphasize wisdom because we are them. So if we get this part of the job done, we’ll be very proud of this. We’ll be very proud of our efforts.Grace Shao (15:40)Actually, one thing you just mentioned, how do you connect these founders with investors? What’s the strategy there? Because that’s not a product strategy. Is that just your connection, your network?Bei (15:50)More so than that. So there are multiple approaches. ⁓ number, think about this in a few different approaches. So number one, this is actually interesting challenge because our founder friends are, most of our founder friends are struggling looking for investors and most of our investor friends are still struggling and looking for quality startups, even though they might be in the same room. So that’s still a ⁓ friction. we tackle this in a few different layers. So many of the founders are not effectively connecting to the investors because they’re not ready. They’re not ready. first, we want to make sure Tanka has the capability for them to chat with the team, for them to carry through all the conversations, and making sure all the minute details are reflected in the business plan and the pitch deck so they appear. They are more buttoned up.So that’s where we do the effort in preparing them to be investor ready, because investors are ready in the other end of the room. So that’s low-hanging fruit. And then separately, we are very active in the Bay Area funder communities. ⁓ So if anything, we have no lack of is there is an abundant funder communities here in the valley.And we’ve been actively in the community facilitating the conversation. We’re inviting investors to give advice on how we can build a tool to better empower the founders. We’re doing this in different directions. So in a way, by having a presence in such communities, we’re already acting as a connector between the two parties. And furthermore, what do we do have on the product roadmap. our features like investor database and the investor matching, because that’s a low-hanging fruit. We do want to provide the founders more value by making it very easy for them to see that based on their business plan and the sector, who might be the right person they should be talking to. we are also evaluating the options such as the data room analyzer or the even warm intros because we’re even discussing with the actual human expert fundraising agencies as a potential layer because we do believe this is, AI is not ready to take over the world yet, As awesome as AI can ever be, humans do bring tremendous amount of value. So on a needed basis, there needs to be a human layer on top of the AI workflow.And even if the human layer just evolved for 10 % of the time, we believe that’s where potentially the 90 % of the value may come from. So this is where end of the day we foresee we likely will build ourself into a hybrid solution where 90 % are conducted by the AI or focusing on this path addressing the problems many of the tools are really not addressing specifically.And we’re connecting the human brain, the different part of the party much, much closer in solving this problem. So yeah, does that make sense?Grace Shao (18:58)And you know where else founders should be talking? They should be talking on my podcast because that’s where investors are listening as well and media is listening. And that’s how you get your story out there as well.Bei (19:07)They should. Investors should be listening, too.Grace Shao (19:14)Investors are listening. Actually, my main audience are investors in the US and Europe. I think, you know, interesting founders should be DMing me now. But on a more serious note, I think you just talked about like, agents can do what 90 % of work, you still got to have 10 % of human quality control, right? So end of the day, what are things at least at this point, or the next, say, 12 months, we can delegate agents, what are things that we still really need that human touch or humanity to kind of guardrail, the kind of progression of technology or our workflow or the usage of AI.Bei (19:49)Yeah, we’ve been thinking this day in and day out. So definitely when it’s related to the information gathering, information collecting, the document generation, document refinement, and web scraping. So without saying the features, that basically meaning how you turn from your conversations and inputs, documents, team chats into the pitch deck, into their data room documents, and how to scrape online, how to go to the linking. Those delegatable missions, those missions that tend to be competitive but yet time consuming. If it’s a delegatable, if you can put into a ⁓ SOP or standard operating procedure, we should try our best to let AI to do this as much as possible.However, we do acknowledge that sometimes it takes a lot of judgment in this process because when the funders are so early, would the investors invest into the project or are they investing into the persons? Most likely, earlier they are, the earlier the investors are putting their weight on the persons. But many of the persons’ attributes and experience are not quantifiable.So there are certain things that I cannot build into the AI agents to automate everything. So that’s where we do need a human to better probably connecting with the macro, better putting in the latest reflections, and better just to step in, making sure we’re not misjudging certain startups in either of the directions. And also separately, I would say, we also, Even with all the AI tools out there, we also had very, very top-notch founders who are deeply in the research world. So they just don’t have time. They are very busy. They do want to focus on building their product. Can they learn how to do the whole fundraising business plan or so? They surely can. But it’s more valuable for them to focus on what they do best.That’s where I think sometimes often it just makes sense for the human layer to just step in and take it over. And it could also be entirely 100 % human touch, which could be well suited for the situation. But just wanting to make it possible whether the human touch is 0 % or 10 % or 100%, it is how this startup works. And we should build our product to be seamlessly connected and adapted to the reality here.Grace Shao (22:22)And I think it’s important to kind of note, like, you know, as you mentioned, as we’re all hyping up the AI agents right now, there is some mindfulness to be said to have to about the potential risks, right? So when people are using AI agents, I think this is as an AI agent question as a whole, not just Tanka but who audits the process ensures there are no mistakes, right? When the machines are starting to complete tasks, how do we actually ensure or how do we human ensure that we minimize the mistakes and the risks that they may come with.Bei (22:55)Yeah, it’s increasingly a more critical question as the adoption rate for the AI are increasing. So I don’t have a perfect answer. I don’t think anyone has really found the answer yet. I would say it’s the process. Process meaning when we’re building the product, because we’re building Tanka to be very deep thinking, deep researching, and working on very, very serious projects.We try to use our best model, most expensive one that does the deep thinking and the reasoning to the best extent. So we don’t try to save money using the cheaper model for faster speed, ⁓ which might be introducing more errors. We’re carefully balancing that. We would rather deliver higher quality at a higher cost, but for higher quality. So that’s number one.And number two is because that’s really actually where the memory comes in. Whenever we build a 10-cut AI to help brainstorm with the founders on next steps, we make sure it all ties back to the prior memory or it ties back to the traceable sources. for all the conversation and the generations, there is a link back to where you can point out to.But that being said, it’s not 100%. It’s not like we can disregard any human efforts not to look closely. We still are constantly calibrating, and sometimes errors happen. And that’s even because the LLM, sometimes because the core server, it has variations. Maybe a question from the same LLM vendor may generate different answers.One is more correct than the other one. So I would say it takes both efforts, even though that’s why we do want to emphasize the value of the human, because here’s AI. And we as a human, we still need to be very carefully guarding our own outcome. And then we introduce the human expert to further enhance the quality. I mentioned a lot of fundraising, and we actually have a lot of friends and mentors and advisors from other areas, such as sales and go to market, tax and legal, who are actually ready to engage and looking to find ways to help out the founders. So we’re not building us as a marketplace yet, but essentially we do want to, our vision is we do want to make a Tanka to be the center console where the founders work with Tanka, but also using other tools where it applies. We’re not here to replace anyone.And we would definitely encourage or we may build a bridge between the Tanka with the human experts so that the human and the AI and human harmonically work together to further minimize the hallucination and the errors.Grace Shao (25:40)That’s really interesting. didn’t realize it’s kind of like building up an in-house incubator or like a consultancy, right? Like you have Tanka as your main touch point, and then you expand into your human expertise. Actually on the technicalities, I want to ask what models are you using and how is that decided by the agent? What I put in a prom when I’m using your agent, how does the backend look?Bei (26:00)Yeah, so I’ll say, maybe without disclosing a specific model, we do use a combination of the top tier models. maybe that’s the best way to say it. Using the AI memory, I was too aspect. The AI memory layer is built a little differently. It is called EverMind. It’s actually went open source a few days back. So we built our own prior proprietary and memory layer using a set of the algorithm. And that’s one. And when we build our Tanka AI agents, we do have a router option. We do build a AI. We do have a few preset prompt. Whereas depending on the type of the questions and depending on how the different steps of the agents that can execute, you will automatically pick the best model for the task. So it’s not just one, one deal, right? The kind of large language model will vary. It depends on whether you’re asking to generate a rough idea or whether you’re generating a very buttoned up business plan. So it’s different. And separately, I do want to say because of the memory, that’s where things are a little different, right? So because we do have the AI memory,The large language model is capable of working with ever evolving context and the memory. So even the same question would absolutely yell the different answer the more you engage, the more you evolve with the AI. So I would say the LLM is a commodity. They’re very powerful. They are the necessity. But that’s where at the end of the day, we do think it’s probably going to be safe, whether you’re using Google or OpenAI or Cloud. At some point, it’s going to be indifferentiable, So that’s where, how to make sure it works for you, right? Not for a general purpose. It’s more critical.Grace Shao (27:49)That’s interesting. think that’s what a lot of the AI agents companies been saying as well. Like eventually, you know, the user experience will not, the users will not be able to actually differentiate which model they’re using, but it’s really just on how the interface interacts with the user and if it’s for a specific task. So I kind of want to go in on the product itself. Walk us through the product surface. Like, what is the experience like when I’m a user, I’m a founder, when I go on Tanka what should I expect?Bei (28:16)Yeah, we put in so much sense into the product, but if we, let’s say, we simplify, as a founder, you go into the Tanka, first of all, there is a place you can work with Tanka agent one by one basis. So on this cases, it’s essentially not too crazy different compared to the other AI agents out there, right? You still interact with the agent, you still ask all the questions, right?Further develop your initial idea into a very buttoned up plan and further refining and fine tuning on that. Again, the main differentiation is ⁓ our window never closes. Our window stays always on and never worry about missing any information. So that’s the one. And then let’s say you as a founder, you get an idea from ⁓ a raw impression into something more tangible, you need to work with your team, right? And if today, whether the team is your co-founder, or whether it is your friend or your son-in-law, right, advisor, there needs to be a joint effort because often the wisdom come up in the conversations, right? So that’s why we have a second portion of the tanker to be a chat, right? Whereas we, whoever you invite into the tank to discuss the ideas, to hear the feedbacks, whether positive ones or constructive ones, and whether you both share or you all share any external references. All those conversations are precisely memorized and processed to be the high definition by the AI agent. then whenThat’s essentially where ideally your business plan will evolve from your own work. And with the other AI agents, you would have to reprocess the information. You will have to bring all this conversation into a prompting and making sure, let’s say, that GBT understands what you have talked about. But it was tank up because the AI is sitting there. The AI is sitting there with you in the conversation. After you finish the conversation, after you are aligned,You and your partner or your mentor are aligned on certain solution. Well, you can simply tell the tech to say, go make the next version. In that case, there is no transfer of information. And then there is no loss of communication in between. So that’s the next step, because we see the collaboration being a very core part of the founder. Very few people can pull off the one person team, even for one person, assume, right? You as a super individual, you probably collaborated with many, right? To develop your own business, right? And the last but not least is we are building Tankard to be a very open platform, right? Because this is where we fully acknowledge that everyone will probably use some other tools, whether it’s Slack, right? Whether it is Minos, right? My favorite tool. Again, we’re not trying to compete, right? We’re trying to say, if those critical contacts happen in other platforms, we want to make sure there is a way to bring those contacts into the Tanka so Tanka agent can sync with more deeper memory in mind and thus generate more high quality contents, right? And then the other direction is also true because we actually keep the memory well organized. If at some point the organization or the startup outgrow the Tanka capability, and we are building the MCP to make sure all the memories are exportable to the next tools you’re trying to use. So we are here for the specific purpose. And then there is a beginning point and there is an end point. We’re not trying to do everything. Again, we try to do the best in the part of the problem we’re trying to solve.Grace Shao (31:59)That’s super interesting. I was just going to ask you, where do you think founders can outgrow Tanka? Because you’ve been really focused on saying, helping them out in the very early stages. it’s interesting that you’re quite mindful that eventually, if a company grows to certain size, there is potential that the company or the founder himself might outgrow your app and they will move on to the next agent, next tool. I guess on that note, I kind of want to end on a big picture question, which is,Bei (32:05)So yeah.Grace Shao (32:24)What do you think is the future of work for knowledge workers, especially startup founders, what you’re witnessing in Silicon Valley? I think you alluded to this a little bit, that there are more and more of these called super power or super one-person bands, whatever. But what should we expect? Are we still going to see the kind of startups of couple of people with different technical skills kind of coming together, founding a company, to scaling it?Bei (32:38)Yeah.Grace Shao (32:50)and then becoming a big corporation or are we going to see complete that mode, complete transition revolve.Bei (32:56)Yeah, it’s a loaded question. again, we’ve been very actively thinking along the lines of that too. So here are a few things we believe the future will evolve to. Well, there definitely will be big organizations. That’s just the case. Some businesses are better to be at a bigger scale. Let’s say if you build a robot company, you better be. You need a scale. However, we do see that with all the tools empowering people to go from ideas to the apps very quickly, we definitely see there will be exponentially more super individuals. And when we say individuals, it means either one person or either three to five person. Because eventually, everyone, we do see the traditional roles being very blurred, right? There will no longer be like a PM or front end or back end or marketer, right? Essentially one person likely that’s gonna pick up multiple roles, right? I assume, Grace, you probably were many, many roles at the same time as the owner yourself. I think that’s incredible. And then there will definitely, many of the businesses don’t have to be that big. We do see many companies will probably stay it’s pretty small, right, 5 % or 10%. For instance, Gamma achieved a $2 billion valuation at 50%. That’s incredible. And I think there will be more and more companies like that. So that’s what we are inspired to solve for them. And also, adding one more thing is we do think the future collaboration will be multi to multi, right? That meaning is no longer going to be one person being employed by one company for a long time. Because hey, when everyone can do so many things, if that person has a capacity, why couldn’t he work on multiple projects with multiple teams? That’s also where we are creating the tank to be not constrained by an entity. You don’t have to be the same entity because we fully expect anyone can work with anyone. And we want to embrace that and empower that too. And last but not least, again, there are many good thoughts. I think it’s probably a book worthy if we had more time. So I do think this is where, for the first time, in AI can, in the past, in order to value whether the workforce or organization, whether it’s effective, you kind of have to wait until the quarter end or year end to see the outcome, to see that. Because many of the information are not really recorded. But now, because everyone used so many tools, and also AI has a memory, and AI can understand how things work, I think the efficiency will be exploding. Because the AI is able to catch where the inefficiency, where the blocker is happening. That’s also, again, that’s why we built AI to be not just one-on-one, but to be in the team, just so AI can discover, right? It can observe what’s working, what’s not working, and making sure that the team always work. Whether your own team or whether the cross-functional team is always in optimal status before too late to essentially the performance review happening every second. So that’s also back to the super individuals, right? The super individuals can measure their own success in real time and furthermore be more successful.Grace Shao (36:21)All right, Bei, we’ve had a wonderful conversation. I have one last question for you, which is a question I ask every single guest that comes on my show. What is one differentiated view you have or something unique you believe in about the industry, about the future of tech and AI, or even something just in general in life?Bei (36:37)Let’s see, I have a couple, but I’ll pick one. even as an agent, we’re probably, I think it ties back to our conversation today, right? We are building, I am actively building the AI product, right? So we want to build a co-founder or even a super powered AI solutions. But I do want to acknowledge that the penetration and adoption of the AI in the real world is so, so low. And chasing after a technical advantage, going after, I think ⁓ sometimes it’s just almost a wrong direction for builders to say, let’s make this PowerPoint generation even more smoother or nicer. And while ignoring that, there are massive amount of human workforce are not even closely in leveraging even basic AI to do things. They are struggling with the basic data scraping. They are suffering with the basic information gathering and to be truly embedded in their workflow. This is actually tied back to our chat in the whole fundraising journey. I we’re talking about the most capable, the smartest, the bravest, the most ambitious founder who can build everything. But I mean, why are they still struggling in knowing where to find all the investors? Who is the right investor to work with? Am I ready for the investor conversation? What else do I need to prepare? How good is good enough? And what to anticipate?Why there are so many basic questions that are not solved. Sometimes I think it’s a, I don’t know whether it’s a differentiator. I just want to, we are doing practicing ourselves. Sometimes we try not to be ⁓ bad at in how we can build this tool to be better than the other competitors. But we go back to the basis on what problem are we solving? How is the problem, how people are tackling the problem today and how we can leverage the technology to best solve the problem. Because I definitely observe when we go chase after the technology advancement, we’re going after like 0.1 % improvements. But when we go back to the basic problem resolution, when we look at how the real world is being operated, we’re looking at like 90 % of the problem are not even remotely empowered. So that’s where I’d like to put out here.There is still a long way to go. And there are so many things to be built. So I’m very excited about the journey and all the possibility and all the value we can bring to the community.Grace Shao (39:11)Thank you, Bei. That’s really thoughtful. And I think that you do highlight a point where I think when we’re all so embedded in the tech and AI scene, we assume people are all adapting and adopting it. But to your point, actually, the general mass is really not up to speed with it. And there’s so much work that needs to be done in terms of educating them and actually working together and actually a lot of issues are not solvable by technology, but it still requires that human expertise. So really appreciate that. Thank you so much for your time today. Is there anything else you would like to share with us before we hop off?Bei (39:44)Well, first of all, thank you for the time. I love all the very thoughtful questions. It’s been a pleasure chatting with you and I’m grateful for the opportunity to organize the mind and the share with you and your audience as well. The last thing will be just any recommendation, any suggestions is welcome from you. I I hope this is a...This is the start of the conversation, more so than the end of the conversation. And again, you’ve been a super individual. I want this product to be helpful for you. And also, I would love this product to be helpful for your audience, whether they are investors or they are founders. So I’m just putting, I’m definitely very, very open. We’re a sponge. We’re a sponge. We’re here to take on any suggestions or feedbacks and that’s the only way we can get better and really focus on the right problem to solve and we need everyone’s help. So thank you, thank you, Grace and thank everyone in advance for all the nice thoughts. Get full access to AI Proem at aiproem.substack.com/subscribe
• From Coal to Compute: China’s Grid Meets the AI Boom with David Fishman 18.11.2025 47min

  David Fishman is a Principal at The Lantau Group who advises on energy development, infrastructure, and electricity markets across East Asia, with a focus on China. His expertise spans power-sector policy and economics, grid development, project bankability, and transaction support, backed by regulatory and economic intelligence across China’s solar, wind, coal, nuclear, hydro, transmission, and power markets. He has led work on policy forecasting and tariffs, renewable-asset due diligence, China business matchmaking, and green-power procurement for multinationals.In our conversation, David unpacks how China’s decades-long planning underpins its energy transition and how renewables, storage, and grid build-out are looking to be able to meet AI-era compute demand. We also touch on China’s East Data West Compute and how it leveraged strong geographical planning, as well as discuss the cultural and commercial reasons behind the global retail adoption of solar energy. For me, the most interesting point he brought up is that electricity used to be bound to scarce resources, but as the saying goes, the sun shines, wind blows, and water flows everywhere. Access to reliable power will become more evenly distributed, which can raise living standards in places left out of prior industrial revolutions - and Chinese technology is driving that change.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.ChaptersIntroduction to China’s Energy LandscapeThe Evolution of China’s Energy DemandNuclear Energy: Pros and ConsData Centers and Electricity ConsumptionMain Drivers of China’s Growing Electricity DemandChallenges of Renewable Energy for Data CentersGeographical Dynamics of Energy Supply in ChinaInfrastructure Challenges in Southeast AsiaCommercial Reasons for Renewable Energy AdoptionChina’s 2030 Renewable Energy Goals and BeyondThe Transition to an Electricity CivilizationTranscript generated by AIGrace Shao (00:00)David, welcome to Differentiated Understanding. Thank you so much for joining us today. I have been following a lot of your work on X and LinkedIn and you’re such a prolific writer yourself. And thank you so much for dissecting the industry, but really also breaking down the jargon on energy-related industry policies. So today I think we’re going to cover quite a broad range of topics, but really starting off from the high-level China energy planning, how it came about, and why their leaders are right now.How that plays into right now, obviously, the energy competition within the AI boom, and then the companies that are backing these developments and growth, whether it’s like PV or solar. But yes.Thank you so much. So to start off with, why don’t we start with your work and what you do at Lantau Group?David Fishman (00:44)Yeah, so I’m, I’m a principal, I’m a principal consultant at the Lantau group. We’re an energy economics consultancy. We’re focused on the commercial and economic aspects of the business of electricity or energy broadly around the region. And in China, I’m focused entirely on the business of electricity. That could mean either working with generators, right, producers of electricity, or those who invest in generation projects.It could mean working with the markets, the grid, either the physical infrastructure of the grid or the commercial or virtual infrastructure of power markets that help connect power generators to power buyers. And then the ultimate user of the electricity, the end user, which could be a large producer of physical goods or IT infrastructure like data centers. Anyone who has a lot of exposure to electricity as a buyer or seller would be somebody that we work with in our business.Grace Shao (01:40)Perfect, so you are the perfect person to go to help us understand China’s energy buildup then. We always hear about China being the biggest emitter, But then at the same time, they’re now the leaders in renewable energy generation. How did that really come about? What’s the kind of dynamic there now?David Fishman (01:56)it’s all driven by the need for electricity consumption demand, which has been rising incredibly rapidly at the same pace as the Chinese economy, right, electricity or energy consumption really closely correlated with GDP growth. So as long as your, your economy is doing more things, it starts needing more energy as well. And we went from a period where, you know, the the pace of the growth was even outpacing the ability of energy sector players to meet.that need. They were not able to build enough electricity infrastructure. They were not able to find enough primary energy sources to meet the growing demand. We’re talking about the 90s, the 2000s. And then in the last 20 years or so, we swung around more towards being in a position of relative abundance on the energy side, where it’s possible to have more energy than the economy is currently calling for. And that’s where we got to where we are now. We built up a huge energyelectricity generation base that was primarily powered by what at the time was the best option, which was coal. Coal-fired power became the backbone of the entire Chinese electricity grid. And because China is huge, whatever is the leading share of something in China becomes just massive in the world. It becomes massive overall. The renewables didn’t really come on the scene until about 2011, 2012 is when the first really strong installation capacity subsidy programs were put into place to really encourage generators to build wind and solar farms and started ramping up the scale of the industry overall. But at that point you already had tons and tons of coal fired generation that was, you know, the backbone of the entire fleet. And that wasn’t going to go away so quickly. So over the last 10 years, we can add tons and tons of wind and solar, but it doesn’t, you know, it only stems the growth of coal. It hasn’t really even started taking away from the total generation of coal fired power. So that’s how you end up having, you know, the largest coal generation sector in the world, the most emissions in the world, and the largest renewable sector in the world all at the same time.Grace Shao (04:01)But how did China become so dominant in solar and all the renewable manufacturing? was like scale, was it cheaper capital, cheaper labor, policy support? How do we understand this? Because right now we’re seeing that the US is talking about transitioning to a more, you know, green energy, but like economy.David Fishman (04:18)Yeah, well, mean, every to have an industry that is somewhat speculative, relies on new emergent or unproven technology. And in 2011, China didn’t necessarily set out to say, we’re going to go all in on wind and solar, and we’re going to count on this becoming world beating in the next 10 to 15 years. At the time, it was just, you know, we’ve built up a lot of production capacity or a decent amount of production capacity for for solar panels and wind turbines, and we’ve been exporting them.And now we would like to start installing some of them domestically. But it’s not going to be so competitive or so profitable if we do it right now. So let’s make sure there are very comfortable, good incentives in place so that anybody who builds a wind or a solar farm will be guaranteed a good rate of return on their expenditures. So you start out by saying state support. We need to offer state support to incentivize certain types of things to happen that the market wouldn’t want to build on its own.⁓ And then we need to be able to, you know, apply pressure throughout the value chain, wherever there would be somebody who’s unhappy about not getting an acceptable rate of return on their activities, generating electricity or producing solar panels or lending money to people who want to do this. Everybody needs to be incentivized to participate in this game. We’re trying to create electricity from wind and solar and it’s not very economically competitive right now. So how do we help them, right? That’s where you get Yes, your subsidies, your state support, your low interest loans, your affordable land access, things like that, all of those things, those help thing scale up, right? Those give you scale. And once you start getting that scale, you start to enjoy the economies of scale, you start to enjoy the effects of competitors on the production side entering into price wars to try to maintain market share. You get these jumps forward in innovation. I’m going to squeeze an extra 1 % out of my solar panels. going to I’m to beat the other guys, right? The scale turns into a bit of a snowball where all these other effective, enjoyable benefits, your economies of scale, you’re increasingly lower costs, you’re increasingly more attractive technology are all piling up. And all along the way, you’ve still got the state presidents hanging out in the background saying, will help, will make sure that your rate of return is acceptable. If things look like they’re getting out of hand, we’ll tweak things so that you still make enough money to keep yourself solvent. And then finally, one really important factor in all of this is the state owned enterprises themselves. The state owned enterprises are mostly involved in the capital intensive section of the industry. So that means building and operating the solar farms and the wind farms. That’s the really risky part of the entire value chain.And so they’re willing to take on a lot more risk than a private developer might be able to. They’re willing to accept lower profit margins than private capital might be willing to. And that acts as a great big lubricant for the whole system so that you can continue keeping capacity numbers high even if rates of return sometimes are a little questionable.Grace Shao (07:20)So, how should we understand this, actually? Most of the energy players, whether throughout the supply chain, are SOEs or do we have major private players as well?David Fishman (07:30)Well, so the equipment makers are almost all private companies, right? Your solar panel manufacturers, your wind turbines, your batteries, those are all private companies. They get support from the state. They have subsidies, they have land grants and things like that, but they’re all private companies. And then on the generation investor project development side, you have mostly SOEs. Private companies are certainly happy to participate when they find good opportunities with acceptable rates of return. But for everyone else where maybe the rate of return is only a 6 % return on capital, right? That’s not attractive to private equity. That’s not attractive to a fund usually, but state-owned enterprises are happy to do that kind of work because that’s their mandate. So it really is a partnership between the private sector doing certain things and the state sector doing other things.Grace Shao (08:21)That’s quite commonly actually seen in China, Like across a lot of industries. How do we understand the grid mix right now in China?David Fishman (08:28)Yeah, so right for for 2024, which is the last year we have full full data sets for I think was something like 58 % coal fired power and that’s been declining as a percentage even as it increases in total volume, right? Because other sources have been growing faster than coal has been growing but 58 % coal and then we had wind water solarWind, solar together came out to, I think it was about 32%, something like that, with wind and solar being in the 18 % range and hydro in the 14 % range. Hydro has a lot of annual variability, depending on whether it’s been a good year for rainfall or not. And then the balance, that last 10 % is made up of nuclear, which about five, five, six percent. You’ve got gas to power, which may be two or three percent, and then other, includes things like biomass, waste to power, experimental technologies like that, but still very fossil fuels dominant.Grace Shao (09:28)How does that compare to other major economies? We might look at Germany or even the US.David Fishman (09:33)Yeah, it varies a lot depending on their natural resource endowment. So the United States has been shutting down a lot of its coal capacity and instead leaning heavily into gas to power. And gas to power is now one of I think it should be the largest generation source of the United States. There’s a lot of gas in the United States. Then you take a large country like ⁓ France, right? France has nuclear, built nuclear decades ago. Nuclear is still the largest contributor to France’s electricity mix. You switch over to somewhere like Brazil. Brazil’s got lots and lots of hydropower and hydropower is the main driver of Brazil’s economy, of their electricity economy. you know, natural resource endowment really matters. Major countries that are still using coal a lot, like China, you’ve got places like China, India, ⁓ Indonesia, those are the ones that come to mind, especially in East Asia.And then throughout Europe, course, Poland is well known as being a huge, huge coal consumer. But if you’ve got access to gas, you use gas to power instead. South Korea uses a lot of gas. Currently, Japan uses a lot of gas. Gas is very common throughout Southeast Asia, usually imported gas or LNG. So it’ll be quite, quite expensive if you’re using that for electricity, which can sometimes contribute to high power costs, lot of gas in Europe as well.Grace Shao (10:51)That was really good context. I actually want to double click on the nuclear topic. It seems like there’s quite a bit of controversy with a lot of the SMRs being like the small ⁓ nuclear plants that are being reactivated right now. How do we understand, I guess, the pros and cons of nuclear? Is it really still quite dangerous for local communities or is it like potentially a big possible solution for us as we’re seeing electricity shortage?David Fishman (11:16)That’s like three different, very different questions. So I’ll try to answer them in order there. So SMRs are next generation kind of experimental ideas for nuclear that they could be maybe more cost effective or more flexible in a smaller format, something like 100 or 200 megawatts instead of a thousand megawatts. Right now there are only a couple of SMRs operating around the world for commercial civil power use.⁓ China’s got one, for example. and mostly in the United States, the conversation has revolved around restarting some previously retired or mothballed but not shut down or decommissioned, nuclear power plants. That now that there’s this energy crunch, an electricity crunch that some of that retired generation that maybe wasn’t competitive in the market landscape of whatever year it was, decommissioned in or retired in, mothballed in,Currently the market climate has changed that there’s such demand for electricity now in such a scarcity that the buyer of the electricity is willing to finance the refurbishment of the plant and willing to finance the long-term operations of the plant by becoming a buyer of electricity. So when you look at your major tech companies that have just signed an agreement to restart a nuclear power plant, that’s because the operator, the owner of the nuclear power plant has secured a very lucrative contract to sell electricity from that power plant to that data center for a long period of time. On the safety aspect, yeah, look, we’ve had a couple of notable kind of headline grabbing, world attention grabbing accidents over several decades of the nuclear industry’s operations. I always consider it to be kind of like one of those airplanes versus cars thing, right? A lot of people are afraid of flying, although it’s incredibly safe relative to driving on a statistical basis, right? But you know, an airplane crash grabs the headlines in a way that a car crash never will. And that’s, you know, a similar situation with the safety of different energy types, right? A nuclear power incident once every several decades grabs headlines, but the long term damage to human health and livelihood caused by combusting fossil fuels has been immensely larger, incredibly, incredibly larger. So everything has its own kind of trade-offs and how you evaluate it is up to you, but I do invite everyone to think of it as an airplane crash versus a car accident risk scenario.Grace Shao (13:40)That’s a very interesting way of framing it. I think like you said, a lot of times the headlines really focus on what’s big or what’s more exotic right? And that’s kind of in the case that nuclear is just not as commonly heard about, therefore everyone actually pays attention to it more. I really want to double click on what you just talked about on the big headlines of big tech buying up, or reviving plants to power data centers. Because when we were talking before this interview, you joked you said,Hey, look, I’ve been an energy guy forever. No one really wanted to talk to me that much before. For now, all of a sudden, everyone wants to talk about energy, right? We are in the midst of an AI boom. And the bottleneck, especially for the US, or the choke point right now, for a lot of these big tech companies, is securing enough energy to power data centers that are required to power their training or influence whatnot, right? So help me understand this. The AI boom is capital intensive.So is energy. What, how do we understand the relationship actually between energy and data centers and AI right now? Why don’t we start with that big question?David Fishman (14:42)Yeah, well, mean, there are a few other productive activities in the world that rely so much. On electricity as a primary input as the operation of chips in a data center. I mean, maybe the only thing that’s similar is Bitcoin, right? Mining Bitcoin because the operations you’re doing are just constant calculations that require consumption of electricity to be performed. data center operations are similar. In the physical world, the only thing I think comparable is something like ⁓ non-ferrous metals smelting, like aluminum or copper smelting is also just, we call it solidified electricity.That’s what aluminum is. So data centers are constantly performing actions and tasks that use electricity. It’s a direct relationship to do what it needs to do to be productively useful at all. needs to consume electricity. so from an economics perspective, you just say this is a demand driver of electricity, almost a direct demand driver of electricity. It’s not an aluminum smelter, it’s a data center, but we need more electricity to serve its needs so that it may serve its function, its customers, which are asking for computing power. And so from the electricity sector perspective, I just see that as a number that used to be 1 % year on year demand growth, and now it becomes 3 % year on year demand growth, or 5 % year on year demand growth.And that’s something where, if we build one or two generation assets, we will meet the anticipated demand for the future. Now it’s, need to build eight generation assets or 10 generation assets, or I can’t allow that large electricity user to actually connect to the grid and start demanding power. Cause I don’t have enough. I don’t have enough electricity to serve them as a grid operator. Maybe that’s my perspective, right? You’re not allowed to connect to my grid, you need too much electricity. And when you’re in that circumstance as a data center operator, you’re saying, well, I got to bring my own power, right? Bring my own electricity. And so that’s kind of the way they’re thinking about it now. If I want to set up my compute in an area that, you know, for whatever reason, it’s beneficial for me to be sited here, but this local grid doesn’t have enough electricity for me, I got to come up with my own solution. I could build my own assets. I could have a captive power plant, perhaps I could support ⁓ local generators to build something and sign a long-term contract with them so that they’ll build the asset, I'll build all the power. Or maybe I’m looking at something like, you know, something more creative where I say, Hey, there’s already an asset. It’s in the grid system. It’s nearby. It’s this nuclear power plant that’s been mothballed for 15 years. Let’s get that going again. I’ll buy all the electricity. Whatever it is, it’s it’s, you know, because of this very strong direct relationship.A data center operator is a wonderful customer for an owner or operator of a power plant, right? I’m in the business of making electricity, you’re in the business of using electricity, like, surely we can work something out. So that’s where you how you end up with this kind of like, we call it a PPA, a power purchase agreement. That’s the direct relationship that’s established between the generator and the consumer.Grace Shao (17:48)Who are you seeing as biggest consumers of power right now amid this AI boom that you’re seeing in China specifically.David Fishman (17:54)Well, yeah, it’s going to be your big Chinese tech companies, right? Your Tencent, your Alibaba, and ByteDance, companies like that that just have massive need for tech, you know, and then your emerging AI companies. And of course, the big tech companies have their own AI plays always. And then they’re going to be independent or third party AI companies that are looking to train or whatever it is. And they’llThey’ll need a lot of that too. So in the IT space, that’s who you’d expect. The big tech companies and your frontier model creators.Grace Shao (18:27)And what do you think, if you had to put a number on it, how big is this AI driven load as a percentage in terms of the incremental electricity demand we’ve seen, I guess, in the last three, four years and comparing it to next three to five years of production?David Fishman (18:40)So it’s interesting, it grabs all the headlines, but remember, Chinese power consumption is already growing at a stunning rate for every other reason already. So this is just one additional thing. So in the last five years, think it’s been ⁓ a modest amount of the growth can be attributed to ⁓ data center needs.⁓ Looking at something less than 15 percent I’d say maybe 10 10 15 percent of the growth can be directly attributed to data center needs And then over the next five years I’ve seen forecasts from like state-grade energy research Institute where they’re saying you know of all the different sources that are driving power consumption growth Maybe maybe 20 % will be attributed to to AI so or to data center so not not like nothing, but also not as much as you guess or expect, certainly not compared to other countries where it really is maybe 50 % of the load growth or more can be attributed in that way.Grace Shao (19:36)I see. What are actually the top drivers for China’s demand, growing electricity demand, other factors?David Fishman (19:42)Yeah, so among those, the same report that I saw that said, you 20 % can be AI. In that same report, it was another 10 to 15 % or so. it was 10 % is charging for EVs as a driver of growth. And then another 10 % was electrolysis of hydrogen. So you can produce hydrogen by running a current through water, essentially electrolysis of hydrogen, which replaces our other ways of producing.Grace Shao (19:57)Okay.David Fishman (20:09)hydrogen, which are usually quite dirty and fossil fuels intensive. And hydrogen is very useful as a as a input for for production of ammonia or methanol, many other very useful chemicals. So that’s about 40 % of the growth was attributed to those three things, these three emergent sectors. And so the other 60 % of the growth will be traditional sectors. So that’s your heavy industry, and uses tons of electricity all throughout it. That’s your services or your tertiary industry. So you’re looking at growth of AC use in shopping malls and hotels and things like that. And then the smallest portion of that is residential power use. All three of those traditional sectors are all rising in China as well. And in addition to these new three sectors that are emerging as part of the clean tech revolution.Grace Shao (20:58)That’s really interesting. I wish we hadn’t time to go into the EV talk today, but see if we do later, but let’s focus on AI first. I actually want to understand why can’t we go full renewable with these data centers right now? What’s the issue with, you know, using renewable solutions instead of traditional solutions? Cause there’s been a huge debate around that, right?David Fishman (21:16)Yeah, well, mean, and so when you have a power load, doesn’t matter what it is. If you have a power load that draws on power pretty continuously, the easiest way to meet that kind of load will be something that generates power pretty continuously. It’s not impossible to meet it otherwise. It’s just the easiest, the most straightforward way to meet that load. If it has certain types of flexibility in the way it operates, then you can also start to address it with more flexible generation sources. But overall, a ⁓ data center load is a more stable large load. And large stable loads are really good matches for our conventional generation sources. Our hydropower short, but you know, coal, gas, nuclear, great matches for the way a data center needs electricity. Now we can meet those needs with a combination of solar and wind and storage and flexible generation sources like that. But operating those in that way generally incurs a larger cost, either a direct cost in terms of operating cost of the assets or a systemic cost.And in order to have all those variable assets in there, we also needed to keep some gas generators on standby. We needed to pay some capacity payments to a battery station, something like that. Those are systemic costs that also usually end up being assessed to someone, ideally the one who caused the need. But sometimes it’s assessed to an end user who didn’t cause the need for that variable generation, but they still paid for it. So when we look at can we meet the ⁓data center needs with renewable, certainly we can. It’s just more complex and it takes more planning and maybe it incurs more system costs and sometimes the system costs aren’t really well tracked. So China does intend to start meeting a lot of its data center growth with renewable sources. We talk about where the new data centers are going in, the parts of the country that they’re going in and there’s this program called the National Hub Nodes.Which is part of the East Demand West Compute program. And so under this scenario, any new data centers added into those national hub zones need to be consuming at least, I think it’s 80 % of their electricity needs to be renewable. It’s a pretty high percentage. So if it weren’t possible, there wouldn’t be such a requirement, but it is trickier, it’s more complex. And so that’s how China is proposing to drive most of its data center build out now is with these renewable energy sources. It’s a way to avoid this growth of a new load sector resulting in increased fossil fuels consumption.Grace Shao (23:57)I see. I’ve heard about like, you know, the challenges with battery solutions right now for the intermittent kind of nature of renewables. Do we have actually strong enough battery solutions right now to solve that issue?David Fishman (24:10)I mean, batteries, the question is, can you you scale it cost effectively? Right? Can you can you get your battery storage to be able to cover what two hours, four hours, eight hours? How long of a backup solution do you need there? How what type of gaps are you expecting to need to meet in the context of all the other power that you’ve contracted for or that you have available to you? If if you’re in a region where there’s tons and tons of solar power, it’s sunny all the time. The sun is up most days and it’s producing well most days. Okay, well, you’ve got a solution for part of the day. And then it’s like windy in the evening. So you’ve got a solution for the evening. But how do we cover that evening peak period when it’s not particularly sunny or windy? Okay, batteries, sure. How long are the batteries good for? Can we be sure that the batteries will be charged and ready with no loss of load incident every day? Or at least at a very, very slow failure rate, something like that. When you start doing some type of probabilistic assessment, how many batteries are enough so that we can ensure a failure happens once every 10,000 operating years or something like that, whatever your threshold is, you start to realize, woo, it’s a lot. I need a lot of backup. I need a lot of storage. And so maybe that’s where you run into some of the problems where the project economics might not work. They might not make sense anymore. Ideally, I’m describing an extreme scenario, but ideally you can find a solution that involves not using massive, massive amounts of batteries that you barely use most of the time.Grace Shao (25:38)I see. It just sounds like there’s a lot more operational risk, right? I want to understand East data, West compute. From your perspective, like how should we understand that, I guess, from the energy framework? And can you help us understand the geographical importance of having the energy suppliers on Western regions of China?David Fishman (25:57)Yeah, it’s absolutely driven by geography. maybe you’ve seen the famous line before, you can draw a line through the center of China and everything on the west side of the line is 7 % of the population and everything on the east side of the line is 93 % of the population. It’s roughly half and half of the country. That’s just where all the Chinese people, all the load, all the industry is, is along the coasts in the east and the south of the country.And so your need for electricity, your need for, your demand for any energy use of any kind is mostly going to be out there. But the good energy resources in China are in the West. They’re in the North, the Northeast, and the Northwest, and the West. So you’ve got to find a way to get them to each other. You want, you know, the wind blows all day, the sun shines all day, and the mountains are just heaped with coal in Western China.So we got to connect them to Eastern China. In the case of data centers, you’ve got two options, right? Either we’re going to bring power lines from the West to the East, or we’re going to bring fiber optics from the East to the West, right? So either I bring the electricity to where the load is, or I send the compute to where the electricity is. And they are doing a mixture of both. So this is where you end up with a bifurcation of the types of computing needs that you have.The short term, very rapid response compute needs to stay in the east, close to where the demand is. And so for those, we bring the electricity to the compute. And then for like a longer lead time, you know, maybe we’re training a model or something like this, we can send it out to the west, send it closer to where the cheap electricity is. So this is a consideration that makes sense for China, because China is huge and it’s got very different ⁓ dispersion of its resources. A smaller country could do it on a smaller scale, of course. I know in the UK they joke about London data computed in Scotland or something like that because they’ve got the offshore wind up there in Scotland. So it’s a similar concept, just in China’s case it has to happen on a continental scale.Grace Shao (27:56)I’m also interested in actually how the relationship between the provinces are. How do they work together? Do they have different mandates for each of the provinces or is it just a very top-down kind of mandate from the federal level?David Fishman (28:10)Yeah, so in China, is a tiered, a hierarchical relationship between kind of national government, provincial governments, and municipal governments. Generally, you’d expect planning at the more macro level to come out of Beijing for the whole country, and execution is going to be left to provinces and municipalities. So when they say you need to enable cross-regional power transmission, okay.Like that’s that’s something that comes out of Beijing, but now it’s up for the individual provinces to work out how to do that. Eventually, they’ll execute their trades through the state grid power exchange, which is up in Beijing. But all the negotiations and all the interactions have to happen at the provincial level or even lower at municipal levels. That’s that’s broadly true for for almost all Chinese policy, not just the power sector and not just data centers, but in general, broad strokes laid out at the top and then executional happening at provincial or low.Grace Shao (29:05)That’s interesting to hear. I kind of want to shift gear and kind of double click on the private sector right now. You you just talked about along the supply chain, lot of the coin makers are actually private companies. We’ve got the PV makers, the battery makers. think, you know, people in the West will probably know of CATL, Longi, Trina. There were some also even international ambitions of these companies kind of set up manufacturing hubs in the US or even across Southeast Asia. Can you kind of help us understand their global ambitions and where they’re at now.David Fishman (29:34)Yeah, mean, so the initially the Chinese market was very initially, they were built in China to be sold internationally. And then the Chinese market got so much larger than everywhere else that many of these equipment makers had plenty to do just selling to the Chinese market. But because of the very, very stiff domestic competition, there’s always that interest of like, well, could we could we diversify a little bit out of this incredibly competitive environment by selling?broad as well. And so initially they could sell abroad from China again, as some of those trade barriers started going up that really incentivized the expansion into Southeast Asia, into Vietnam, for example, that you could have Trina solar panels produced in Vietnam and exported to markets that were maybe closed, or tariffed for for Chinese exports. And then we saw, I mean, with the United States and its recent tariff policies, that tariffs were placed on the Southeast Asian countries as well. And then specifically if there was evidence of trans shipment of Chinese panels heading to Vietnam and then going to the United States that they would be taxed or tariffed even more. So a lot of that has, you know, eventually ended up coming down to just one place, right? If you’re a Chinese solar producer, and you want to sell to the United States, you’d better be ready to produce in the United States. That seems like the way I’ve talked to a couple of them, and that’s the way they’re thinking about it. If they had production already, they say, okay, we’re gonna keep producing for the American market from our production facilities there. If they’re thinking about starting a new, opening a new factory, well, it’s been a scary environment for that kind of thinking recently, right? Of course, there was the recent⁓ raid on a South Korean under construction battery facility in the United States, right? There’s a lot of concern about, you know, that kind of capex heavy investment for a market that just maybe isn’t geopolitically that friendly, and it would come back to bite you in the end. There was another case, I think, where a Chinese solar panel producer built a facility in the United States, and then it wasn’t going to be eligible for the IRA tax credits and they were forced to sell it to an American company, which is now producing solar panels from that facility. So if you’re a Chinese solar panel manufacturer or battery manufacturer and you’re looking at this kind of geopolitical climate, certainly you’re willing to consider building an overseas factory to get around tariffs, whether it’s in Southeast Asia or even in your target market. But some markets seem like they’re a little bit more geopolitically favorable than others right now.Grace Shao (32:08)It’s obviously quite challenging for a lot of these companies trying to sell to the US right now. Just like you said, given the kind of backdrop. Okay, I want to ask about Southeast Asia. We know that a lot of these companies are building out, you know, manufacturing hubs, even data centers in Southeast Asia. A lot of the big tech companies, Chinese big tech companies are now said to be the biggest hyperscalers across Southeast Asia as well. Like they’re buying up all the data centers.But there are some concerns around infrastructure challenges in Southeast Asia, right? So there have been complaints about, you know, the utility prices rising. There’s issues in Johor, Malaysia, where frankly, local infrastructure such as water, land, roads, it’s not really there yet. How can you, how do you help us make sense of all of this? Like, who are the buyers? Who are the investors? Who are the users? Is this really helpful right now? Are they actually helping the local economy? And are they actually, are these companies getting what they need in those regions?David Fishman (33:00)Yeah, so anywhere in the world, including China or even very, very transparent power markets ⁓ like parts of the US or in Europe, it has proven to be remarkably difficult.to attribute changes in the price of electricity to any one thing. Data centers have become a ready and available target. And I think it’s surely true in some cases that it is the thing that’s driving higher power prices for that region or for that node. But in other cases, I think it would be more difficult to make that assessment. And it’s something that is speculated in the media and then kind of becomes true by default without really having a good investigation into it. I haven’t looked at Malaysia or Vietnam enough to comment whether I could attribute the data center, no matter who owns it, the data center demand to rising power costs. It does make sense usually in an economic sense, right? Your demand load has increased, your supply hasn’t changed, and so we end up at a higher point on this cost curve for more periods of, you know, throughout the day.So that’s, mean, it’s something that is theoretically a problem. If it gets to the point where the data center load starts to really affect the competitiveness of electricity for other parts of the economy, either the industrial, other industrial segments are saying our energy inputs are too expensive now, or you’ve got residential power costs skyrocketing, you’ve got a developing economy, and you’ve got, you know, residential people can’t necessarily afford the costs of higher energy costs, then you’d expect the national policymakers, regulators to step in and say, if you’re going to be allowed to operate in this way, we need you to bring your own electricity, for example, ⁓ or you need to be willing to cover the increased costs of other sectors of the economy through some type of ⁓ payback scheme or kickback scheme, right? That kind of thing is a reasonable thing to ask of a data center operator when they’re coming in and creating those types of problems for a generation overall. But I mean, remember, they are also creating their own economic benefit. They’re not just leeches on the system, which I think many would argue something like cryptocurrency mining is that you’re just sucking up tons and tons of electricity, and you’re only creating value for a smaller group of people who believe in the value of the cryptocurrency. Data centers, on the other hand, are performing tasks that are considered to be productive and valuable tasks for their clients, for their customers.Maybe you could argue you think that using AI video generation is not a productive and useful task, but hey, like somebody thinks it’s valuable and useful. Enough people think it’s valuable and useful. So that’s what you weigh the two against each other, right? You’ve created a burden on the grid system, a burden that is having knock-on effects on other parts of the economy. You need to account for yourself. You need to take responsibility for the burden that you’ve created on the system.But also, you you’re not a leech. You’re not a parasite on the system. You are in fact doing a useful thing and it’s the responsibility of energy planners and economic planners to try to make energy available to you. So it’s a delicate balance. And it’s, I wouldn’t, it’s not fair, I think, to attribute blame to anyone unless they’re intentionally trying to get around policy or screw over other people.Grace Shao (36:28)Yeah, I have a question that’s a bit more anecdotal, but when I drive through Europe or China, right, you see like solar panels on private citizens houses across the countries, right? Like, you know, in Spain, Italy and Germany, you really see across China, like Hebei in Songsu, you really see people embracing it as a society. Can I do you think there’s a cultural difference in terms of kind of embracing renewable in the US versus China or Europe or what kind of incentivizes them.David Fishman (37:02)There’s a cultural difference, but in the case you cited there, there’s a very strong commercial difference as well. In Europe, residential solar is primarily motivated by the decision making of the people that own the house, that own the rooftop where the residential solar goes in. In China, rooftop solar is driven by developers who don’t see the they see the rooftops as real estate.So developers go in, they knock on doors and say, I see you have some nice unoccupied rooftop space up there. I will pay you to lease your rooftop. I would like to put solar panels on it and sell electricity back to the grid. So the Chinese business model is very unique in that case. I would say the United States and Europe are more similar in the ways that the mechanism for motivating, motivating residential installations. And then China is off doing its own thing because it has this incredibly effective, but unique scenario where developers are the ones promoting solar panel development to residential users who weren’t very aware of, you know, solar energy at all. And then also, they don’t even have much electricity consumption, frankly. So they’re not super motivated to go install solar panels. ⁓ residential electricity costs in China are very cheap. So it’s not like they’re trying to offset a high power bill.The whole motivation for the residential sector in China is ⁓ collecting a rent on their rooftop space so that some developer can generate electricity and then sell it to the grid.Grace Shao (38:31)I see. I totally thought it had something to do with just kind of this overall societal embrace of new technology and this 2030 renewable energy goal. Can you actually, I guess the last question I have for you, can you help us understand what is this grand 2030 goal of China to really convert, you know, transition majority of the energy consumption to renewable? How do the EV sector, the battery sector, and various sectors related to energy really play a role in this?David Fishman (38:58)So the 2030 goal is the peaking goal, the emissions goal. So they want to peak carbon emissions by 2030. And then they say carbon neutrality neutralize the emissions by 2060. So we got to reach a peak, then we got to draw down and neutralize. So by 2030, I mean, maybe we’ll be down to a 50 to 55 % share of coal in the power sector. I don’t know if it will slip below 50%. By that point, it’ll be getting close to it.And then you’ve got all your other non power sources of emissions, right? Transportation, building, heating, and industry. That’s the major one, industrial inputs. So by 2030, the goal was to just peak the emissions across all emissions generating sectors. Now it looks cautiously like we might have reached an early peak this year. I say cautiously and I hedge my phrasing here because, well, there’s lots of different segments that create emissions.Power is maybe gonna stay flat, maybe gonna be, you know, peaked by 2025, 2026, but we’ll see what happens to new capacity for wind and solar next year. If wind and solar and hydro and nuclear can’t keep up with the pace of consumption growth, then the only way to meet the additional consumption growth is by using more coal, right? So that’s the power sector. Fingers crossed that they managed to figure it out. Transportation fuels maybe is going to peak in this year or next year. We’re talking about the EVs coming in such massive numbers that they’re replacing petroleum ICE vehicles.Grace Shao (40:28)What’s the percentage actually? What’s the percentage of EV cars on the roads right now, do think?David Fishman (40:33)It’s you know, it varies widely by province in Shanghai. It’s over 50 % now for sure. Major cities is 50%. But you know, I was recently in, like Shanxi, a northern, you know, Chinese province, and it was probably less than 10%. ⁓ So it’s I don’t know what the number is nationally. It’s it varies so incredibly by city, I think it’s or by province, I think it’s best to look at that on a provincial level. But the idea being that newGrace Shao (40:38)Wow.David Fishman (40:57)New EVs should have already started outpacing the sales of ICs and that should continue so that we would be able to, know, more vehicles are going on the road but the EVs are being added more rapidly and that petroleum consumption for passenger transportation should be peaking, you know, within the next two years or so. And you’ve still got aviation and maritime fuels and long-distance trucking and all the other things. Excuse me.And then finally, you’ve got the industrial sector and there’s different aspects. Some are growing, some are rising, know, steel is maybe flat, cement is dropping for the next 15 or 20 years, but hey, petrochemicals and coal chemicals are rising for the next 15 or 20 years. So you’ve got all these different, you know, sources of emissions to consider. That’s what’s all going into that 2030 goal. So if they manage to peak by 2030, like wonderful, lovely, enough of those sources have peaked and started drawing down that we can make up for the ones that have not peaked and that are not drawing down. And then we’ve got another 25, 30 years beyond that to figure out how to get to neutrality after peaking in 2030. If they happen to incidentally peak this year or next year, I don’t think they’re gonna claim it. I don’t think they’re gonna crow about it. I think gonna stick to the 2030 emissions peak, give a little bit of time just in case a little bit of buffer.Grace Shao (42:08)Really?David Fishman (42:13)But it seems very likely that we’re flirting with the peak a little bit early versus their own targets.Grace Shao (42:20)What’s the practical impact on society and the economy then?David Fishman (42:24)Well, the ideal goal is to have as little impact as possible, right? That business as usual growth as usual continues, economic abundance of energy is compatible with energy transition. That should be the goal that the only outcome in fact is that the business of the energy transition is good business, that it contributes in fact to GDP.⁓ installing solar farms and building wind turbines and upgrading industrial equipment is all like good economically productive activity. And so that should be the only impact. That it’s a good impact and that otherwise everybody just goes on driving cars, they just happen to be EVs now. They go on using electricity, it just happens to be increasingly clean electricity. They go on using steel, it just happens to be green steel. That’s the best case scenario.Grace Shao (43:16)That’s a good world to be looking forward to. Okay, David, thank you so much for your time today. What is one differentiative you you may have? It could be about your industry or it could be about something else broader in life.David Fishman (43:28)One differentiating view that I have.Yeah, well, I mean, I’ll talk about my other great passion, which is kind of economic development in general. I got into electricity because I cared about development. And I stay in electricity because I continue to see it as important for development. And I am pretty agnostic when it comes to the different isms of the world, the different political science factions that say it’s this ism or that ism.jWhat works is the best ism, whatever it is, I don’t care what label it has. And my observation right now about where we are as a civilization is that we are right on the precipice of moving from the fossil fuels phase of humanity to the electricity phase of humanity that electricity will be an interim period, because eventually we’ll probably figure out fusion power and harnessing plasma, and that will take us to the stars. But in the interim phase, we’ll be in the electricity age. We went from biomass to mechanical energy to fossil fuels to electricity generated from the natural elements. And that this electricity phase should be the phase where energy stops being additive.And starts being substitutive instead. All the previous energy phases that we went to, we had to use more energy to upgrade. But to upgrade to electricity should allow us to start using less primary energy. That using wind and solar and hydro and nuclear are less intensive for primary fuels, which means we’ve freed up space for developing nations in the world to start to enjoy a fraction of the energy abundance, the civilizational abundance that so much of the rest of the world is already enjoying. They’re already there, right? As long as we stayed locked in the fossil fuels era, they’re not allowed to enjoy that abundance. They’re locked out of it because we have no more carbon budget to spend. We need to work on this, you know, civilizational existential challenge of transitioning our energy. And for the moment, there’s no more carbon budget for the developing countries of the world. Getting into the electricity era, the electricity civilization era, frees up that carbon budget, enables them to start pursuing the same type of abundance that everyone else is already enjoying in developed nations. That’s the great thing that motivates me throughout all of this, that the electricity civilization era is one where regardless of what ISM introduced it or ushered it into the world, is one that is going to be incredibly beneficial ⁓ for humanity. And at this moment, I think I’m working in the country that is most likely to usher in that era. And here’s the controversial portion, right? I think China is the one to bring the world right now into the electricity civilization era. And that will be such an incredible boon for humanity in general.Grace Shao (46:27)That’s a really, really interesting way of framing it. I really appreciate that. Thank you so much for your thoughtful answers and just your time and your insights today. I really appreciate your time and just your generosity in sharing with us.David Fishman (46:38)Thank you for having me. It was a great pleasure.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• Is this the Cursor of China? Alibaba's Qoder team on agentic coding, Qwen, and international ambitions 10.11.2025 50min

  “So our philosophy here is to integrate the globally optimal models and give users the best results.” — Hang Yu, Head of Product at Qoder, AlibabaThis is the first episode in a series of founder and builder dispatches, featuring interviews with the people creating the future. If you are a founder, builder, or investor in this space and would like to share your story, please reach out.Today, I am joined by two guests from the Qoder team at Alibaba: Hang Yu, Head of Product, and Christian Hu, Head of Global Marketing and Operations. The Qoder team launched just over two months ago, joining the likes of Cursor, Warp, and Copilot to make coding more agentic, so today we get to learn from them directly about their unique positioning being part of the Alibaba ecosystem.Hang discusses the thinking behind designing Qoder, how it differentiates itself from peers currently available on the market, the future of agentic work, his fears and excitement about the pursuit of AGI, and finally, challenges the notion that the future of AI may not be based on Transformers.Christian walks us through Qoder’s business positioning, global ambitions, how it fits into the Alibaba ecosystem, and the reasons for routing between models, beyond just Qwen.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Topics we covered:Product* Introduction to Qoder and AI Coding Agents* The Transition from Copilot to Agentic AI* The Future of Developer Productivity with AI* Addressing Developer Bottlenecks* Multi-Model Strategy and (Qwen) Integration* Differentiated Views on AGI and AI’s FutureBusiness* Understanding Qoder’s Positioning in the Market* The Competitive Landscape of Coding Tools* Qoder’s Role in Alibaba’s AI Strategy* International Ambitions and ChallengesTranscript (AI-generated)A. Hang Yu, Head of Product at QoderGrace ShaoAnd again, I just want to say thank you so much for joining me today, Hang and Christian. So today, the first half of our conversation will really focus on the product design of Qoder and the transition that we’re seeing from Copilot to Agentic. And then we’ll move into the second half of the conversation, which will really focus on the business strategy of Qoder, international expansion goals, objectives, and then how it really fits into the bigger Alibaba AI plan and the bigger AI playbook. So with that, I just want to bring in Hang. Hang, it’s lovely to meet you and thank you so much for joining us today. Let’s start with the very, very basics. What is Qoder in plain language and what does it actually do for developers day to day?Hang Hey, great. Thanks for having me. So yeah, so in one sentence, Qoder is an AI coding assistant that helps developers maintain and improve existing software system, not just build some new stuff from scratch. And I think that distinction is actually really important. So when we look at what developers actually do every day, we found that 95 % of professional developers spend their time maintaining what we call real software.So commercially valuable, long-lived systems that are not building new projects every day from zero. So that’s why we design code around that reality. We are optimizing for the messy, important work of understanding existing code, making targeted improvements, and maintaining production systems.Grace Shao That’s really interesting. you know, there’s for me again, not a technical person. The hype we hear about is all this vibe coding. And then there’s even these marketing phrases calling it like ⁓ AI coding, build your app from a single prompt. That’s not what’s really happening, right? How does it actually really work?Hang Yeah, exactly. actually, that’s the flashy use case everyone talks about, like building one app from a single prompt. But if you’re working at a real company maintaining a five-year-old code base with 500,000 lines of code and 20 different developers who have touched it over time, you need help understanding what’s actually here. So you need help making careful modifications without breaking things.That’s where code delivers value. So in terms of how we think about the AI developer relationship, we see it evolving through three stages. So the first is assistive programming. In this stage, AI helps the developer while human leads. So like code completion, fixing syntax errors. And the second stage is collaborative programming, like co-pilot. Like the human and AI works together like pair programming.And the final stage is autonomous programming. In this stage, AI takes on complete tasks independently. So the developer can delegate work, and the AI runs in the background and comes back with results. That’s what our Qoder Quest mode does.Grace Shao So the really unique bit of Qoder Quest is the autonomous piece, right? And I really want to dig into that a bit more, a bit later. So right now I’m curious. So when you say the user experience is intuitive even for non-technical users, what design choices really led to that? Because honestly, a lot of developer tools are pretty intimidating, especially for people like myself who’ve never done any coding. But what I’m hearing on the street is people are going out of their way, even as non-technical people, building their own apps now with the help of AI. How are they able to do that?Hang Yeah, great question. So we have this philosophy. So don’t make users think about things they shouldn’t have to think about. So for example, if you look at some products, they have like 40 different AI models in a dropdown menu. Honestly, that creates a lot of cognitive load. So developers end up becoming model select or instead of focusing on building their own product. So our philosophy here is integrate the globally optimal models and give users the best results. So we will auto select the right model based on the task. So we believe model selection will be better than human selection. And the same thing with context management. The users shouldn’t have to manually figure out, OK, which files to include, what tokens to optimize. Our context engineering handles that automatically. So the goal here is to remove the cognitive overhead. And let developers focus on what they are trying to build, not on configuring the AI tool.Grace Shao I have a really dumb question, but is there a latency then between me prompting like, can you help me build this versus the machine telling me which model is optimized? Is there like a latency? It’s automatic.Hang No, no, there would be no latency. Yeah, it’s all automatically. The user will not feel it.Grace Shaothat’s amazing. Okay, let’s talk about kind of the hype right now, the copilot versus the agentic transition right now in the coding space. Many tools are being called assistants or some are called copilots, You know, the cursors of the world. And cursors essentially been leading this. So where does coders autonomous capability really truly defer or is unique or different? And what’s really the big breakthrough we’ve seen here?Hang Yeah, this is the defining question, So Cursor has ⁓ perfected the co-pilot approach, real-time help while you’re coding. Their tab completion, tab, tab, tab, is industry-leading after two years of their custom model training. But I’ll say this. So Cursor’s tab completion compatibility is catching up fast. We have made significant progress in recent months and rapidly closing the gap.But here’s where we see the real future, the autonomous programming. So you delegate a complete task, implement this feature, fixing this bug, et cetera, whatever. And then the AI works on it in the background. You don’t have to watch every line of code being written. You just come back and review the result. now sophisticated autonomous coding and production skill is still involved.Hang We think it’s about two to three years out, maybe 2027 or 2028, before it’s really mature. But that creates an opportunity window.Grace Shao That’s actually quite soon. So what does that really mean practically?Hang Yeah,so it means With autonomous coding you can actually Delicate your work and make and the the AI agentic will can work it background in the cloud so Like you can close your laptop and the AI keeps workinggo to a meeting, go home, do whatever you want, the coder is still running. So in other words, you can spin up 10 parallel sessions working on different tasks, and it doesn’t slow down your own machine. As one colleague said, he said, I’m managing 10 agents. My productivity went up 10 times, and it didn’t mess up my work-life balance. So yeah.Hang So this cloud execution model is pretty similar to what cursor recently launched with their cloud agents feature. So both approaches let you handle your tasks to agent running remotely. The key advantage here is you are not tied to your machine. You can dedicate the work and then close your laptop and then come back to complete the result. And then when theHang Yeah, and then when the agent finishes, we just need to review the results.Grace Shao So that’s actually my question. Like when you talk about reviewing the results, is it very obvious to kind of find the issues only in the result or do you have to go back to the process? Like how do you audit the whole process actually?Hang Yeah, so the agents will present to you summary of, this is what I’ve done. This is the result of the question, or this is the feature. You can see it through the browser, or the agents will submit a PR to your GitHub to your routing workflow. And you can easily check, OK, is it done good, or it needs to be modified again.Grace Shao So does that mean that junior developers in a way will be replaced then? Because essentially you only really need people who understand a higher level of the code and the execution can be actually outsourced so easily.Hang Hmm. Good question. So simple, agentic task works well today. You can already dedicate things like write a unit task for this function or add log into this module and come back get some good results.So sophisticated, agentic coding at production scale, where AI can take a high level of business requirements and then autonomously design, implement, test, and deploy a complex feature across multiple systems, that’s still two to three years far away. So what this means for software development is a fundamental shift in how developers spend their time. ⁓Hang So right now, developers spend maybe, I think, 30 % of the time on creative problem solving and 30 % of the time on mechanical work, like writing border plates, fixing syntax errors, ⁓ updating documentation, writing tests, et cetera. But as agentic tools mature, that flips. So AI handles the mechanical work. Developers ⁓ can focus on what AI can do yet, like understanding what actually needs to be built.Hang making architectural decisions that require domain knowledge and validating that the code actually does what it should be.Grace Shao Mm-hmm. So in the end, it’s still like, there are still aspects of humanity that can’t be replaced by machines yet. Let’s get a bit more practical. I want to understand how precisely Qoder can actually turn a prompt or product spec into working code? What types of tasks or repos does it handle best today, right now?Hang So Qoder handles best what we call a bounded, well-defined task today. For example, implement authentication for our user login system or add relimits to our API endpoints or generate a dashboard for monitoring system health metrics. So these are tasks where requirements can be clearly specified, the scope is contained, and the success criteria are measurable.Grace Shao Mm-hmm.Hang But what’s harder today are tasks that require deep domain knowledge or ambiguous requirements like improved user engagement, that’s too vague, or like a required entire authentication system, that’s too large, that’s too large and risky for autonomous execution.Grace Shao So it seems like a lot of the tasks it can do is still pretty much something that’s very easily verifiable. It’s a bit more like a black and white kind of answer kind of task, but not so much things with nuance, right? I want to understand better. So, thinking about the developer workflow, and correct me if I’m wrong, there’s the planning stage, the code writing stage, the running tests, the debugging, version control, and deployment, right? Where does Qoder natively sit the strongest, like in the most helpful, and where do you hand off to other tools?Hang Great question. Let me break it down. So the first is plan and design stage, So Qoder is strong here through the spec generation. We can help the developer translate business requirements into technical specs. And the second stage is writing code. This is our core strings. The Qoder can write code across multiple files, handle complex logic, and generate boilerplate. And for the run and test stage,Qoder can generate unit tests, integration tests, and run them either locally or in cloud sandbox. That’s building feature. And for debugging stage, Qoder can diagnose its errors from test results and fix them autonomously. But for production debugging with live user data, you still need the traditional tools. That needs to be taken care of. And as for version control or call out, we integrate with Git, GitHub, and GitLab. So a coder can create branches, commit changes, and create pull requests. But the actual code review and collaboration discussion happens in your existing tools, like GitHub, GitLab, whatever you use. But yeah, in the final stage, the deployment, we hand off here. So deployment involves your CI-CD pipelines, like infrastructure, monitoring system. So Qoder creates the code and the tests, but you own your deployment process.Grace Shao I see. I kind of want to take a step back and understand another big kind of hovering question a lot of people have right now, which is if we’re going to have agentic tools really implemented the work process, how do we understand developer productivity for the future? Will developer productivity still be measured the same way that it’s currently being measured?Hang Yeah, I think the fundamental shift here is this. So AI changes what developers spend their time on, not just how fast they work. So again, right now, developers spend maybe 30 % of the time on mechanical work, like writing boilerplate, debugging syntax errors, searching documentation, setting up environments. And only 30 % of the time goes to creative high-value stuff, like understanding what needs to be built. Like understanding what needs to be built, making architectural decisions, and validating that solution actually solves the problem. But with AI, agent AI flips that ratio. So AI handles the mechanical work. Developers focus on the part that actually requires human judgment. So when we measure productivity, we are not counting lines of code or tickets closed.So we are looking at, can developers spend more time on higher value work? Can they ⁓ ship features faster without burning out? So what we are seeing is promising. 99 % of our paid users actively use agent model. Nearly 99 % of our paid users actively use agent mode. So it’s become core to their workflow.80% renewal rate tells us that people see the real value here. And enterprise reports two to three times improvements in the deployment frequency. So the real metric here is simpler. Developers tell us that they are less frustrated. They are not stuck debugging environments or writing repetitive code. They are solving interesting problems. And that’s the productivity gain that matters.Grace ShaoI see, I see. That’s really interesting because I think as a writer, when people really initially rejected AI, the idea was also that there was a lot of mistakes, AI slop, hallucination, whatnot. But then what people started realizing is that you can actually use it as a productivity tool. And like to your point, it doesn’t change how you think as a writer building out the framework, using your critical thinking and really still rely on your own creativity. But you know, you what you’re outsourcing is actually just like the execution that was a lot of the grunt work frankly. Right. Interesting, interesting.Hang Yeah, yeah, exactly. Yeah, you free up your hands and you, yeah, yeah, you free up your hands and focus on your head.Grace Shao Okay, so I wanted to ask another question on productivity and bottlenecks. So what is, I guess, one of the top bottlenecks or what are a few that you see currently in this space for developer right now? And I guess can we actually separate the separately answer this question, the first half being what are professional developers bottlenecks and what are kind of the new age, like casual developers bottlenecks and how are you helping these two differently?Hang Good one. Let me hit the main ones. So usually the bottlenecks, like the first big bottleneck is about the environment setup, So this is a huge time thing. The coders, and then the coder can help you set up the environment automatically, whether you are running it locally or in the cloud, pre-configured environments, automatic dependency resolution.So you don’t have to spend two hours debugging your Python version complex or just trying to build your dev environments. And then for the flaky tests, coder can also detect flaky tests by running them multiple times and identifying inconsistency. So you can also suggest fix based on the failure patterns and the test outcomes. And then for...For another bottleneck here is the legacy code. This is where the Ripple Wiki comes in. So remember how I said the documentation is always out of date for developers? Ripple Wiki uses... So Ripple Wiki delivers value here. So Ripple Wiki features here can use AI to generate up-to-date documentation from the code itself, plug the Git history. So it’s not just...Here’s what the function does, documentation. But here’s a business logic and why it was architecture this way and what changed and why documentation. So the documentation itself stays fresh. When code changes, the documentation regenerates automatically. So we have measured about five times speed improvements, so from 60 minutes down to 12 minutes for team documentation.And the last but not least bottleneck here is the context limit. So this is a technical challenge. Models have tokens limits, So we have our context engineering figures out what’s actually relevant to the task. So we don’t just dump the entire code base, which will crush your token limits. We intelligently select what the task needs and what the AI needs. and we ⁓ gain a better solution from the ⁓ context we select.Grace Shao That’s really interesting. So I think I want to talk about the models and orchestration here. So Qoder is a multi-model platform, right? You guys use Qwen, your in-house built models, but you also use other models like you mentioned earlier, use whatever is like kind of state of art model for the right task, right? How do you route between the models, thinking about latency, cost, evaluations, languages? Do you usually like have a preference for Qwen or the third party models when you’re assigning these models to tasks?Hang OK, so what we are trying to do here is integrate the globally optimal models. ⁓ So what we are trying to do here is to integrate the globally optimal models and give users the best results. We are not limited to just Alibaba’s model. So if one frontier model is better for a specific task, we use that. So like if GPT Excel somewher, we use GPT. If Qwen is the right fit, we use Qwen. So now, why does this matter? So first, it best has time. I’ll be direct. So Qwen model isn’t the best model in the world for every coding task yet, but it’s improving fast. So by using the best global models today, can serve users well while our own models can catch up. And if Qwen becomes the best model globally in your year, is our goal, then naturally we’ll use it more. And secondly, yes, sir. Yeah.Grace Shao But actually, just want to jump in on that. But actually, Curser is even using Qwen isn’t that fascinating that we just found out recently.Hang Yeah, so the cursor’s composter is... They didn’t officially admit it, it’s... So the community thinks they’re fine-tuned and post-trained based on Qwen or some Chinese open source model. Yeah. So first is the best time. And second, the cost optimization without compromising quality is our second goal. So not every task needs the most expensive frontier model. So simple completion is a smaller, faster model. But for a complex reasoning, we use the frontier model. And last but not least is about the reliability. So you don’t want your entire product to stop working because one API provider has an outage. Here we saw OpenAI has multi-day outage in 2023. we want to our reliability to our customer. Yeah, that’s why we use multi-modal strategy. Yeah.Grace Shao I see, I see. I was actually going to ask you on that. What’s the thinking behind designing the product using a multi-model design versus only proprietary Qwen? I guess you already answered that partially. ⁓ I also wanted to kind of double click on the fact that Alibaba is throwing 53 billion US dollars into AI infrastructure right now. know, isn’t it, is there not some pressure coming from up top to really hone in on Qwen or, know, like, I guess the question’s more about - Is Qoder really developing based on what’s best for the user and use whichever model that’s best for them? Or is it more focused on being part of an ecosystem, another tool out of the Alibaba set of tools that they provide? Does that make sense?Hang Yeah, great question. So it comes down to serving users best today while building towards to the future. look, Qwen models costs us like 1 fifth or 1 sixth of what Frontier model API costs. So that’s an 80 % cost reduction, right? That’s a structural advantage. But Qwen isn’t as good as the top Frontier models on context reasoning tasks.That’s why we use a multi-modal setup at the balance cost, capability, and reliability. But the key in size is that this isn’t a permanent state. Qwen improves, as coding capability, sorry, as Qwen.So the key inside is that this isn’t a permanent state. As Qwen improves, as its coding capabilities get stronger, the balance will shift. So we are not philosophically opposed to vertical integration. And we are actually pragmatically choosing what serves users best today while building towards our ownership tomorrow. So for enterprise customers in China, components actually require a domestic model anyway. But for international customers, they often trust well-known frontier models or GPT models. So multi-models board let us serve both.Grace Shao I think that’s a really interesting business strategy because end of the day, to your point, it’s a very pragmatic approach to actually serve your customers best and to serve your customerbest, you actually get more business. That’s just how it actually the virtuous cycle works, right? Instead of building up these guardrails and the paywalls. I want to kind of pivot to strategy soon, but before we do that, I really want to ask you a few other questions just on more, a more broad general question. One is, you’ve been working in this space for a long time. There’s a lot of hype around Agentic tools, not just a Agentic tools in coding, right? How do you view a Agentic AI in the next 12 to 18 months? And how do you think it will actually affect what even the general mass think of AI use AI?Hang I think agents will swallow the whole market. ⁓ People will use more more agents in their daily workflow. So when you’re trying to use a chatbot, you need to copy paste a lot of data, your contacts, your own data, your domain knowledge.To chatbot and then generates some specific task results. But by using agent, agentic AI, you don’t need to do that. The agentic AI will live in your workflow. It knows and it contains your domain knowledge. And it knows your context by nature. So this means agentic AI knows ⁓ what you’re trying to do, what you did.What you are trying to do and what you want. That’s actually a big difference between the big difference between the LLM and the agent AI. So I think, yeah, yeah.Grace Shao I see. I think that that leads me to the next question, is like, so for teams who are customizing tools or agents, like you just said, they would have domain knowledge. They already know your work processing really well. And how does it actually work? What’s the plug-in API story with Qoder in this sense again?Hang Yeah, so this is about the domain knowledge question and the accessibility work. Let me break down how the domain knowledge and the accessibility works. So first, we have MCP supports. We support the modal context protocol, which is becoming the industrial standard for connecting AI agents to external tools and data source. So this means the coder can ⁓integrated with thousands of tools in the ecosystem, database, API, version control, and project management through a standardized interface. And then we also support subagents and skills. So teams can create their own specialized subagents. Think of this as task-specific AI workers with their own contacts and capabilities. So ⁓ you want a code review subagents turned into aSo you want a code review, sub-agents turn to your team standard. You can define it. You need a security scanning sub-agents for your compilers requirements, build it by your own. So these are the version controlled and shareable across the whole team, and can also run in parallel. And as for the domain knowledge, you can inject your own documentation, design system guidelines, and your coding standards directly into coders context.And when Qoder generates code, it will follow your rules automatically. So now, ⁓ here’s the key part. How you actually use all this. So you can think of Qoder like a Stripe for payments or Tailor for communications. You just embed it into your existing development workflow. You don’t need to replace your IDE. You integrate Qoder into VS Code, into JetBrains, into Temrano, into your CI-CD pipelines, whatever and wherever you already work. And there’s actually a precedent here. So Alibaba’s Model Studio platform has enabled over 800,000 custom agents across different domains. And we are also trying to bring the same extensibility here to coding workflows.Grace Shao That’s really really interesting. Thank you so much for sharing all that and I do apologize if any of the bit that I didn’t like sound that smart because it’s so technical but I really appreciate you breaking it down for me and really explaining it to me in very simple language. I have one last question for you which is something I always ask all my guests. ⁓ What is one differentiated view you hold? And this doesn’t have to be about work. It could be. It could be about your space could be about agenda coding. It could also be about anything else in the world. It could be about your experience in Silicon Valley versus China. Just a differentiated opinion review of something that you think is not that mainstream or it be a bit against consensus.HangSo we talk about something related to agentic coding. actually, I don’t think the transformer-based LLMs in the future. Because I actually do agree with the point that LLMs is just a compressor of the tags. They don’t really understand what they are talking about.But I think as the develops, as the technology goes, one day we will have the truly AGI. But it’s not based on Transformer.Grace Shao What does AGI mean then?Hang AI means the AI can really understand what it’s talking, what it’s doing, and really have emotions. Yeah.Grace Shao Fully human-like, emotionally, intellectually functioning being.Hang Yeah, like human are…From this perspective, they have emotions, have logic, they understand the work, the 3D work, and they have logic of, okay, why I’m doing this, and what I’m going to do in the next.Grace Shao Is that scary to think? Because in many ways, they would be more powerful than us, right? They can speak every single language on Earth. They will know more knowledge than us as an individual. But can we though? What if, you know, what if there’s charge and they can charge themselves these days, these humanoid robotics?Hang we can cut off the power. You mean to take off and charge by themselves?Grace Shao Yeah, like, do you ever worry about this AGI achievement going rogue?Hang To be honest, I’m both a bit excited about this and a bit scary.Grace Shao Yeah, I think especially for the next generations, right? Like what does it mean for them? Yeah. Thank you so much for your time. I really appreciate it.B. Christian Hu, Head of Global Marketing and Operations at QoderGrace Shao Christian, thank you so much for joining us. I just spoke to your colleague, Yu Hang. he was super helpful in explaining to me the technicalities of Qoder and the design of the product. My interest right now is really shifting towards the strategy and business side of Qoder. I’m really glad that you can join us today. So why don’t we start with the beginning? Why build Qoder inside of Alibaba? What was the thinking behind that? And what unique advantages does that really give you?Given that you have someone who just threw $53 billion into AI infrastructure that’s backing you, right? In terms of distribution, your infrastructure, your data, your model access, how does that really advantage you in many ways?Christian Okay. Thank you. Thank you, Grace. Thanks for having me here. You know, Alibaba has a very big plan and has very big ambition for AI ⁓ and his position in the future in the AI industry. And as you know, Alibaba has a full stack. We call the full step strategy from the cloud to model and to application. So that’s the full stack strategy. And for, ⁓ you know, a home computer, you know, we don’t build Qoder in a vacuum. Yeah, we build Qoder from the real context because we can find, we have got so many, maybe thousands of engineers from inside Alibaba. They are facing very real software problems in real software development. They have many issues in interface and how to fix the issues. But for the existing AI coding tools, they are maybe reactive. cannot...resolve some existing problems, maybe some complex and sophisticated problems in the real software development process. So they are trying to refer to a new product. we build Qoder just from inside. We have so many inside, the mind from the inside engineers. also, as you know, be part of Alibaba give us, mean, have so many advantage because first of all, we own models.Because you know, you know, we have early access to Qwen and that’s how I’d propose large-scale models We don’t just use Qwen we also reinvent and feed the Qwen with our real data from our AI coding Qoder base. So we just really reinvent the Qwen and we are coordinated with the Qwen team to improve the performance of Qwen models andSecondly, we own the cloud. I mean, the bottom layer of Alibaba’s AI strategy. that’s the cloud. Every digital tools and maybe even every AI coding tools should be based on large cloud infrastructures, consume very large computing powers. So cloud is also a very important factor in AI coding platforms.The third, think we, I think the most, last but not least, we also have the advantage of Alibaba has so many enterprise customers and so many business, you know, we got some business lines maybe from the consumer to the from the SMB to large enterprises, from the consumer to industries. So we got so many data across, as real data for us to, how to do involve the, the area AI to, you know, to ⁓ upgrade our coding platforms from the real data. that’s the, so, ⁓ so back to the, the origins of the Qoder So we just not, you know, we’re just not create another AI coding platform just for the tool, which we want to build a Qoder as a new platform for a Agentic platform for the real software. We want to build the Qoder for real software, not just for fun, not just for the fun making, but for the real software.Grace Shao Thank you, that’s really helpful for everyone to understand. I think saying we have a model is the most humble thing anyone can refer to Qwen as, because you guys have one of the best leading open source, open weight models in the world right now. Actually, so on that right now, you know, we’ve been hearing a lot of news about people adopting Quinn globally, and it’s really being used not just in China right now. ⁓ Who are the people using Qoder actually?Are they mostly Chinese developers or are you guys actually expanding globally? That’s, I guess, the first half of that. And the other half is, are these mostly professionals or are they like students or are they enterprises? Like, how do we understand the demographic here?Christian Yeah, good question. We are building Qoder for global developers from day one. Yeah, so that’s not just for Chinese developers, but for global ones. And for now, ⁓ actually for now, our users are mainly come from the, we call them individual developers, not just the enterprise users because know we the enterprise editions on it our way we are just we want to start with the individual you just at the first and and the funding you know geographically from China to the overseas market so we want to some difference between you know preference different preference for different you’re just in different regions yeah from in China they would prefer to the more, more integrated coding system existing with the existing customers. It may be fully integrated with existing systems. And in the Western countries, maybe in some in US or maybe some in ⁓ Malaysia, the users will prefer to more flexible workways. So they’d have a different preference for different coding languages. And for now, we are trying to solve the real software development problems. the most of the users come from the professional users. Because they have a real problem to resolve in the real software development process. Because they doing their work. Because they doing their work and they want to increase their productivity levels to solve more problems. So the main users come from the professional users.But we also find the increasing adoption of the individuals and some new learners, maybe some product designers, maybe some UI, maybe some UX designers, they want to deploy the AI coding tools to to expand some ⁓ new interface or maybe some new mini apps to increase their productivity is, yeah, so that’s the trend. we believe, uh, Qoder is created to solve the real problem real software development problems. But we also find that the, our agentic, uh, models and the questing models can allow, uh, more users, just like the new learners and new indies, new learners to, uh, new individuals to use our software to create something new, something more powerful. Yeah.Grace Shao So my understanding is that Qoder itself came out of the desire to help professional developers. But Qoder Quest, that mode you can go into, can actually help people with less technical backgrounds to be able to play around with it and potentially still build their own thing.Christian Yes, Quest Mode is maybe the key to unlock more space for web coding. What is web coding? Web coding is for non-professionals to use the Qoder tools to create something new. I think the Quest Mode may be the new way to unlock more web coding. But our Quest Mode is different from the existing agent model of other competitors or maybe other coding platforms because our Quest Mode is always is really is also a Agentic is a Agentic about also ⁓ we called a delegated and you can control the workflow and control the result on the question model. So that’s that’s for real software. Yeah, so the Quest Mode of coding is also is for real software.Grace Shao I see. I think I have a question for you. I asked Yu Hang earlier as well. But basically, my question was that, you know, there are a lot of tools out there already, like the co pilots Curser co whisper Warp you know, there’s there’s a whole array of them. How do I understand where Qoder sits so Yu Hang explained it to me in the technical sense of where it sets between co pilot and a agentic. Can you explain to me where Qoder sits in terms of like the business positioning.Christian Okay, we have a slogan for a Qoder We Qoder a agentic platform, a agentic coding platform for real software. So we got two keywords, Agentic and the real software. I just explained what is the software. I can explain more about the software because many developers, mean, no matter professional developers or non-professional developers, they found something critical in their developing process because they need to know the existing Qoder base. They need to understand what the existing Qoder means, what the existing documents mean for the codes. So they need to understand so many documents, so many files to understand what the coding process will be like.So we created a Ripple Wiki and a non-context memory function to understand the codes, understand the documents, understand the behaviors of the developers. So for the new entrants, for the new developers, they know how to get the real software down. So that’s the issue in real software development context.Yeah, that maybe seems different from what we call the rubber coating. So that’s for real software. And for Agentic, for most coding platforms, we call it your prompt and the Agent, the React. for Qoder we call it your delegate, then Qoder delivers. So then Qoder that deliver the real software and real results and the real apps for your delegations. So that’s the difference from the existing, I mean, maybe some most of the competitors to agentic platforms. So we are just not want to respond because of the delegation that can, you know, to free you from the desk. mean, for agent, for most agents, you also need to speak to the agent.You need to communicate with the agent while you are sitting alongside the desk. But the dedication can free you. You just don’t need to communicate with the agent. don’t need to interact with it for line by line, word by word. You just need to monitor the per science and maybe some need to check with the workflow so that the Qoder can deliver the result. That’s the difference between way from the other competitors.Grace Shao Actually, let’s take a step back. I’ve noticed that when I was doing some research on you guys that, you’re, it’s not just Alibaba that is creating these coding tools. You know, we have even just in China, we have ByteDance creating something similar. Obviously, Microsoft has things like has Copilot. So then we are looking at obviously then the startups and the Frontier Labs all kind of swarming into coding tools. Why is that? What is really the reason for focusing on coding right now as the next kind of use case?Christian Yes, actually, coding, as we just talked, coding may be acceptable for every engineer, maybe a non-engineer, maybe you’re just a digital learner, maybe just an analyst for the AI. So coding is the most certain way for token consuming, for the ⁓ infrastructure consuming, the tool to the cloud infrastructure. Maybe the coding is the most important way. And for different players, they are playing different games. For giants, you just mentioned the Bydance, maybe Microsoft, maybe AWS, they have very large cloud infrastructure. They just to integrate the new AI application to the infrastructure. And for SRAPAC, maybe some other front-end labs, are... ⁓They are trying to find the new path to the developers, maybe some to the application levels. for Cursor, they are just a new service. They are deploying the large-scale models to make it accessible to the new developers. Cursor has...I think for new startups, have their weakness, maybe they have their cost structure weakness. mean, they are the users of the large-scale algorithm. They can’t dominate the large-scale algorithm model and maybe they will be at risk to disengage with the large-scale algorithm model owners. So that’s the cost structure that is at risk for them.And for Qoder know, as I just mentioned, the Qoder is born within Alibaba. Alibaba has full stack from ⁓ cloud infrastructure to model to application. So I think that’s a great advantage. As you know, ⁓ as you just mentioned, Qwen is very popular around the world. some, you know, I just kind of found the news that Airbnb, Airbnb,Grace Shao Yeah, Brian CheskyChristian Yeah, yes, I said adaptedQwen open source model to the real invent their own infrastructure That’s very popular and for it for cursor. Yeah, it’s just anything for a cursor itself They just came up with a new composer composer model because it’s a small small model for coding but We have a speculation that the this small model comes from the Qwen or maybe some cheap. Both of them are from China. So we will find that the full integration and full connection with the model and the client structure will benefit a for Qoder in the future.Grace Shao Actually, on the point of how Qoder sits in Alibaba, can we kind of zoom in on that? Can you help us understand the big picture? Where does Qoder sit in the Alibaba Grand AI strategy across the stack?Christian Yeah, I think Alibaba AI strategy is a very big picture and also very long roadmap for what we call the super artificial intelligence. We call it super artificial AI. Yes, wait, wait, wait, ASI, that’s the roadmap. so from the map, so when we look at the map, I think it will form the bottom to the top level.I just mentioned is the cloud infrastructure. Yeah. The middle level, mean, the Qwen, maybe some other large Niagara models in the middle layer. And the Qoder, I think it sits on the application level alongside with some other applications. Maybe, you know, Dink Talk, maybe Quark, maybe some other applications on the application levels. think that’s the full, we call, so we call the full stack strategy for Alibaba in the near future. Yeah.Grace Shao I understand you lead Qoder, international operations and marketing, right? And that’s a pretty big title you got here. What’s your GTM focus right now? Where are you kind of focusing on essentially selling your product to? Like which markets, through what channels?Christian Yes.Grace Shao How is this kind of working out for you?Christian Yeah, so I think it’s a tough task. I know because we are new, we’re just a baby. We just launched our products about two months, no longer than three months. We’re just a new one in the market. our ambition is here. We want to ⁓ become the top global coding platform for global users.So for me, think the go-to-market strategy for ⁓ us to go global, think we need more partnerships, we need more integration with the local communities. So for me, that’s why I’ve followed to different regions to meet up with local community developers, maybe the leaders of local communities. We want to talk about it.What kind of preference for them, what kind of products they want to prefer in their context, in their context, maybe within the enterprise, maybe within the individual developing context. And also we are aiming to the global products, we are not going to separate.China with the global markets. We want to offer two different products. We just offering one product, one platform, one user interface for all the users around the global. So that’s the challenge. But we believe we can do that because we are trying to interact directly or maybe talk directly with ⁓ the developers from the different corners of the world.Grace Shao Yeah, that’s definitely quite different from how Chinese companies used to sell their products abroad, right? Because it used to always be a one app location or one interface, domestic one interface for the globe, for the rest of the world. So actually on that, think my final question for you really is just how are you navigating the current climate? Obviously it’s not.Probably not easy given just current situation with geopolitics, with the competition, with everything, right? How are you navigating differences between China and the global markets in terms of adoption, compliance, data residency requirements, and even developer culture? Because I know you were just in Singapore, like you mentioned, you’re flying around a lot, you’re meeting people from different parts of the world. How are you making it all work?Christian Mm-hmm. Mm-hmm. Yes.Quite different. I think it’s quite different. I have seen huge difference in the developers in China from the developers in Southeast Asia. Because I just flew to Singapore and met up with some local developers. The difference comes from different angles.User interface and the language and maybe some other features of the software. So the difference, is huge difference between the different rappers geographically and demographically. Yeah, so all the differences are very, very huge. And we also find some, because we are trying to offer our enterprise edition in the near future. So we also started a veryWe have made a study for the enterprise levels in different markets. We find quite different stories because in China, the developers in enterprise, maybe in the small companies, in large small companies, they would prefer the company to purchase the software. They prefer to use the existing software. The truth is using software can be embedded into the existing OA system or existing software ecosystem. So they prefer the existing, they prefer the integrations. But in other countries, maybe in other markets, they prefer more flexible and more athletic applications. They don’t care about the integration, they don’t care about the...I mean the sophisticated integrals, maybe sophisticated workflow and controlled by the company owners. So that’s the different culture. So you can find it’s very common for individual users to buy our software, buy our products in other countries, mean outside China. Yeah, so that’s the huge, that’s the huge difference.Grace Shao That’s really interesting. I would have thought that integration part is something that everyone kind of wanted because it makes your workflow so much more seamless, right? And that was the selling point for all of Microsoft’s tools essentially, right? In the software era. Interesting. Yeah.Christian Yes. But it doesn’t mean that we don’t need to make an integration or some ecosystem with the... You just mentioned Microsoft, mean, the OA or maybe some Workday, some other software platforms. We also want to explore more connection with existing software.Grace ShaoThank you so much, Christian. I really appreciate your time. Thank you.Christian Thank you. Thank you so much.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• E-Commerce Evolution: AI and Live Streaming in Retail, with former Alibaba executive Sharon Gai 04.11.2025 55min

  “Retail is simple. Retail is just how do you sell something, and make someone’s eye light up. AI or any technology you add to it, is just another way to do that,” — Sharon Gai, retail tech and AI expert, former Alibaba executive.Joining me today is Sharon Gai, an expert in AI and innovation, with a focus on retail. She was an executive at Alibaba, where she advised brands and heads of state in crafting their digital strategy with programmatic marketing and AI. In this conversation, Sharon shares her journey from working at Alibaba to becoming a consultant in AI technology for global companies. She discusses her experiences in e-commerce, particularly the evolution of live streaming and innovative marketing strategies in China. Sharon emphasizes the importance of AI integration in retail operations and the future of shopping with AI avatars. The conversation concludes with insights on simplifying retail to focus on core selling principles.Sharon was selected as a RETHINK Retail’s Top Retail Expert and a LinkedIn Community Top Voice in 2024. She has two books, E-commerce Reimagined and How to Do More with Less Using AI. For more of her work, go to sharongai.com.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Chapters02:27 Experiences at Alibaba: The Global Leadership Program05:11 E-Commerce Evolution: Insights from Tmall and Live Streaming07:54 Innovative Marketing Strategies in Chinese E-Commerce10:23 The Rise of Live Streaming in E-Commerce31:33 The Evolution of AI Avatars in Retail34:09 The Impact of AI on Shopping Experiences41:00 Challenges in Retail During the AI Revolution43:22 Integrating AI into Retail Operations48:43 The Simplicity of Retail: A Unique PerspectiveTranscript [AI generated]Grace Shao (00:00)Hey Sharon, thank you so much for joining us today. It’s great to reconnect with you. We met like, I think four or five years ago back in Shanghai and you were still with Alibaba, right? So why don’t you start with telling us about your journey? Like you grew up in Canada, you worked in Hangzhou, Alibaba. I know now you live in New York. How does it like, how did you kind of bring together all those expertise to what you do today, which is help?Sharon Gai (00:27)Sure. So when we met, I was working at Alibaba still. For me, as somebody who was born in China and raised in North America, and then I chose specifically to go back to China to work for a bit, the reason why I did what I did is I just knew that going forward 10, 20 years out, the two major superpowers would be the US and China. And I already had a pretty good understanding of things that were going on in North America.⁓ Going back to where I was born and getting the chance to work there at one of the tech companies there really opened my eyes to how both countries work. I think in the future it’s going to be a ping ponging back and forth of I’m sure different ⁓ globalized companies and projects between the two places.⁓ And so that had blended pretty well to what I do now, which is a lot of writing and keynote speaking and consulting for global companies that both have a footing in China and the U.S. So it all ties together pretty well now, but definitely as I was going through my through line or life trajectory, it seemed very confusing in the beginning phases.Grace Shao (01:35)Yeah, tell us a bit more about your time at Alibaba, because I know you were part of quite a special cohort. was a of a test and trial group of international cadets, per se.Sharon Gai (01:47)I’ve never been in cadets before. But I guess with anyone joining Alibaba, it feels like going into entering some, a ⁓ corporate army of sorts. But the program was originally set up by, or it was a brainchild of Jack Ma’s. He had always, I think he had similar thoughts, which was, you know, eventually you’ll hit the ⁓ bottleneck of about a billion or so. ⁓Sharon Gai (02:09)internet users in China, where do you then grow the company beyond the billion users? You have to find it outside of China. And so the first place of external search was Southeast Asia and then into the Middle East, Africa, Europe, and then eventually the US. And so his long-term vision wasto recruit people who came from those places, those corners of the earth, to get them to come to China to be in green with Alibaba’s culture, a way of working, and to bring them back out again, and then ping pong back and forth, just as I thought. So I think his vision and my own personal vision aligned pretty well. So that’s what got me to join the program. And yes, it was definitely an experiment. There were many, what I would call seasons of us or cohorts.every single year there were new people that came in, from different cultures based on the strategy of the company at the time. think in certain years, they really wanted more, a certain language to be spoken. So they, really hired for that specific language. and it, definitely changed at different versions. but the idea was to bring in.people who are bicultural, multicultural to eventually lead some of the business units that was trying to expand outside of...Grace Shao (03:22)And actually on that point, what were you doing at Alibaba? I believe you were involved with Tmall, right? So the international business, flagship business of Alibaba’s e-commerce sides. Could you tell us a bit more about that?Sharon Gai (03:34)So the first sort of business unit I was in was called Tmall Global. Our larger BU was called Tmall Import Export. of course, as you would know, Chinese tech companies always like to change names and just change things. Embracing change is one of the values.So at first it was called Tmall, import and export. And I was first on the import side. And then I went to the export side, which is what we call Tabout Tmall world, where there’s about 50 million Chinese diaspora around the world. And they also will use Tabout as a shopping app. At the time was also the growing footprint of the Shiians and Tmus of the world, where these local Chinese e-commerce apps were trying to leave China.And so Taobao Tmall World was also part of that exercise. And so those were my main two. the first was, or sorry, one of them was Taobao Tmall Import-Export. And then I moved to Tmall Classic, which is the domestic side of Tmall.where the brands, most brands were either Western brands selling into China or Chinese brands selling to local Chinese consumers.Grace Shao (04:47)So you’ve really got a good look into how the retail e-commerce digital evolution happened in China. You’re super plugged in. And you were there for like six, seven years, right? So when you left, you published a book, you published E-Commerce Reimagined. And I believe at that time it was COVID, pre-COVID, and just everything kind of changed again. And we saw the rise of e-commerce really.⁓ being part of the daily lives of North American consumers as well. Tell us a bit about like, I guess, first your book and then tell us about what you witnessed over the 10 decades, sorry, about six or seven years while you were Alibaba.Sharon Gai (05:25)Yeah, so in 2017 when I joined, was the height of live streaming starting in China. When I joined T-Mall Classic, I was actually one of the first teams to set up a live stream and take it to the US. So funny today, I’m back in, funny today I’m in Kuala Lumpur, Malaysia where I’m joining you for this podcast, where Jackson Wang had a concert here yesterday.And he was one of the celebrities that we collaborated with first to do one of those live streams. At the time when we were trying out and testing out live stream rooms, we didn’t know about the flow, how to direct things, what questions to ask him, how do you showcase the products in a very natural, organic way. All of those were tests and experiments that we figured out throughout the process.But during that time, I mean, in a lot of, so I do a lot of keynote speaking today and in a lot of the keynotes that I do, I always start off with comparing just the size of the consumer economy of China, where it is the largest one in the world. It has the most number of internet users. It also just has a very voracious consumption habit. It also has the highest,internet penetration in the world, the number of mobile users in the world, and people, and out of those users, people who are buying things online. There’s a high amount of trust online because historically from Tmall, from JD, there’s been very, very high standards by merchants. So when merchants enter a marketplace, there’s usually very demanding.terms for them to host returns, be able to accept returns, to deliver things on time. And that standardization eventually increases trust in the marketplace that even if there is a new seller, new tab out seller that emerges, the consumer will most likely trust them versus if you had that same transaction happen in the US, there’s a lot less trust in the marketplace.So 2017, 2018, we’re laying out all of these foundations. And I think what I took away is the immense competitiveness of that space. And so out of competition, naturally there’s more innovation because as a merchant, you’re fighting for the same eyeball that your competitor is. So either you’re going to lower your price orbetter your brand or better your quality. There’s some sort of lever that you have to tweak to be better than your competitor. I’m sure you’ve heard the term involution. In China, it’s an involutionary environment in schools, in academia, and definitely in business. And so from a brand’s perspective, there’s just a lot of...⁓ playful and competitive things that they can do and that’s definitelyGrace Shao (08:20)Can you give us some examples?Like what kind of playful gimmicks or tricks would people have to use to kind of draw more eyeball over?Sharon Gai (08:28)So back then I was part of ⁓ the Tmall food team and I was in category management at the time. And this is a traditional sort of key account structure in all sorts of countries or categories where your top accounts will produce roughly 70 to 80 % of your total GMB.And then the next tier, mid-tier type of brands will produce about 10%. And then the long tail produce another roughly 10 % or so. And so in...Grace Shao (09:01)So is that just in China for Taobao and Alibaba? Is that like a reflection of the whole industry similar kind of to how the marketplace play out in the US?Sharon Gai (09:10)It’s similar in the US for sure. If you look at the large FMCG companies, Procter & Gamble, Unilever and Nestle, the large three will take up about 70 to 80 % of the market share. And then depending on the subcategory or the category that they’re in, they will pretty much dominate the market or they’ll dictate a lot of how the market shakes out.Grace Shao (09:12)Okay.Sharon Gai (09:36)So during that time, or let me preface by saying that’s what you see in a traditional offline heavy space. So in a supermarket, for instance, where things are a lot more traditional. So in 2017, 2018, around that time, it was when online, there was that huge push into online. And in the...food space, there was a lot of DTC brands that sprung out of that. At the time, China didn’t have that green light, red light policy. So the consumption sector was really robust. lot of consumption startups, FMCG related startups raise a lot of VC dollars and wanted to IPO. So at the time with their VC dollars in their hands, they were able to create all sorts of new brands.So the Chagees, the Nice Nose, these sort of tea brands and bakery brands that you see often in Southeast Asia or in China, those were all sort of created around that time, 2017, 2018 time. And I remember...going back to your earlier question, so what sort of creative things did I see? There was this milk brand that experimented with adopting cows. So in China, depending on where your listeners are, the dairy industry is very traditional. And in China, there’s Meng Niu and Yili are sort of the two very, very archaic, hundred-some-year-old companies.been in China just for so many years and everyone knows them. And around that time, would not really do much in the e-commerce space because they knew they had that foothold in the market. And so out came this new brand called, in Chinese it’s called Zhinyang Yitou Niu, which translated is Adopt a Cow. Funny name, and I think part of their success was alsothe cleverness in developing that name. I wrote about this case study in my book too. Their goal or the goal of their founder was traceability. So a couple of years earlier than that, there was this huge milk powder incident where several babies, I think infants died because of this milk powder was poorly made. And so his...Focus was could we have every family in China quote-unquote own a cow that they can milk from afar? And that when they get their milk delivery, it’s from that cow that they’ve raised or that quote-unquote they adopted So that was his ultimate vision That’s how that’s how the name came about⁓ and the things that we did was, selling milk cards. we, so, so traditionally on an e-commerce, have skews where you’re buying the actual product. This was the first time where we’re buying, a, almost a gift card, but this card would just be for this one brand and it would be sold at a much steeper discount. so.That’s why on the consumer side, you would want to buy it. On the merchant side, you would want to sell it. And we would roll this product out to a lot of other categories for a lot of other products where people knew they would be spending that amount of GMV. So anything like toilet paper, rice, flour, milk, health supplements, any sort of category where you knew you were gonna spend that much.in a year’s time for you yourself and your family because it’s just one of those products that you frequently shop for. And so that was something that eventually the entire platform rolled out to the entire team all rolled out, but it started with that one company. And the reason why that company did that was because there were two very large incumbents.And the only way that he could compete is if he was innovative enough to think of something else. And throughout my time, I worked very closely with that team to see them from day one where they had two followers on their store to now if you go and look at that brand, it’s the millions of followers they have. think by now they’ve also IPO’d because they were able to IPO before the red light, green light policy.and they set a new standard because, out from out of that, there was a new type of, DTC, all sort of mindset that came out of T-Mall where, you know, if you were a no-name brand, if nobody knew who you were, as long as you were able to think of attractive enough and innovative enough things to keep attention and to keep your consumers.to for you to for first of all your customers to discover you and then for them to come back again and again you you were able to Survive even if you were a newcomer and even if there were very large established incumbents in your category and so that was a slice of what I learned and witnessed and there was many many more examples, butIt’s a that one’s a pretty notable and memorable one for me.Grace Shao (14:47)That’s like so interesting. have so many questions, but I don’t know if we’re gonna be able to double click on everything. Number one is, do you think it had anything to, this is like, we don’t have to go fully into this, but I do feel like 2017, 2018, like you said, it was a peak of like also private equity in China. And like, did you see consumer brands were just like going nuts? Like whether it’s like makeup, cosmetics, like FMB, definitely like a lot of brands peaked.Sharon Gai (14:49)No.Grace Shao (15:12)I’m surprised that this one’s still living, which is great, but it’s just like a lot of the makeup brands, like, you know, Perfect Diary also kind of did this, like they kind of just fizzled out, right? There was like these coffee shops all across the country that were coming up that were selling like specialized filtered coffee, whatnot, right? So it’s interesting to see that this company was able to utilize innovation in their marketing and branding, but actually sustain this business.Number two, comment is how can they actually track which cow? It seems pretty crazy. Like surely they weren’t actually getting the milk from that specific cow, right?Sharon Gai (15:49)Which one should I answer? On the note of the cow, that was his vision. No, today, think, well, I also stopped following the brand after I left the company. To my knowledge, it did not go to that extent. However, the extent that it did go to was the founder did do a lot more beyond just the cart.Grace Shao (15:53)Yeah, so it’s not actually like played out, right?Sharon Gai (16:14)He also wanted to and I’m sure you also probably covered a lot of it 2017-2018 was a big push in agriculture from a Chinese government standpoint. He also wanted to create these he wanted to turn his factory into like a like a like a touristic activity where people could go visit the cat like your cow or you could go andGrace Shao (16:35)Okay.Sharon Gai (16:36)and experience how milk was made, how it was pasteurized, the entire process. It’s actually a very, very complicated process before it gets, know, we sometimes take for granted like packaged or milk from our fridge is taken and drink it, but it goes through so many steps. he, so within that sort of touristic factor, you could see,certain cows were penned off and there was a video camera on each of them and each cow was numbered. So you could technically see it, like see your adopted cow. But to date, I don’t think it eventually, or that model stayed, but there were many, experiments that he did. And I’m sure there were...Grace Shao (17:11)Okay. Okay.⁓Okay, we’re going way off track.session with these cows. I’m just like bring it back and I’m going to look up these cows afterwards. But it reminds me also this Alibaba thing where people try to gather points and they can own their own trees and they can plant a tree on Alipay. Right? Okay. So anyway, let’s bring it back. Okay. Let’s rewind and go back to retail and AI. and technology and innovation and retail. So live stream, you mentioned it, 2016, 2017.Sharon Gai (17:37)Mm-hmm. That’s happening.me.Grace Shao (17:51)You’re one of the first, you know, part of the teams that were kind of, you know, really, I guess, leading the frontier of live streaming technology or even the strategy. It’s still not that mainstream in the West. So to start with, could you tell us what even is lives from e-commerce and how has live stream shopping really change in e-commerce in China? Is that something that you’re seeing? And like, I guess, ⁓adoption in the US right now with TikTok or anything else.Sharon Gai (18:22)Mm-hmm. So live streams, e-commerce is when a seller and you can be a brand or you can literally be a person who owns something that you want to get rid of on the internet. You turn on your camera and you showcase this product live to an audience. And this audience could be some could be your actual followers or just be strangers. And then this sale is made technically the old in the old days on the in the western side.The first adoption of live stream shopping was on Facebook live where Facebook live used to be a big thing. People would go live all the time. And then you host a room and then you make a bunch of sales and you’re actually recording a lot of the addresses and whoever bought you bought something by hand. It was very, very manual. And then on tab outside, they started with live streaming in app.And in the US, think, you know, the Walmarts, Amazons tried their shot at live streaming. They invested a lot of dollars in creating these very posh and professional looking streaming rooms, setting up streaming studios. And all of that didn’t really go anywhere. So it had its spurt of interest. I think I think at the end of the day, it’s a it’sOne, it’s a timing thing and two, it’s sort of the way that it’s done. Yeah, and then I think the third, the host has something to do with it. So first timing in China, when I played with it, it was in 2017. That was still newish in China. China didn’t really take off with it. It didn’t really become mainstream until 2019-ish. That’s when we really understood. That’s when every single.brand, merchant, you that, and if I’m outsourcing my operations to a TP, my TP better have a streaming room in their company. A ⁓ TP is a team all partner. if you don’t want to run your own e-commerce operations, you can outsource this to a company that knows this space very well. And really,Grace Shao (20:12)Sorry, TP is like a, a TP is. T-Mall partner, yes. And to give the audience, sorry to interrupt, but to give the audience some context, like how big are we talking about live stream e-commerce? Like give people like the headline numbers, like a leading influencer during Double Eleven or their annual GMV. What are we talking about?Sharon Gai (20:40)I think in total, I don’t know the exact numbers, but it’s north of a couple billion per year for sure for some streamers. Which is crazy.Grace Shao (20:55)which is crazy. This is like a single salespersonif you think about it.Sharon Gai (21:00)Well, on camera they’re one person, but behind them is hundreds of people. And I’ve been through their product selection process. I’ve worked with them throughout the night. They do not sleep. Their streams start at 8 p.m. That’s not actually when they wake up. A streamer’s day is at 2 p.m.ish in the afternoon. They’ll wake up.Grace Shao (21:03)Yes.Sharon Gai (21:21)and they will start getting prepared. Their director will usually tell them today, is the, you know, we’re going through the final list. This is the final price that we’re gonna sell all these products at. At 8 p.m., usually 8 p.m., maybe sometimes six, sometimes nine, the show starts. You’re streaming for about four to five hours, so you usually finish at midnight-ish. And then midnight.Starting midnight to around 5 a.m. You start to review what you just did because you have literally just talked for 45 four to five hours You sold a bunch of product products during double 11 It might be tons of millions of dollars that you’ve just sold per show And then tomorrow it’s gonna start again. SoYou’re going to review everything that’s worked, that’s not worked, what you said was right or not right because some of these products repeat again the next day. So they’re always fixing their script and they’re always fixing how to say it better.Grace Shao (22:15)they buy the inventory or they take a cut from the brands? How does a business work?Sharon Gai (22:19)They don’t buy theinventory. Streamers will not buy the inventory. The only fee that’s given to them from a per brand basis is a, is a, a, is is a, like a slot fee. So if I’m a brand and you’re the streamer, I’ll pay you $5,000 for you to show my product. But beyond that, you’ll also take a 10 to 15 % commission per product that you’re selling.Grace Shao (22:48)So this is very differentfrom how Instagram influence is kind of where my point is. Like for Instagram influence, it’s like, okay, you place this product here, that’s one off fee versus a live streamer in China is actually making money continuously as long as like people are buying through their link, right? And people sometimes don’t realize that.Sharon Gai (23:07)They don’t. Yeah. So this very traditional financial model for streamers in China is not the same in the US. In the US, it depends. Some streamers are just hired to talk for two, three hours at a time, and they’re almost paid on a per hour basis. Some streamers are paid.They’re paid a set fee for the entire stream. It’s not even commission based. Their whole job is to, if they’re selling clothes, they just put it on, turn in front of the camera, comments will fly in and say, can you try this on another color? They take it off, put it on the other color. And so it’s almost very robotic and you sort of follow what the comments are saying. It’s not commission based. So they think less. They have to think less. They’re a lot less strategic also with placement of products. Which one goes first? Which one goes second?because that also changes, that also impact GMB. So the two financial models are pretty different. But going back to your question earlier, is this happening in the US? It is absolutely happening in the US. It’s actually, the last numbers I checked was about $60 billion in the US, and that was last year’s number. So it should be a lot more now. The biggest US player is this app called Whatnot. I think that’s the most successful app.They’re about as there. think they’re a series D startup now. They started with collectibles and I know some of their early collectible sellers. They sell trading cards. That’s the start of live streaming in the US is trading cards. Like there’s a huge fan base for these trade. There’s this Italian brand that makes this specific card and there’s a bunch of also intricacies with how to detect if a card is fake or if a card is real.⁓ but, that’s the starter product is very niche.Grace Shao (24:52)So still quite niche. It’s not mainstreamand prolific in China right now, like how everyone knows what live stream e-commerce is, like any auntie on the street will know about it. So it’s just a very different kind of model. I wanna bring it back and talk about what that means. Does it mean it’s like...shopping felt more personalized or did it mean like you said it was more rapid? Did it mean more interaction? Like I just want to see does that kind of foreshadow the future of technology and shopping experience?Sharon Gai (25:23)in the US? ⁓ I think it does. does. The way that I look at it is at the end of the day, you just want to be more, you want to reduce your barrier between the merchant and the buyer. The reason why live streaming works so well is because it’s live.Sharon Gai (25:42)Um, it’s so the, in some of the keynotes that I do, one slide that I have is the evolution of e-commerce. So in the nineties, you Craig list type listings, where it was just a product title and a description. There were not even photos because you didn’t have to put any photos. So some sellers were lazy and they just wrote down that I’m selling this leather jacket for $200. Um, and then Amazon came along and enforced the a plus content or, um, their, their product detail pages have to be.put together in a certain way. So it became photos and then there were, was mandatory to have product videos. Then it was mandatory to, well, that part wasn’t mandatory, but then e-commerce became different influencers showcasing it. And then TikTok started Instagram, all of that. And so it’s pretty natural that things will just evolutionize. I think this goes back to competition.which is China is just more competitive. So they jumped into live streaming a lot faster because we noticed that the merchants that live streamed did better than the ones that didn’t. It was as simple as that. And so the ones that didn’t would start doing it and they would have to, they were forced to learn it even if they didn’t really want to because that’s just what worked. And so in the U.S. it’sit’s worked for these collectibles, toys companies. And I think other brands see that. now have, now what not has become cross category. Now it’s, they started with collectibles, but now a lot of fashion brands have jumped in, food brands have jumped in. Supplements also. And then in the TikTok shop world, it’s pretty, it’s more and more common for D to C brands to start streaming also.⁓ to outsource their streaming to. Now there’s a new term called TTPs, which are TikTok Shop Partners, to do this for them.Grace Shao (27:32)I have a question. seems like in some ways that the China side, you said, the influencers, like not influencers, the live streamers themselves actually don’t need to be influencers. In many ways, it’s not that people go to them. It’s not like they build any, they bring any extra credibility. They’re just the avatar or, you know, the mannequin in many cases. Whereas in the U.S. it’s e-commerce is so heavily reliant on like the actual influencers, like, know, which celebrity endorses what.you know, makes a big deal. Do you think we’ll continue to see that kind of divergent path? Or you think, you know, people will actually care less and less about who the influencer is?Sharon Gai (28:12)It’s actually kind of the reverse if we’re just talking about live streaming in that in the US, the people who are streaming are sort of no name people. And then in China, it’s the well known streamers, not the celebrities. I mean, on both sides of celebrities, let’s take out celebrities because they, ⁓ you know.Sharon Gai (28:35)They’re their own world. Also, some of them doit sometimes. Yeah, there’s definitely pro, there’s definitely very well known live streamers like the Austin Lees of the world. They’re now a semi celebrity themselves because they’re so famous now. Like that type of profile is non-existent in the US where they’ve just gotten so good at the streaming side of things where they’ve in the branding side of things.Grace Shao (28:53)Mm-hmm.Sharon Gai (28:59)So will it diverge? think for the US at least, you’ll have more full-time B2C, more traditional Instagram influencers jump in and try it out. And it’s definitely gonna be a word of mouth type of thing.⁓ like what I see on tech talk today is, there are certain influencers who, who are traditional influencers and now they’ve gotten in, they’ve jumped into the live streaming side. so I think in the U S it’s still a testing the waters type of thing. as for China, I think the big concern is the AI piece, where now if you go to any branded room,If it’s an off hour, if it’s an off peak time, it’s some sort of AI avatar that’s streaming. To date, I’m sure you’ve heard of the Loyong Hall live stream room in 4.6.18. I think it was this year where he did the AI live stream and he sold X million of dollars. And it was more than he would do when he streamed himself.in person as a human live. And that in the media space, there was a big wave of, are we going to be replaced by these AI streamers now? I think that was definitely a big, was definitely more of a PR push. Like it’s very easy to tweak those numbers so that you do make it seem like your AI version can sell a lot more. ⁓Grace Shao (30:24)Mm-hmm.Sharon Gai (30:25)You can play with pricing, the product that it sells, a lot of things. Yeah. So, but so in China, I think the future is going to be, are those big name streamers going to want to employ an AI version of themselves to conduct live streams? I think that is something that a lot of them are thinking about. Like once we let it happen, are we going, is it, how is that going to change the industry?Grace Shao (30:50)Actually, that’s where I really want to kind of take our conversation to next, which is like, we’ve talked about live streaming and that was, that’s kind of like the next technological, I guess, breakthrough in e-commerce for the US given that like, you know, Asia, East Asia is already like really, have, e-commerce already really, really, sorry, I want to say is like this live streaming model is already very prevalent across East Asia. ⁓And like you said, the next stage of technological breakthrough for e-commerce in China or even including East Asia is really in avatars or AI avatars. And we already kind of saw Alibaba with their kind of like fake avatar, like these like cartoon avatars for a while now. Can you tell us a bit more about that? Like actually give us the context of likehow the AI avatar technology within Alibaba has evolved over the years and what we’re going to potentially see in the future. Because what you just said with the loyal health thing is essentially this frees up his actual free time, his time, and his avatar can just sell for him. Then do we still need him? Or, you know, is he going to become the brain behind it? And what’s the future of AI within this, this whole conversation?Sharon Gai (31:57)There’s I think there’s a difference to note between a real person creating an avatar of themselves and Being like the the Wizard of Oz the Oz and like puppeteering their avatar Versus I think what you were talking about is the Aya ease of the world where it’s like very futuristic looking She is a completely fictitious person personGrace Shao (32:02)Mm-hmm.Yes.Sharon Gai (32:23)Her face is made up, the things that she wears is all made up. That type of, maybe for a better term, that type of sort of metaverse looking, futuristic looking thing is not so much an avatar. But it does happen. There’s also, Lil Mikaela is similar in Brazil.Grace Shao (32:40)Mm-mm.Sharon Gai (32:47)She’s also like, yeah, she’s like forever 19 years old. She’s worked with all sorts of major luxury brands. We’ll work with her now. So that model is proliferating around the world. I think that way of selling something or the evidence of that in the market was pretty big in...the 2020-21 time of the world, it’s definitely died down a bit. To date, I have not seen new versions of Ayayi because technically if Ayayi worked really well, there should have been many, many versions of her. from a brand standpoint, we should also see, like each brand would have sort of their own version of Ayayi. think this brand called Florisys.Grace Shao (33:22)Right.Sharon Gai (33:39)which is this makeup brand. created one too, but she’s also kind of disappeared. So I think that that... Right. It’s like another look. Yeah. Yeah.Grace Shao (33:46)It’s almost like mascots, the mascots, right? He had to see like, right? Like, and they kind of just died and yeah, but people actually still wanta human looking thing, whatever, like a human being to try on the product. So to your point, like the Loi Yonghao example, it’s like people still like the fact that it looked like him, like a human, but it’s not him, right?Sharon Gai (34:09)Well, I think the Lui Yonghao specific one was just the novelty of it. yesterday, I was trying to buy something directly in Chachi BT because of their partnership with Etsy. I just wanted to see if it’ll get delivered, when it’ll get delivered versus if I just bought from a traditional dot com website. Like there’s a novel and I just came out of Capgemini.panel a couple of days ago before that where every single panelist was like, we’re all shopping and try GPT right now. like, how, I don’t know how long this actually gonna last.Grace Shao (34:42)How good isexperience good? Because I can’t do it in Hong Kong. Is experience good? you... Is the whole...Sharon Gai (34:47)Yeah, you can definitely buy things within the app for sure. Not every single product.Grace Shao (34:55)But is it a good experience? But is it a good experience? how does it differ from going to a Alibaba.com versus like a Taobao.com versus like Amazon.com?Sharon Gai (35:04)I think it’s very dependent on if you’re very sure of this product because the weakness of ChadGBT and just shopping via LLMs is it’s way worse display. Because usually product photos are very beautiful and they hired models and background sets to picture this product very nicely. But in ChadGBT, a lot of these pictures are compressed.Sharon Gai (35:29)If they were originally long, they’re like short now. Like a lot of these photos are, there’s no standardization. And then also it cannot show videos. It can show YouTube videos inside a chain of chat, but there’s no, like if there was a product video, it can’t show the video. So if it’s something that you already know that you already buy all the time, then I think that’s, it’s.Grace Shao (35:29)Yeah.Sharon Gai (35:51)It’s easier to buy through an LLM, but if it’s something that you kind of want to browse and look at and maybe like flip through a couple of pages of reviews or photos, it’s not the best shopping experience. I would still go back to our traditional dot com website.Grace Shao (36:04)So basically like the agent will just go from your intent all the way to purchase all like in one go. You don’t have to click through a bunch of buttons. You just say, buy me this using this credit card go. Kind of thing. Okay. Actually tell us about more about that. Like are you seeing more brands that you’re working with right now adopt like AI within their purchasing process?Sharon Gai (36:28)⁓ From the perspective of a brand, I think all brands right now, the big to do for all brands right now is AEO, if you’ve heard of that term. So it’s a differentiation from the old SEO world where if I wanted to buy like a winter pea coat, I put it in Google, a bunch of blue links show up. I click through the Macy’s link, the Bloomingdale’s link, maybe another one.Ralph Lauren or something. And I look at it and I browse through it, maybe at Dakar, maybe at Leave, I go to another site. That type of experience is being changed because now a lot of people are searching for their products through a ChachiBT, a Claude, a Perplexity. And so now as a brand, how do you not show up as the blue link, but how do you show up in the answer? So if I’m putting in the same query,So I want to buy a winter p-coat and this is my budget and this is my type of style. I want it for this warmth because New York is usually zero to eight degrees in the winter. What are my choices? And then there will be choices that show up and some brands will show up and some brands won’t. So to get your link or to get that skew to show up, that’s called AEL or stands for answer engine or optimization.Grace Shao (37:49)And how do you get on that list? How do you make sure you’re on that list?Sharon Gai (37:50)There’s a lot of things you have to do. And it’s also something that all brands and also startups are trying to figure out right now. So it’s like an empty space for startups. in, give it another two quarters, you’ll see a lot of seed stage startups that do exactly this, which is an AO. ⁓AEO product, like use me and I will make sure you show up in an answer engine. It’s not a sort of, a, or fortunately, it’s not a very straight cut thing. Like if you do X, then you will show up. There’s a lot of things you to do behind the scenes from just reorienting your product detail page to make it more crawlable.Some websites want to wear off bots. Whenever you see the Cloudflare pop up where it says, check if you are a human. Most websites don’t want bots to crawl all over their site by default. And so you have to remove that. You have to also make sure that your product detail pages, or any web page rather, is very, very intricate. So much so that if anybody searched for a certain ⁓ question,that whatever your product detail page says will pull up. So for instance, even with the example that I just said, a peacoat in the winter for New York, and it’s perfect for zero to eight degrees. That was part of my question. It’s very specific because the way that we search has just changed. We went from typing a couple of keywords to Google to now, even for yourself maybe when you look into China GPT or when you’re looking for an answer, your prompts are getting longer and longer.Grace Shao (39:15)It’s so specific.Mm-hmm.Sharon Gai (39:32)just what you are looking for is more and more, because it’s easier and easier to find information that directly hits what we want to find. So in the zero to eight degree example, if one coat seller had that listed, would surface in the answer engine versus the coat seller that just said, this is a beautiful coat. And like providing no additionalGrace Shao (39:49)Right.Sharon Gai (39:57)⁓ intricate detail for an answer engine to pull off of. But beyond that, there’s so much more. There’s more on reviews, how well your brand is guarded, general sort of reviews, Google reviews of your brand. And then back in the day,Grace Shao (40:12)the credibility.Sharon Gai (40:14)Back in the day, these LLMs pulled a lot from Reddit. This month or these months, they’re removing that Reddit portion more and more because people say random things on Reddit. Also, Reddit is 50 % bots anyway. So they’re cutting down the amount of information they’re going to pull from Reddit. They’re now going to pull more from very more credible sources. But to get your...Grace Shao (40:18)Mm-hmm.Yeah.Sharon Gai (40:38)a coat to show up, a lot of brands used to do Reddit campaigns. So they would hire an agency and create a subreddit for their brand and make sure that these skews from their new collection was very well reviewed. a lot of them were fake. But that at least it was talked about in the Reddit community because Reddit content was very highly regarded by LLMs. And it’s sort of a moving target.Grace Shao (40:57)Mm-hmm.Yeah. Would you say like,yeah, would you say then like getting on the AEO is probably like these retailers biggest pain point or choke point right now?Sharon Gai (41:11)It’s definitely on the forefront for a lot of them, for sure. ⁓Grace Shao (41:15)What other issues are they kindof facing during this like kind of this AI evolution or revolution?Sharon Gai (41:21)⁓In regards to, well, right now for retail, it’s just a very tough quarter. Like this year, we had no growth in retail in the US ⁓ and the Q4, we might go backwards a bit. So we might go negative one percentage point, but that’s mainly caused by just people want to cut back on spending. People are buying more private labels instead of original, like being loyal to brands. They just want something more.Sharon Gai (41:48)bang for the buck, more quality over or same quality for same dollars that they’re spending. And then with all of the inflation, all the tariffs, it also makes goods. There’s just a lot of instability in the market. So everyone is trying to hold on and just maintain and survive. There’s also so many retail chains that’s closed andGrace Shao (41:49)Mm-hmm. more expensive. Yeah.Yeah.Sharon Gai (42:14)just gonebankrupt in the past year. So 2025 has been just a major shakeup for retail. I think a lot of brands want to, their CFOs are telling them to AI-ify their teams and to start employing agents instead of more people and increasing headcount in their teams. But I think for a lot of retailers,Grace Shao (42:35)And what kind of agents?What kind of agents are we talking about?Sharon Gai (42:37)Agents for all sorts of every segment of the value chain from product creation, the merchandising piece to buying. So how can you AI-ify your procurement team? Maybe not completely remove the person, but at least stop them from increasing the team and to start outsourcing a lot of the work to AI. To the redesign of websites, writing, copy, creating.product detail pages, creating marketing content, how do you sell, how do you market this product better to your consumers. All of that can definitely be, your productivity can definitely increase from ⁓ using AI and that’s what a lot of brands are thinking about doing and implementing this year and probably the next.Grace Shao (43:08)And that’s where a lot of the work focused on that you do. You help them integrate a lot of these AI tools and the AI agents into their work processes.Sharon Gai (43:34)Mm-hmm.I think a big, actually, missing link from the retail side is there’s no one... When people learn about AI, it’s through today, I read this headline. Tomorrow, I read that headline. The next day, I took an executive education course with MIT and I got a certificate. It’s all... It’s very...It’s very sporadic and it’s very disparate and everyone has different information points. There’s, I think there’s an absence in learning the foundations and the fundamentals and the applicability, the application part for your business. think those three core things, a lot of retailers are not formally doing because from the tech company side, now you have AI skills as...Grace Shao (44:20)Mm-hmm.Sharon Gai (44:25)mandatory, a mandatory piece of your job. But retail is so retail is a very traditional business, actually. So they’re there, unless you’re a D to C type of company, so your your team is naturally younger, everybody’s in their 20s, they’re already playing with Sora on their phone every day. You won’t really be part of that sphere.Grace Shao (44:39)Mm.Sharon Gai (44:48)So what’s needed is a level set of just learning about AI. Yeah.Grace Shao (44:53)Yeah.So if you are, I like this question and I always, I thought it was quirky, but if you were, if you were given a hundred million dollars to really help revamp a traditional retailer, like let’s say Macy’s right. Or Bloomingdale’s or whatnot. how would you actually,use and spend that money in terms of adopting AI and technology? What areas would you actually spend the money on if you want to make your company future proof in the next five to 10 years with this whole AI revolution?Sharon Gai (45:22)I would, instead of looking at where to spend money first, I think I would look back on our business metrics and what number do we want to shoot up more or tweak? Do we want to build our awareness? I think for Macy’s probably not because everyone’s already very aware of this retailer. Do we want our revenues, revenue numbers to go higher? Do we want to increase the number of locations?Do we want to increase profit? I think I would zone in into an actual business metric. And then from there, cut into what is our bottleneck? So what part of the way that our current team is oriented, are they spending the most time on? Per role, what part of your job should be? Per role, if you were to do a task organization exercise whereThere are, an e-commerce manager for instance, in your day-to-day, there’s definitely things that you do repeatedly because every time you upload a new SKU or you’re onboarding a new SKU, there’s definitely repeated things there. And there’s also things that are never repetitive where, know, if for instance, your SKU was just magically picked up by...Grace Shao (46:27)Mm-hmm.Sharon Gai (46:37)a Sabrina carpenter and now it’s, you know, floating all over the internet. That’s like a one of a kind thing that AI is not really good with in handling and capitalizing on. And so if we figure out the bottleneck and the repetitive pieces of your job to then find the tools that will take away that repetition.Grace Shao (46:57)Mm-hmm.Sharon Gai (47:02)to free you, to give that time back to you, and to free you in whatever else you can do to make the business better. I think that’s where, that’s sort of the stepwise sequence of things that I would do. And I think that in terms of a dollar figure standpoint, honestly, a lot of these AI companies don’t really know how to best price their products.Sharon Gai (47:24)The whole industry is just in a huge moment of experimentation right now, where a lot of people don’t know whether to price it by the more traditional SaaS models, or should we do this in a... A lot of companies are going for a consultancy, like a consulting services play, because Palantir popularized that, ⁓ or like a per time, per query.Sharon Gai (47:51)way a usage number and I get a lot of emails from partner vendors from their partner teams and every quarter or so they’re like, we we changed numbers again. This from now on, our pricing is gonna be this for this many queries that come in. So a lot of things are in flux.But I think the important things that shouldn’t be in flux or that would stay is that those foundational three things.Grace Shao (48:20)That’s a very actually thoughtful answer. I was just looking for a headline. But no, I think that that’s really, really fair assessment. There’s just such an early stage in the whole industry right now that these tools don’t even know how to price themselves. But eventually they will figure out their own business models. And then I think for the users, whether these retail companies that you advise or not, they will figure out how to price that into their business.Okay, mindful of time, I do want to ask, what do think is the most overhyped idea in retail tech right now? This could be AI related or not, but just in general, where do you think people are putting a lot of attention in and whether it’s overhyped or not? What do you think is an interesting retail tech that people might not know about? Actually, that’s a better framing of the question.Sharon Gai (49:02)something they might not know about is, well, I have mixed feelings about this, but in 2017, when I, before I joined Alibaba, well, it was right around the year. They actually set up this traveling road show called, Gateway. If, if you’ve heard of that, they did it first in Michigan. So I’ve ever re used to really want to bounce.Grace Shao (49:19)No, I’ve never heard of it.Sharon Gai (49:24)bond themselves with the Republican Party and go to all these Republican states and say China is a huge consumption power. We’re going to import so much meat and a lot of chicken and a lot of soy and grain. And so they had this traveling conference called Gateway and it would go through different states. And I remember there, there were these Alibaba developers that would make everyonedo this demo of putting on a VR headset and looking at Taobao, like three dimensional. So you can browse through Taobao and through the headset. I think that is something that people might not know about. It’s not so much hyped, I guess. It’s definitely something in the future, but it’s definitely not something here in the next three to five years. I think what we have.Grace Shao (50:13)You think like wearables, like wearables for shopping.Sharon Gai (50:16)Yeah, wearables for shopping or anything where it’s like a digital 3d store. Like all of the things that we talked about in the metaverse, like those are very, very far away from us. Yeah.Grace Shao (50:31)Yeah, I feel like those like VRsets just like didn’t really take off. They had like a year of hype and then the technology wasn’t good enough. But I do see your point. Like even for myself, like if I could just wear something, I’m sound so lazy, but I can try out outfits and then like, you know, click order. That would be really helpful because right now it’s like I shop on Net-A-Porter and it comes to me that I have to like send it back, right? If it doesn’t fit, but that technology would really change how to shop.Sharon Gai (50:57)I yeah. Amazon has that, where you can change your outfit. And then you can also.Grace Shao (51:02)Mm-hmm, but it’s not feel likeas real yet, right? Like it doesn’t fit my body, my changing body and you know, like just like, it’s just like a fixed avatar, right?Sharon Gai (51:13)Yeah, it’s two dimensional, you, but you, it’s hard to do that too with a VR headset.Grace Shao (51:15)Mm-hmm.I see what you Okay.Sharon Gai (51:20)Yeah, you would need some sort of mirror to do like that sort of, that was, that was a tech that was big too and then also died, which is like in-store interactive displays where guests implemented it, where you don’t have time to try on all those outfits. So they uploaded every single.Grace Shao (51:36)Yeah.Sharon Gai (51:43)item into this mirror and you can just plop plop the mirror and you’re standing in front of the mirror so you can see, so you can check out your outfit without actually physically putting it on. ⁓Grace Shao (51:55)Yeah, Lululemon had the fitnessones, right? The fitness mirrors. then they kind of, it was the same technology behind it, but it also kind of died after COVID. Although I think a few Chinese companies are trying to make it happen again. They’re saying it’s AI empowered. But right now it doesn’t feel like much more than a big screen where you get to see reflection kind of just following the structures and instructions and that’s it, right? It’s not huge breakthrough yet. Yeah. I have one last question for you. Cause we were kind of going off on ourGrace Shao (52:22)tangent again, the two of us, they first started with cows and now it’s like the mirrors. What is one unique belief or differentiated view you may hold about industry that you think is quite different from others in the industry?Sharon Gai (52:35)on the note of retail, I think.Grace Shao (52:37)on the note of retail and or retail and technology.Sharon Gai (52:40)I think retail is pretty simple. I think people complicate it too much with all these bells and whistles. At the end of the day, retail is how do you sell something and make the other person interested? That’s all. How do you take any product that you have, you can put any sort of shell over it, and someone’s eyes light up.That’s what retail is. I think all the AI that we add, the VR things, the personalization and this or the other, those are all ways to do it. But I think it’ll just change. So I think...I think not enough retailers go to the heart of what they’re selling or what they’re doing. And they often, it’s also understandable because, know, it’s most like in any Pyramidic industry, you always look at what the top key account or top brand is doing and they’re usually producing the most GNV. So you say, ⁓ they’re selling a lot. I’m going to do what they do.And then everyone kind of swarms in that direction. And it’s not that, I think that’s for any industry, that’s just how I guess the world works. But I think the right way to approach it isn’t to look at that model. I think it’s more just to answer the question of what are you selling and how are you gonna sell in a way that attracts attention.Grace Shao (54:06)I think that’s like pretty applicable to like every industry. Like when you’re saying that, like really I was just thinking about like the business of media content or anything, right? Like fundamentally go back to like what, what is the problem you’re trying to solve? What answer are you trying to provide? What are you solving? Right? Like, and then, you know, go find your core versus all the kind of frivolous things outside of it.Those are just tools. anyway, really, really appreciate your time. And I actually really love that you just walked us through the evolution of technology and retail, the intersection of technology and retail from, you know, what we know of e-commerce marketplace to live streaming to, you know, AI adoption. And we kind of even touched on our VR.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe
• The Visible Hand: China's Strategic Economic Planning with EIU Chim Lee 30.10.2025 47min

  Joining me today is Chim Lee, Senior Analyst at the Economist Intelligence Unit. He works in EIU’s China and Asia teams, and is based in the company’s Beijing and Hong Kong offices.He leads EIU’s research on China’s advanced technologies, Climate change, Energy, Semiconductors, and Artificial intelligence, and also covers how China’s industrial policies link up with the broader diplomatic and macroeconomic dynamics.Our conversation starts with China’s newly announced 15th Five-Year Plan proposal, which reveals the country's next priority and how it may impact its economy, society, and trade relations with the rest of the world.We then dove into the current involution 内卷 issue, particularly zooming in on the solar and EV sectors. Then we look at the data center build-out driven by the AI boom and how local and regional governments are making sure involution does not hamper this sector.Finally, Chim reflects on his work and his analysis of China’s economic planning and innovation direction.--In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.They delve into the strategic priorities outlined in the plan, including self-reliance in technology, maintaining manufacturing dominance, and the role of private and public sectors in driving economic growth. The conversation also touches on the challenges of overcapacity and the evolving landscape of China’s international cooperation.--* China’s Five-Year Plan signals strategic priorities.* Focus on self-reliance in technology and innovation.* Maintaining manufacturing dominance is crucial.* Private sector plays a key role in economic growth.* Overcapacity remains a challenge in various sectors.* International cooperation is evolving in China’s strategy.* AI and new energy are critical emerging industries.* China’s economic planning involves both public and private sectors.* The plan addresses geopolitical tensions and trade flows.* China’s approach to technology is both strategic and adaptive. Get full access to AI Proem at aiproem.substack.com/subscribe
• AI Plus: Understanding the Intersection of AI and Economic Growth 13.10.2025 48min

  In this conversation, I spoke with Tom Nunlist from policy consultancy Trivium, about China’s AI Plus plan and its implications for the economy and society. We discussed the role of digital infrastructure in AI adoption, the transformation of production relations, demographic challenges, and the government’s role in connecting academia and industry. The conversation also covers the complexities of navigating China’s regulatory landscape, municipal and provincial implementations of AI policies, and the measurement of AI’s economic impact. Tom shares insights on how MNCs can better align corporate strategies with government objectives during the AI growth era, and talks about the emerging AI pilot zones and how China balances between innovation and regulation. Tom Nunlist is the Associate Director of Tech and Data Policy at Trivium, a leading China policy research consultancy. Tom’s work explores the intersection of politics and technology, with a focus on data and artificial intelligence. His hands-on consulting work with Fortune 100 clients covers policy analysis, risk assessment, government relations, and communications.In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.For more information on the podcast series, see here.Chapters04:27 Understanding China’s AI Plus Plan10:55 Transforming Production and Society with AI15:54 Government’s Role in AI Development24:59 Measuring AI’s Economic Impact27:12 Local Adaptation in Policy Implementation28:01 Understanding Chinese Policymaking for MNCs28:59 Aligning Corporate Goals with Government Objectives31:19 AI Pilot Zones and Innovation Hubs33:26 Promising Use Cases for AI Adoption35:56 Balancing Innovation and Regulation in AI42:52 Shifts in Government Priorities for Technology45:56 Tracking Real AI Diffusion in the Economy48:57 The Skills Gap Created by AIAI Generated TranscriptGrace Shao (00:00)Joining me today is Tom Nunlist, Associate Director of Tech and Data Policy at Trivium, a leading China policy research consultancy. Tom’s work explores intersection of politics and technology with a focus on data and artificial intelligence. His hands-on consulting work with Fortune 100, clients, covers, policy analysis, risk assessment, government relations, and communications. Tom, it’s so great to have you here and it was lovely meeting you online actually a couple weeks ago at one of the panels we were on together.So today, will unpack China’s AI Plus plan, what it means for the real economy and how AI governance is viewed by China, sorry, viewed in China and compare that to what’s happening really in the US. But to start, tell us about Trivium and what’s your own professional journey. How did you kind of end up in Shanghai?Tom Nunlist (00:45)Cool, thanks, Grace. That was a nice introduction and likewise, good to meet you and good to be here on your podcast. Definitely flattering. I’ve seen your upcoming guest list and lots of exciting personalities coming up to be on the podcast. So yeah, a little bit about Trivium. We were founded ⁓ in, I think, 2017. So we’ve been around about eight years now. We are a China-focused, or right now China-only policy consultancy.⁓ And so we really our kind of like value ⁓ is that we really know how the sausage is made here in terms of policy and politics in China and we help our clients mostly multinationals and investor clients understand that. So, you know, for example, a new policy like AI plus, you know, comes out, you know, we can come in very quickly and, you know, help inform our clients, you know, what this is, where it comes from, its overall context.and then forming scenarios for how it’s gonna play out and kind of what they might wanna do. As you mentioned, I’m on our tech team, but we cover a lot more ⁓ than tech, really kind of the whole nine yards of ⁓ policy making, be it from economics to labor to kind of everything in between.Tom Nunlist (02:01)Yeah, as for for myself, I think I have a pretty typical China story, you know, insofar as, you know, long time expats. I came here in 2008, you know, more or less on a lark as a study abroad students, you know, to figure it all out and, know,then life happened, got interested in it. I moved here permanently in 2013. My undergrad background is in journalism. So I studied here for a bit. Then I worked at a business review magazine and then eventually kind of made my way into the consulting space. Not too much of a very strict career plan, but again, one thing sort of leading to the next and here we are.Grace Shao (02:43)Awesome. So I think we’re going to go straight into it. What everyone’s interested right now is in the AI Plus plan. So that rolled out in August this year, believe, late August. It’s quite new. think people are still trying to understand what it really means. So the Chinese State Council published a high level paper that was basically pushing all sectors to really embrace AI.It’s said to be the most comprehensive blueprint for AI development domestically, and even touches on China’s international ambitions or diplomacy as well. So to start off with, can you just tell us, high level, what is this really about? How do we understand this policy?Tom Nunlist (03:24)So this is the second high level AI policy to come out of China. The first was some years ago already back in 2017. That was about, you know, it was about the new generation of AI. AI Plus, the concept has been around for two years already now. It was originally announced at the two sessions. Hopefully our listeners know what that is. an annual meeting like that sets policy. It was announced two years ago, talked about again this year, but you know, not manydetails were revealed about it. There was some assumptions which they were correct that it would be a bit like a former policy from 2015 called Internet Plus, was kind of following in that same vein. And just to sort of like set the stage of like what a document like this is, so it’s called a State Council opinion, you know we’re referring to it as a plan, it’s called Plan in the name, but it’s really a sort of like high levellike political document that is setting the priorities for what the nation wants to do, right? Like here’s the direction we’re gonna move in. Here are some like broad, you know, over the horizon kind of KPIs. This is where we all wanna go. ⁓ And then, you know, the plan will cascade kind of down from there. We’ll get more details over time. In terms of the name, AI Plus, that’s AI Plus or added to everything.So everything in the economy, in all sectors, in society, we wanna see, the government wants to see AI make its way into there to get the most use out of it, to really ultimately transform the way that the economy and society works. It’s big, it’s a big vision is the answer.Grace Shao (05:04)think it’s really interesting you mentioned Internet Plus because I remember when that came out. So you said roughly 10 years ago, It was really embraced by every part of the economy and society. So have you seen any attitude changes or shifts or how people view AI Plus just being on the ground in China right now?Tom Nunlist (05:22)Well, in terms of views on the ground, like in terms of people talking about it publicly, not a whole bunch. I think what’s really happened here is these both internet plus and AI plus are responding to things that are already happening. So the internet certainly did not arrive in China in 2015. was ⁓ like.Grace Shao (05:42)Yeah.Tom Nunlist (05:43)very, very much going strong and was already one of the world’s leading digital economies at that time. And so what it was really seeking to do is kind of take the momentum for things that were already happening and push that further. So obviously, in 2015,you know, the consumer internet, know, Alibaba, Taobao, like those type of, certainly WeChat Pay was introduced in 2013. You know, these were making waves, making big changes in the way that society just kind of basically works. And then Internet Plus was like, yeah, let’s take this momentum and apply it to everything else. Let’s have Internet Plus healthcare. How can we use the Internet there? You how can we use it in government services? And again, AI Plus is sort of doing the same thing. You know, the, you know, this is an introducing the AI wave to China, the AI wave is here, it’s happening, everyone’s using it, everyone’s excited. And so this is getting behind that momentum that is naturally already here and attempt to build a policy framework around it and like, yeah, really, where are we going with all this momentum, right? What are we aiming to achieve?Grace Shao (06:52)Yeah, I think actually on that, I am curious, has China’s highly digitalized society and the infrastructure made AI implementation or diffusion any easier in your eyes? Or how has that kind digital infrastructure played a part in just the mass consumer adoption of AI we’re seeing right now in China?Tom Nunlist (07:05)Mm-hmm.Yeah, it’s a huge part. mean, from just the consumer side, China is like the US or Europe, just an extremely connected society. Everyone, even in the most remote places in the country, has their smartphone, has probably even has like 5G.WeChat is something of a national infrastructure at this point. It’s a messaging app that everyone uses for work and life. It really is absolutely indispensable. And so having that infrastructure already there, or having everybody with a phone in their pocket automatically makes these tools accessible. I think...any age, any person you might come across, you know, do they have DeepSeek on their phone? Chances are, yes, they probably do. On the back end, which I think is, you know, just as ⁓ important. So the past few years, or as we all know now, right, the...know, the biggest part of the biggest spend of the AI boom is building out, you know, massive, massive data centers, right? And making that kind of infrastructure work. It’s a huge race right now in the United States. And so there was already a national plan to have ⁓ a nationwide network of data centers, you know, put in place as kind of before this big AI wave.It’s a bit to do with some broader reasons of internet and energy and having some of this infrastructure in place. ⁓ Actually, in terms of energy, that’s one of the ways that I think a big leg up that China has in terms of the US is the amount of energy infrastructure it has built out compared to other parts of the world. So they’re ready to sort of do this, where I think other places maybe a little bit less so.Grace Shao (08:57)Yeah, and I think we can kind of talk a little bit more about the East data, West compute and all these different government initiatives that’s really boosted the data center built out. like, to your point, priority even the AI boom. But actually, let’s take a step back and kind of look at how the policies really affect a society, right? I think in September, one of your colleagues Kendra, her shape for a road, a blog post saying,The state council’s AI Plus directive is to reshape the paradigm of human production and life. When I read that, was like, what does this mean? It seems kind of crazy. Like, are we going to make AI babies now? What does it mean to promote a revolutionary leap in productivity and profound changes in production relations and accelerate the formation of a new intelligent economy? So how do we kind of like break this down? What does it mean? And, you know, when we were chatting prior to this podcast recording, you said this is a grandiose term.It’s three-shaping human production, but what does that actually mean? Like, is this quite literally reproduction? Like, how do we understand that?Tom Nunlist (09:54)Well, I don’t know, hopefully reproduction will stay traditional. in terms of, know, these types of policies sometimes have these like really big grandiose framing. know, again, back to what I said earlier, the point of this, it’s a political document at the end of the day. It’s establishing a vision, right?and the promise of AI, which is not... ⁓news to a Western audience is that it will be transformative of society that will kind of like change sort of how things work. This specific language that they’re using there, like talking about transforming production, you know, that’s a bit in the sort of like communist Marxist language, you know, of China. And then the context that it’s kind of living in now as well, ⁓ there’s this like really big deep sense of urgency in China of like kind of likethe need to move from ⁓ an economic model that is waning, that sort of reliance on labor intensive ⁓ economy and land sales and things like that into a ⁓ new area where they’re get new types of growth, new and better growth, the switch from quantity or quantity.⁓ to quality. These goals were kind of already there. There’s another, you know, wonky Chinese policy term called new quality factors, new quality production factors, and what are all of these, you know, types of new things, you know, ⁓ AI, self-driving cars, and so on. And...it wants to leverage these into making new growth opportunities happen, basically.Grace Shao (11:33)Yeah, and I think you touched on one thing, which is like, you know, the traditional economy is very like labor heavy and it really relied on just the mass population of the mass workforce, right? But what we’re seeing right now in China, but not only China, a lot of like actually very developed economies across Asia, including Japan and South Korea right now.is that there are just not enough people. Like the population is declining, people are not willing to have children, right? And kind of given that backdrop of an aging population, a shrinking population, what is kind of, I guess, the goal from the government when we look at the labor force? And how will AI and technology play a role in that? Are we really just going to see like robots implemented or is it more automation? Or how do we understand this? Yeah.Tom Nunlist (12:22)think in some cases, we will see robots replacing physical workers. ⁓ But I think that’s the smaller part of the story. The bigger part of the story is this broader question of actually just avoiding the middle income trap. And so in order for China to take care of its aging population, to sort of weather this big demographic shift that is happening now,no matter what, even if birth rates double or triple next year, it’s gonna happen. And the way to do that, that the government sees is to raise people’s incomes per capita.and do that very quickly over the next 10 years, More profitable companies have more prosperous people, have a bigger tax base. And so that the country is just able to deal with this challenge as it emerges. Again, it’s inevitable. And so back to this new quality production factors or this transformative effect that AI is gonna have, the transformative effect is that it will be a productivity multiplier, right?enabled everyone and companies from big to small to be vastly more productive and vastly more valuable and really help China earn enough and become wealthy enough, maybe right before it ages too much. So I think a bit indirect there, but it’s about the whole economic story together.Grace Shao (13:51)Yeah, and I think the whole approach to AI development and progress has been extremely pragmatic and economic driven for China, which is a bit different from what I get the sense in DC and even for sure Silicon Valley. Actually, on the topic of what you just mentioned, which is the government’s role in promoting companies and companies’ profitability, I have heard of this thing where the government is playing a role becoming a networker between the academics.Tom Nunlist (14:00)Yeah.Grace Shao (14:19)and the researchers and the companies. And I think for the audience, a lot of people sitting in West, we know about Alibaba Tense and Huawei, these mega big tech companies having talent schemes, quite similar to how basically there’s campus recruitment for like Meta and Google, whatnot, right? But how is the government now playing a role for SMEs or even smaller companies in terms of how are they connecting talent and...⁓ policy people and kind of the resources in the public space and the private space.Tom Nunlist (14:51)This is a great question, think, and a really important thing that’s emerged just over the last couple years. It’s not just AI, it’s sort of like all areas of science, technology, and engineering. But what it seeks to do is to bridge the corporate world and the academic and research world in a better way, right? So you can have like...⁓ needs and talents and coms flowing both ways. So for example, this might be setting up round tables or some kind of like platform or any kind of mechanism that brings these parties together. So going in one direction from corporates, setting up links with universities so they can go into departments and say, hey, we’re doing biosciences, we’re doing AI, we’re doing you know, some type of metallurgy, you know, these are the types of talents we need. And can you focus on that? Can you help get us, you know, train that talent that we need?or going the other way, having researchers see what’s going on in the corporate world and having a solution for that, or green fielding their own research, right? They’ve been doing this for a state institution or a university, and now they wanna take it out of there and find the right entrepreneurial partners to do that with.Right. You know, as you mentioned, know, like large companies have done this sort of thing for for a very long time and have prospered, you know, because of those links. mean, indeed, I mean, a lot of the American tech giants, you know, came. That’s a famous story. It came out of a university or dropped out of a university, you know, and now, you know, maintain those links. It’s same in China, but, know, that’s a lot harder to do if you’re an SME. I mean, everything’s harder to do if you’re an SME because you don’t have the resources. Right. So providing that meeting place,facilitating that is I think a really important program and one that I’m pretty confident will see solid results in the next couple of years.Grace Shao (16:49)So what agency or what government entity is actually helping facilitate that kind of meeting right now? And this question is to lead to the next question, which I’m going to actually ask now, which is, for me, I’m not a policy person. I’m getting confused when I read these papers, right? Because there’s the NDRC, which is in charge of the economic planning.Then MIIT, which is in charge of the information technology, the ministry, the CAC, which is a cybersecurity regulator, right? Then there’s the MOST, and then there’s a party central science technology commission. There’s just so many of these government agencies that seems to be all involved in pushing the progress of AI technology at this point, as well as being a regulator for safety and policy work, right?Could you kind of just break that down? Who’s in charge of what?Tom Nunlist (17:39)All right, okay, let’s just address that one, because that’s like a pretty big question. So for this, for the AI Plus plan in particular, the main administrative body for this is the NDRC. So for those of you don’t know, that is the macroeconomic planner. They’re in charge of kind of like setting the big direction.of the ⁓ economy, right? So NDRC is in the coordinating role of this plan, right? So from there, it’ll go to the other ministries of the state council, some of whom you’ve just mentioned, right? So the industrial ministry, that’s MIT, science and technology ministry, most. ⁓⁓ CAC, the cyberspace administration, all across the board. And those ministries will be in charge of taking the big idea and making it specialized or setting specific goals for their various sectors. And we’ve already seen that happen, actually. just a couple of weeks ago, the National Energy Administration, the NEA, came out with the very first ministerial AI Plus plan, which is AI Plus Energy.And we’ll skip most of the details there, but suffice it to say, it is gonna use AI to help make the energy transition happen, which is very cool. From there, it will cascade down further into localities. And localities is really, that’s where the rubber meets the road and where all of the action happens. So we’ll see cities, they’re already AI plus plans.There’s one in Beijing and Suzhou, those are explicit. And then like Shanghai has one basically, although it’s not called AI Plus, but they have one as well. Interestingly enough, those also actually predate the national plan, which is something that kind of happens in China at various points. And so a lot of the like actual like funding decisions and a lot of where the funding comes from will be at the local level.And then there’ll be like, you know, a national pool of money as well that will like help support those, right? So, you know, it’s a top, you know, people say, China’s a top-down system. That’s of course true. And what I just described is how that top-down system works, right? So from the central planner down to ministry needs, down to local level, which has all of those ministries at the local level and, kind of being funded ⁓ from there. And then of course there’s like special national projects here and there.Grace Shao (19:59)So I’m just trying to understand this. In the RSC, the economic planner basically makes a big grand plan and they push out the AI Plus that we’ve been talking about that was pushed out in August. But a lot of the execution that’s done is actually trickled down into localities like the local governments, the provincial governments, city governments, whatnot. And so something like, just taking this as an example, something like the facilitation of maybe a researcher at Tsinghua meeting private company for potential, let’s say commercialization plan, that could be actually led by say the Beijing Education Department or how does that, I just wanna understand how to execute that works. Okay.Tom Nunlist (20:36)Yeah, yes, yes, yes.Those might exist at different levels, I’m not sure. But yeah, the local level would certainly be implementing stuff like that. Or in another more sort of ⁓ direct way, Shanghai has money now where it can like say, companies that are in AI space are eligible for X amount of money.funding for their first year, right? And like that funding decision, that’s made at the local level at Shanghai.Grace Shao (21:03)And that will be decided, I guess, by what the city might mean. So each city, each province, given their strong, they have their own economic factors, right? Like for example, like I think I was researching Harbin, like, you know, people think it’s just like a really cold place for the ice festival, but actually it’s an industrial city with a lot of legacy in robotics, traditional robotics, mechanics, industrial machinery. So their money might be put into developing physical AI.Tom Nunlist (21:12)Yes.Grace Shao (21:29)like embodied AI, right? And then maybe in Shanghai, we’re thinking about like maybe consumer driven products, right? Like just, just kind of high level thinking, but that that that’s kind of what happens. ⁓ So I want to understand how does the KPI work then, like in terms of like, how do we understand, I guess, how these, because what I’ve heard also is like these cities to cities, compete with each other, they compete for talent, they compete for, like, bringing in different businesses, how does that work? And then in terms of like, how do they actuallyGrace Shao (21:59)a measure, right? Like the technology or AI’s contribution. Because we talk a lot about, like people talk a lot about like how companies are trying to measure AI’s like actually, ⁓ you know, contributions to the company right now, the profitability. How do we actually understand AI’s contributions to the economy? I guess it’s two separate questions, but yeah, help me understand that.Tom Nunlist (22:19)Yeah, this is a really interesting question. And I think frankly, it’s one that the government is just trying to figure out itself. For years, of course, it was just GDP. So you win if you bring your area GDP, which is great for encouraging growth until it encourages the wrong kind of growth or encourages the wrong kinds of projects. And so I’m a little bit less familiar off the top of my head, but it’s something my colleagues have looked into. ⁓as well is how these KPIs might be changing. And again, from this shift from quantity to quality, I think at the end of the day, probably something like GDP is simply the easiest thing to sort of see. But certainly, and that’s like if you’re like a mayor, I guess. But for people that maybe work within different ministries or in like...specialist areas, whether or not they do a cool project along these lines, whether or not they brought, they fostered the emergence of a new giant in their district. That’ll be looked on favorably. So in terms of who actually sets these KPIs, I think that would actually go down to the personnel department ⁓ and how they interact, how the personnel department decidesthings to include on there, some of which will be from NDRC’s AI plan and some of which will be from like totally other different things. I can’t tell you what their score rubric looks like. But again, the message here of going this like broad top-down kind of thing, what officials will be doing is, youlooking at the communication of these targets, right? Looking at the messaging and interpreting them for their district, right? So what do I need to do to make that happen here? And that’s the way forward for my career, right? And also to connect this with what you were just talking about in terms of local specialization, right?what’s going on in Harbin, the local conditions there are different from in Shanghai or in Hangzhou. And so I think in the ideal way, and the government uses this phrasing a lot, is to have things definitely specially adapted for your local conditions, right? Don’t just do exactly what we’re saying, like make it work for you.Right. And so in the ideal world, you would have like different things going on everywhere and they would all be complimentary. I think what happens, what tends to happen is that you have duplicative efforts, you know, which of course we see, you know, everyone’s talking about now in the auto industry. my gosh, there’s a hundred auto companies and they’re all, you know, in a giant battle Royale that is destroying value, you know, rather than. Yeah.Grace Shao (25:06)Yeah, the price war right now. Yeah.Actually, how do we understand this? think because for the sake of, know, understanding Chinese policymaking for say, Western investors or Western companies, like say, MNCs operating in China, and in the day is to help thembetter their operations, right? So then how do we understand this from that perspective? Say your client’s M &C and they’re saying, seeing, okay, AI plus is being rolled out on a central level. Then they are like, how do they decide? I don’t know where to put their plan, to build out their operations. How do they kind of make that judgment comparing provinces to provinces? And I think to your point, you kind of have...answer this in the sense of like maybe if you’re industrial machinery you go to Harbin right but if you’re consumer goods you’re Shanghai but are there any other things that companies need to be aware of or investors investing in companies are coming out of these different problems should be aware of?Tom Nunlist (26:00)Yeah, great question. So I think probably the first choice, the first thing to look at is just, you know, where are the hubs for what you’re doing, right? If you’re an automotive company and you’re looking to make any of these, well, might go to, you might go to Enhui, right? Hefei, sorry, I forgot it for second. You might go to Hefei because that’s where a lot of the new energy vehicles are.Right, and then from there, and this I think is a bit more unique ⁓ to China, is if you’re a corporate and you’re trying to be successful here, one of the first things you need to do is align with whatever the government is trying to do. You know, that doesn’t mean do exactly what the government asks you, right? But you know, figure out what officials there want, what their KPIs are, what their existing programs are, and how do you align your corporate goals with that?⁓ And that’s how you get support. That’s how you get buy-in. That’s how you’re ultimately successful, right? You know, as in sure it’s no secret to anyone, you know, the Chinese government just has a much bigger voice in the direction that the economy is going, right? And the things that are happening in the economy and, you know, companies and investors absolutely, you know, have to listen to what that voice is saying.I think for investors as well. So where are these companies collected? Where are the big hubs for the industry that we’re investing in? And also, what is the government itself saying that it wants? And which companies do we think can...Obviously, of course, first deliver on the market promise, like do what they’re saying you’re trying to do, but are there opportunities here? Will they get this kind of support from the government that is a factor that is larger here than it is in other places? Probably maybe any other place.Grace Shao (27:46)Yeah, I think it’s also like the point you’re saying, it’s not really like you have to do what the government says, but it’s like you might as well lean into, like, I guess, lean into it, right? Like there are going to be favorable policies for your industry, certain areas, municipal areas, you might as well lean into it to optimize or to like maximize your success rate or your success possibility, right? So on thatGrace Shao (28:10)point actually, I’ve heard that there are quite a few AI pilot zones. Like, you know, right now, I think for the West, people only know about Shenzhen, Hangzhou being kind of the tech innovation centers, obviously Beijing, Shanghai playing a big role for corporate headquarters and obviously where investors sit, policymakers sit. What are some other major cities that are actually quite relevant to this like AI growth right now or are considered AI pilot zones?Tom Nunlist (28:35)think those would honestly be the main ones. know, Shanghai, Beijing, you said, Shenzhen, Guangzhou, Hangzhou, like these are the places where, you know, a lot like the most action is happening, right? Especially in an area where we’re talking about, I mean, it depends on what we mean, right? So like if we’re talking about just raw AI development, making new LLMs and stuff like that, you know, one of the big, you know, stories is that there’s only so much talent out there that can do that.⁓ and this talent will gravitate towards some center. And there’s only a few of those, only, not everyone can have those people. Not everyone, those people won’t go everywhere.Right, AI, but back to what AI Plus is about, right? AI Plus, all of these other things, right? And having that in various sectors, I think where other cities will excel or have the opportunity to excel is where those hubs are, right? So if we’re trying to add AI into auto manufacturing, that’s gonna happen in an auto manufacturing hub.Right. And I think that actually speaks to the important thing that folks need to be looking out for. You know, at this point, know, we’re, you of course, at the high level, you know, we’re talking about sectors. OK, we AI in the research sector or want AI in the health care sector. But I think what’s most important is going to be looking out for not which sectors it revolutionizes, but which specific use cases, right, are going to be.Grace Shao (29:59)Mm-mm, I see.Tom Nunlist (30:06)most obvious to implement.Grace Shao (30:08)And actually on that point, which use cases, let’s put it that way instead of sectors, do you think are kind of showing the most promising mass consumer adoption of AI, gen AI as we know it? So I’m not talking about like the buildup of LLMs and everything. I’m saying, know, when DPC came out, there was a media frenzy of stories about how China’s like home appliances are even adopting AI, EVs are trying to adopt AI, you know.Tom Nunlist (30:13)Yeah.Yeah.Grace Shao (30:34)I mean, obviously that kind of hype has gone, like, moved past us, but like, in terms of whether you want to use sectors or use cases, where do you see actually China right now really leading in adoption? And where do you think we’re seeing the trend going towards maybe in the next three to five years?Tom Nunlist (30:50)Yeah, think ⁓ it will continue to penetrate more ⁓ on the consumer side, just on of like AI services that are available to everyone. mean, that’s sort of the biggest thing right now. Whether or not we can get consumers to pay in China, I think is a little bit different of a question. But in terms of specific areas, I think it’ll be where we’ve already seen AI ⁓ have quite a bit of traction. So in like logistics and transportation where, you know,with like self-driving is kind of almost here and you know we have the the nev is this it’s a software-defined vehicle and we’re going to be like a ready integration for ai into the features of the vehicle that’ll definitely be one you know another one thinking about ⁓ that comes to mind is is agriculture which i you know ⁓ i can’t name a specific company or or a project but ⁓you know, drones are becoming ⁓ large and, you know, helping to manage big farms, like do things like crop spraying, you know, or inspecting or like, also not just in agriculture, in inspecting power lines, drones are not used to do that. It’s actually physically hard to get up there, right? And so there’s AI use cases for that, right? It can go into like visually inspecting, right? Or visually help, you know, irrigate your crops and so on and so forth.So it’ll be things like that, right? Where we’ve already started to see new things happen, AI being used a bit. And now these new tools and the growing power of these tools will enable it to really actually happen.Grace Shao (32:28)Yeah, definitely. think like, when I first saw and tried out a few of the EV cars, this is even like during COVID, this is like three, four years ago, I was shocked by I wouldn’t say they’re like genuine power, but how tech savvy they already are. had voice control, each of them already had a built in robot, you can control your like windows, you control your heat, like the heat of your seats by voice recognition, voice control. And I think like you said,Tom Nunlist (32:42)yeah.Grace Shao (32:52)implementing GEN.AI into it just means that it can actually embolden it more, right? Do more things or right. So that’s really interesting. I think I want to double click on one question that a lot of people are kind of debating. know, China’s approach innovation often is said to be, you know, innovate and then regulation comes later. Europe obviously takes another extreme case of like hyper or not hyper, but like a lot more.Tom Nunlist (32:57)Yes.Grace Shao (33:16)cautious and safety, you know, safety cautious and like, you know, regulation comes first. And some people are complaining about how it’s hindering innovation or innovation going into production. Right. So I guess my question right now is you’ve been in the AI safety and policy space for a long time now in China. Do you think that actually you must give up safety for innovation or are there other ways that you’re seeing people actually being able to have safety andinnovation co-exist and co-develop and maybe taking data privacy as an example or how did Deepsea come through if there was so or let’s just talk about that space.Tom Nunlist (33:50)youYeah, this is an excellent question. And frankly, I think one of the most underappreciated or even like misunderstood aspects of the AI story as it stands right now in China. I mean, there was a point not too long ago before the EU AI Act, which you mentioned where, you know, China had, you know, the strictest AI regulations on the books in the world. And yet, you know, DeepSeq was still clearly able to emerge here and,you know, become what it is, right? And I think the story here is that, you know, China is, I think, as most people will understand, a very security conscious, you know, country, but it is also highly flexible, right? And the interesting thing, the sort of interesting story, like when ChatTPT first came out, there was this mad scramble.among regulators to get a handle on it, right? Because it was gonna flood the internet, you know, with these tools and man, what are the impacts gonna be, you know, like just a real sense of urgency to try to like write something immediately. And so there was a period of about a year and a half where you had regulation after regulation and, you know, they...you know, if you looked at them in line, you’d think they were different, but they were actually kind of rewriting one another, and it was like all very ⁓ messy and a very confusing space. But then, you know, China was able to kind of like find what its bottom line is.and then be flexible and adapt from there, right? So it was, you know, hurry up, let’s do something. Let’s kind of see what’s gonna work, where it might be too far, and then dynamically kind of like dial back.So one of the interesting, I think probably the most interesting single event of this story was there was a registration system that was created where if you want to publicly release an AI tool like a chat TPT, it has to be registered with the state and blah, blah. And then some requirements started to be built on top of that. And there was a draft that said at one point,all of your data that you need to train your LLM with needs to be verified as true. And the AI research community came back and said that this is impossible. Like if this is implemented, will, know, progress will grind to a halt. There is no way we can do this, right? There was no official response to that, but the final version of the rule did not contain that.Right, was that was walked back. was an idea that was tried out, that was an explored, you know, and eventually, you know, was abandoned because it didn’t work. And so I think, you know, one of the sort of like, again, underappreciated or even unknown strengths of China’s regulatory system is that it can be flexible in that way.Grace Shao (36:27)Right.Tom Nunlist (36:46)in an ungenerous interpretation of this, which you hear from a lot of foreign companies and rightly so because there are drawbacks of this, is that regulation can kind of seem all over the place and arbitrary and you never know what things are gonna change next.And certainly in emerging areas, that is true and very challenging, you know, if you’re in a corporate compliance type situation. But the plus of that is it can be, you know, quite flexible and adapt to, you know, what the perceptions of the needs are kind of as they’re coming up, which in, you know, an environment as fast developing as this one, where again, new problems might emerge tomorrow. I think that’s a really important strength or really useful strength to have.Grace Shao (37:29)We could be quite reactive in the sense that they would actually react to what the industry and the actual practitioners at the leading frontier, technology development, want or need, right? To really help and regulate the technology, yet also not hinder any progress. I think that’s really interesting and it’s a very fresh take on it. I haven’t really heard that before, but I think it’sit makes sense. And it also kind of explains what you said about some people’s kind of complained or misunderstanding of this whole like murkiness. so you said that the AI Plus initiative really it’s been around, like not been around, but like the AI policy or the plan has been around or the idea has been around. And then there was a 2017 National General AI Plan as well,There’s also the made in China 2025 plan, all these big grandiose plans that have been really pushing forward AI or robotics and just technology development in general. as you said, policymakers and regulators can actually be quite reactive. So over the last, I guess, 10 years as these three mega plans been rolled out.How have you seen these things change or how has the policy makers really a change in terms of their sentiment or the attitude towards this technology?Tom Nunlist (38:40)Let’s say the biggest thing, so taking a bit of a longer view, so science, tech, and manufacturing development has been a priority of the governments for a very long time, since the late 90s. It’s been kind of on this top priority list. And so one shift I’ve seen in the past few years is side tech development moving from one of the list of important things into the top thing.like the most important thing. It’s like that is ⁓ kind of an organizational principle, right? Or like a driving organizer of the whole party, right? And again, that’s because of the perception of what the state’s needs are at this point. In the past, in the sort of like last formulation, right? Of like what the country needed, right? It just needed growth.It’s like, it’s the late 70s, we’re into the 90s and 2000s. We need to just grow. We need more people and jobs, we need production. That’s what we need. Now that’s not what they need. We need quality growth, we need to move up the value chain, we need to avoid the middle income trap if we can, expand people’s incomes, become a more efficient and a more technologically driven society. And so the sort of prioritization,and some of the character of these plans have changed sort of in line with that. Some other things I think have stayed the same or strengthened rather, right? So with Made in China 2025, which this not really talked about explicitly anymore because of the political sensitivities it creates in the US, right? But the sort of view, right, was that Chinayou know, doesn’t want to be vulnerable, basically to, you know, always reliant on outside technology and wants, you know, these things for its own, right? Wants them to be secure and controllable. It wants to have, have its own thing, right? That of course, I think, you know,in light of the subsequent ⁓ US effort to strangle the semiconductor sector in China is even more of a priority. So it’s not just move the value chain and get incomes up, it’s also create these fundamental technologies which we absolutely cannot have as a vulnerability.Grace Shao (41:03)Essentially kind of push more honed in on the self-reliant focus than they previously didn’t really have to, right? It was also kind of a reaction as well. Okay, I think I want to go in some quick questions. ⁓ You did answer a bit of it, but one overhyped and one underhyped province or city that you think people are not noticing enough outside of China.Tom Nunlist (41:28)Yeah, again, would say,Yeah, they’re not really as specific over under Hype City that I can think of. But yeah, I would say go back, double down on the point of like, you know, look at where different specific hubs are, right? So right now, you know, especially the US talking about AI development sort of in general, right? Like the rush to AGI, you know, so on and so forth, right? The AI plus plan is about doing things in the real world, right? So I think where a lot of like really fascinating stuff is going to happen is where those real worldthings are in China, right? So like where we have many filtering use cases actually emerge and that’s going to be sort of all over the country.Grace Shao (42:02)Right.Right, like CN maybe for renewables, but like Hefei for you say auto, and then like even like Baoding and Hebei for like auto. You get at least like second, third year cities that are just like actually relevant, but only if you’re in the sector, you would know, right? And that’s a really interesting take. So what is one metric that a policy analyst like yourself should be really tracking or focusing more on?instead of just, you know, maybe what we’re seeing on the headline is like, you know, this crazy chase for like benchmark frontier technology, frontier of LLM benchmarking. How do we actually track or judge real AI diffusion in the economy?Tom Nunlist (42:48)would say it’s probably more along the lines of traditional measures, So penetration, productivity, profitability, wage and efficiency growth. Again, the emergence of those scenarios, Are people actually out there using it in the real world? So I think it’s look for those traditional.tangible things, right? Again, I mentioned that Chinese consumers tend to not want to pay for consumer-grade AI tools. If that’s something that changes, right? If they’re good enough where people are willing to pay for it, wow, maybe that would be an enormous indicator.Grace Shao (43:16)Yeah.I don’t think anyone’s gonna want to pay for like, you know, consumer app. The culture, right? Like no one wants to pay. I don’t know, I switched my brain on and off when I use like Western apps versus Chinese apps. And when I’m on a Chinese app, they pop up, they’re like, pay for premium. I don’t want that filter anymore. I don’t need this sticker anymore. I’m like, I’m not paying. You just have a different mentality, right? Because you do get too many goodies for free already. It’s very...Tom Nunlist (43:27)I think he wants to go to pay for a I know.Yeah.Yeah.Grace Shao (43:51)It’s very hard, think. The barrier is very high. The threshold. I have one last question for you. And it’s a question I ask everyone that comes on to differentiate understanding, which is what is an unconventional view you hold? And this could be about work or something in life, you know? But what is something that you think about and you’re like, oh, maybe I don’t say this out loud, or maybe this is quite different from what my peers think?Tom Nunlist (44:13)What was an unconventional view I hold?I’ll go one with topic specific here. That’s because it’s come up recently in ⁓ fights I get into, Twitter fights I get into with people. There is this interesting and I think not totally off the mark concern with AI that it’s gonna basically make us all dumber. Students are gonna outsource learning to AI. There was a case study that did the rounds about doctors using AI tools to help them spotcertain types of cancer got worse at it, know, like after relying on the tool. And that’s a real concern. I think it’s something that, you know, there’s some red flags that seem to say that that might actually be happening. But I think the real problem might be a bit more nuanced than that. think it might, my hypothesis is that it will create something like a ⁓ skills or performance gap.between different parts of the population and exaggerate it. So, whereas some groups of people might become reliant on it and become de-skilled in their job, definitely. And then in some cases, that might be what we wanna happen. I we don’t want everybody, I mean, that’s sort of the promise, not have to do certain boring things. But I think for a smaller portion of the population, it is gonna be a massive learning and development.accelerator, right, to really help you to get good and improve. And so, you know, I mean, beyond, you know, whether or not I’m right, I don’t know I’m right, it’s just a bit of a guess. You know, I’m wondering where that gap kind of might be and how large it’ll be, right? So is it going to be 90 % of people get dumber and 10 % of people become super learners? You know, or is it, you know, somewhere in between?That’s my unconventional view. It’s gonna create a skills disparity.Grace Shao (46:04)Yeah, I actually kind of agree with I think it would make people who are relying on it for skill set like vocational skill almost like just the art of, know, not art, but the skill or ability to write a press release or draft a basic news piece or you know, build a DCF model or you know, do some quick basic research that might become dumber in the sense that you don’t know how todo it in a traditional way. But I think the arguments also like say 34 years ago, people are like, you have the internet now. You don’t even know how to use a library anymore, which I think our generation honestly, I don’t really know how to use the library very well. Like I go on my loss. I don’t know how, you know, how to find books essentially from alphabetical order and, you know, like finding the topics, but we do learn how to find more information in some sense, right?But I think to your point of like, ⁓ it will help people learn a lot faster, but it will require a new kind of skillset, is like, you can access all this information, can you decipher it? Can you dissect it? Can you actually pick out what is correct? What is actually relevant? Because there’s so much noise and clutter, which is kind of similar again to our generation where we had to use the internet to find information, Versus like our parents generation had to like walk into the library and just like.Grace Shao (47:19)go through like 10 books, right? ⁓ But that’s super interesting. Thank you, Tom. Really, really appreciate your time. This was super insightful. It was really helpful for me to even learn about how to understand how policy was made in China, how it might affect businesses and investors. And yeah, this was just super insightful and a lovely conversation.Tom Nunlist (47:21)Yeah. Yeah.Yeah, thank you, really. It was really lovely to be on the pod.AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to AI Proem at aiproem.substack.com/subscribe