Kyle Chan: I know this is sort of like a, a hackney term now these days, but I really believe that the areas that China is really focused on are the building blocks for the next generation of technologies, emerging sectors that are gonna be not only sort of like the economic growth engines, the job engines of the future, but sort of broader progress is gonna be built on that.

You know, AI, robotics, you know, all that, all the stuff that, that we've all heard about. But I want to bring attention to some of those things, too. It's not just today, it's like the trend going forward that people need to pay attention to.

Steve Hsu: Welcome to Manifold. My guest today is Kyle Chan. He is a postdoctoral researcher at Princeton University. He is interested in many of the themes that we discuss on this show, in particular US China technology competition, AI, globalization, multipolarity. Kyle, welcome to the show.

Kyle Chan: Great to be here.

Steve Hsu: Yeah, it's great to have you. Uh, as I was just saying before we started recording, I've read a lot of your stuff. Two things I'd like to go over in some depth is the opinion piece you had in the New York Times recently and also, something on your substack about industrial maximalism and

Kyle Chan: Oh yeah.

Steve Hsu: A theorist at Beijing University named, I think his name is Lu Fang.

And, those are two things I'd like to get into, further on in the podcast. But, the way I like to start my podcast is to learn a little bit more about you as a person. So, I know that you studied economics at the University of Chicago, and then you did your PhD at Princeton in sociology. Maybe tell us about your, the development of your intellectual interests since, from college, to where you are today.

Kyle Chan: Yeah. So one of the longstanding questions that drives my research that actually has gotten me motivated, like for now, I guess almost decades, is, the question of development. Actually, I. So it's funny 'cause now I talk a lot about industrial policy in China, even in the US, Europe, I mean a lot of countries, India.

And so it's like a big topic now, you know, how can the state, or can the state even do something to help. Move along industrial development. But this older question, I think, is still the biggest question. I mean, everyone says this about their field, right? But I think this is the biggest question out there.

You know, why are some countries rich? Why are some countries poor? Why is that you're, if you were born in, I don't know, you know, some, some country in East Africa or in South Asia or in North America, your life is just. So deeply shaped by just that place of birth and country of birth. And yeah, and it's, it's a puzzling question because it transcends, you know, fields economics is where I began, and I found a lot of the tools there useful, and especially a lot of the statistical tools, which are now kind of standards across the social sciences, that's a good thing.

but I did find that the questions economists asked were getting narrow and narrower. so, you know, I'm always up for a good RCT too, to rigorously test a hypothesis, but I felt like a lot of the biggest variables were not even being considered. And some of those are the hardest to, you know, measure, quantify to nail down.

And so I just wanted to, you know, make that shift over to sociology, which, you know, I, I like to think of it as like the big tent discipline in terms of methodology and the types of questions you can ask. And so, yeah, you can study development. By interviewing government officials, at an organization that deals with development.

You can, you can use corporate financial data. You can, you know, do sort of all these different things. You can mix it all together, and just try to present sort of the strongest case you can, for whatever theory you're, you're arguing. So that's why I, I gradually found myself being drawn to sociology and the questions still remain the same.

I mean, now the funny thing is, you know, questions that I was asking about China, which I think is still also the biggest story in development. So, if I think development's the biggest question, China's the biggest case, and it still, I still think to this day, still begging to be explained, you know, how did this country transform itself?

Basically in a, in a generation, in, in a single lifetime, so dramatically and have such huge consequences for the rest of the world that we're, we're, we're sort of dealing with right now. So, yeah, to me, you know, these were such big things and, sociology, I, I started to kind of,you know, do field work.

So I spent several years doing field work actually in both China and India. The comparative thing was really helpful, because honestly, the first time I traveled to China as an adult, I went to Beijing, like, it was like 2005, 2006 or something as an adult. And I was like, this place is kind of chaotic.

It's kind of like all over the place. I remember getting, you know, a bunch of, you know, sort of, hacked together electronics on the cheap, things like that. And that was kind of like my association with China. And then, you know, I traveled to India. They have their own different issues, their own different strengths, you know, it just opened the world to me.

So that comparative lens has also been very important to me. But, but yeah, that's sort of like the, the longer arc, this, this question of, you know, how do countries become rich basically? How do they get all the things that come along with that better healthcare, better, better education, all the things that I think really matter still.

Steve Hsu: If I'm not mistaken, your dissertation work was on development of railways and comparing India to China. Is that correct?

Kyle Chan: That's right. Yeah. Yeah. Yeah. So that was, that was really fun in part because I gotta take a lot of trains in both countries. And, you know, my question was, how did China build this high speed rail system that is now larger than the rest of the world's, you know, bullet train systems combined, in such a short period of time.

And India actually, you know. As late as like the nineties, Indian railways was sort of on par in scale and technical sophistication with a lot of Chinese railways. and, you know, people I talk to who are, like in the international development world who have experienced working in both countries, they tell me that it's not the engineers.

The engineers are extremely talented in India. In fact, in some ways they had, Longer standing access to English-based resources that, you know, sort of global industry standards. So that wasn't the issue. There was something else about why China could build this high-speed rail network and why India struggled to kind of modernize its own, very, very used, very useful and economically important and socially important railway system.

And yet somehow, building a single rail, a single high-speed rail line was a monumental undertaking. I mean, it would be for any country, frankly. So in some ways the real sort of outlier is China's case, but using India sort of as a foil to understand that. And so, yeah, that, that whole project was trying to understand, you know, at a sort of, I call it like the meso level of analysis. Cause I think there's a lot of people who look at what the top leaders are thinking in any country, whether the Prime minister or the president or the, you know, what is Xi Jinping thinking? That's fine. and they're great people who, who can like, read the tea leaves on that. And then there's some people who focus on them.

Really nitty gritty, like bureaucrat level or at the, you know, single company level. And that can be very fascinating too. But I kind of look at the mesa level organizational structures across the bureaucracy. Looking at, you know, at a time, India and China were two of the last countries with ministries of railways.

So looking at how those were organized and why, you know, on the surface they look the same, but when you start to peek under the hood, the whole structure is completely different. I mean, China's system , ironically, was less top down. In a way a lot of the responsibility was sort of, devolved or delegated down to these like project corporations.

And there, basically, I mean, these were tiny teams relative to the size of these projects. I mean, you're talking about sort of like multi-billion dollar high speed rail projects. Basically run by a hundred or less people ultimately. And what that did is that concentrated a lot of that power and responsibility into, you know, again, not the very top, but still a core group that was, you know, living there, breathing this project all the time.

And in the Indian case, it was much more of this sort of typical matrix structure with cross cutting, lines of authority. You know, basically I can understand the rationale for why they had it the way they had. You have different functional divisions overlayed on top of sort of regional structures, but ultimately, it just caused a lot of confusion in terms of who is in charge, who could finally, who can make the final say.

And a lot of these projects are about. Trade-offs, right? You need to make a decision. You are either gonna be on budget or you're gonna go over budget, but you can actually hit your schedules. So, you know, somewhere along the lines, someone's gonna make the call and someone also has to be responsible when something goes wrong.

So that's, again, going back to the whole like, melee level structure, like that to me was so fascinating. So much happening at that level of analysis.

Steve Hsu: So after years of what I assume, included field work and also, you know, a lot of reading lots of papers and studying lots of other people's work. Have you reached a conclusion for why the Chinese developed differently than the Indian one?

Kyle Chan: Yeah. Yeah. So, the theory I offer is the structure of the bureaucracies in the two countries. Yeah. China sort of reformed its system from a system that looked more like India's more top-down, more centralized, to this one that was based around, yeah. On the one hand, these railway project companies, these sort of specialized state owned enterprises that were formed to manage pro railway projects. And actually a lot of this was learned from the Japanese. There's actually a whole history there of, like, you know, going back to the seventies and eighties, learning from Japanese project management techniques. I hope to actually do one project one day, looking at that historically.

But, being able to sort of like having these nodes, that's sort of the argument that I make. The sort of nodal structure, there's key nodes in the bureaucratic organizational structure where power and responsibility are concentrated and that's how you get things done. Whereas for India, I, I, I argue that it's more diffuse, that you have all these cross cutting structures that a lot of, a lot of very mundane, decisions are kicked up all the way to the top because nobody feels like they can take, take responsibility.

You know? And you know, to be fair, there are many parts of the Chinese bureaucracy elsewhere that still have some of these same issues. So these are not like purely Chinese or purely Indian, but focusing on the railway bureaucracies in particular gave me that sort of leverage to control for, you know, these are two countries with state owned railway systems and you know, yada yada, all, all these things are observable and controlled.

And then, but this, this is something that's different. So, yeah, that, that was sort of the overarching takeaway from that.

Steve Hsu: I, I'm curious whether you've read any of Glen Luke's work. Are you familiar with Glen?

Kyle Chan: Of course. Yeah.

Steve Hsu: He's done really in-depth studies of the financial viability of the Chinese rail system. Do you have any thoughts on that?

Kyle Chan: Well, just that, some of it is not as clear to me. I actually find Glen's analysis very interesting. Yeah, because, but, but overall, my, my take in the whole, financial, sustainability question is that I, I do think that a lot of, A lot of sort of Western analysts are missing the forest for the trees when it comes to understanding this.

I mean, this is not a project investment that's supposed to generate an IRR on the project itself. That's not the main goal. The goal is to build something that will have broad economic spillovers across the entire country. And, and, and that's hard to measure. And there's been very creative economists that have tried to use sort of spatial models, land value appreciation to try to capture some of that.

And I, I look at their work a lot, and find it very useful. But at the end of the day, yeah, it, it's just, it's a different approach you have to take rather than focusing strictly on, okay, is this single project in a return to profit or is the whole railway system gonna return to profit? You gotta think about like, this is a, this is an investment in the entire country, essentially.

And, and, and there you can still, you can solve waste for sure. That's not to say that it all worked out. But to think about the whole system and its broader impact, I think is really crucial.

Steve Hsu: Now after you finished your PhD and you became a postdoctoral researcher, did you, sort of consciously, broaden out your set of research projects that you would be interested in?

Kyle Chan: Yeah, yeah, yeah, exactly. So I started to realize that there's a lot of overlap between infrastructure development and industrial policy that both entail long-term investments and have massive coordination problems. And there in some cases, the state can step in and help, in some way. and, and these are also two areas where China has just been very, very good at, at doing what it does.

so I see that, some of the strengths and some of the issues, come up in both. That is, um, you know, sort of like airing on the side of investing too much, focusing on scale and speed, maybe going a bit overboard when it comes to, replication or redundancies. and, uh, an appetite for risk taking, uh, is one way to characterize it, or waste is another way to characterize it.

But, both infrastructure and industrial policy, I saw that similar patterns, even similar institutions were involved. so, but of course, you know, a key difference is just that, the industrial policy was across, more sectors and involved this combination of, in, in many cases, private sector, private company, you know, sort of shaping, shaping markets basically. So I found that to be really interesting.

Steve Hsu: You know, this, this topic of industrial policy and whether it can work or not. You know, I'm much older than you and so I've seen this play out over a really long period of time. You know, with, you know, the argument shifting, you know, quite a bit over multiple decades. And I think right now I think you're well positioned because I think China is providing an example where it seems like their industrial policy has worked out.

And I think for the last few decades, people in the West have been, or at least an American, maybe American economists have, have been really conditioned to think that, that's just a recipe for disaster, that governments can't really execute properly and can't get things done. And so maybe the pendulum will start swinging back now.

Kyle Chan: Yeah, I think so. I think. China's both, an interesting example that has sort of galvanized other people's interests and maybe caused, you know, us in the US or elsewhere to question our assumptions about, you know, oh, the state intervention is always gonna be, always gonna backfire. The other thing is China itself is a factor prompting a lot of countries to reconsider industrial policy.

You know, you know, now we have rare, you know, all the news, but you can pick sort of many different industries where China's very strong or very dominant. And, there are questions about how to stay competitive in a world with China in it. And I think maybe our attitude towards industrial policy might be different if there wasn't a China type player in the sky in the global system.

Steve Hsu: So when you write a paper or you're formulating, you know, you're, you're, working on some thesis in your research and, and you're trying to come to a conclusion or reach a level in the argumentation where you think, okay, this will be convincing to, target set of, other researchers.

Who are those people? Are they sociologists? Are they at the Kennedy School of Government? Are they economists? Who does? Who do you think of as evaluating the arguments that you're making in your papers?

Kyle Chan: Yeah, that's a really interesting question because I have a number of different audiences in mind. Some are the people who also study Chinese industrial policy and China's political economy. And especially if I'm writing something more, I. sort of academic, then I would have them in mind, as well as other sort of think tank researchers as well, who focus on those areas.

But ideally, I also have in mind a more general audience like Americans broadly, or Americans who sort of care about what's happening in the world or care about what's happening, you know, even in their own backyards. Because I just, I see a lot of things like domestic issues are tied up with US China relations or what China's doing or what the US is doing vis-a-vis China.

And yeah, that's why I, what I try to do in many cases, is try to use language that is not always so specialized for like China experts, or people who are sort of spending their, you know, 24-7 in this area. So, yeah. And, and that's, that's part of like this New York Times piece. I, I just, I realized that there's actually a lot of, I.

Knowledge and understanding that's kind of locked away by, um, scholars and researchers who have been kind of interested in this stuff for a long time. But just talking to each other and, you know, every time I talk with 'em, I'm always reminded that there is just incredible, incredibly deep expertise of people who have been doing field work or traveling to China or studying China in some fashion for, for, for many, many years.

But there's not, there's a huge gap between what they know and what. Folks in DC know what, folks in Silicon Valley know. It's changing. It's definitely changing for sure, but just the broader sort of American business and investment community writ large. So yeah, trying to kind of bridge that gap, I think is, I didn't realize how big a gap there was.

I had been so long sort of like hanging out with my other China scholars, that I just sort of took a lot of these for granted and realized like, oh wow, you know, not only is there a, a, a lack of knowledge or maybe even sort of misunderstanding, but these have really big policy consequences if you have a a, you know, not a, not a fully informed view or a misguided view of like how China works that can lead to really, really, bad policy decisions.

So that's also what I've, what I've been concerned about lately.

Steve Hsu: One we could really go into, which I, I've, I've talked to various guests about at length is the, the AI and chip competition, but may maybe we'll

Kyle Chan: Oh yeah.

Steve Hsu: the podcast. I definitely feel your pain because I feel like I'm in the same position where the people that consume my content might be academic economists or think tank people, or actually, you know, business people that compete with Chinese companies or actually operate in China.

So there's, there's such a broad range of understanding, uh, you know, ranging from people who think they, you know, they actually have a you know, a social score, social credit program to, you know, people who actually, you know, run factories themselves and live in China. So, such a

Kyle Chan: Yeah.

Steve Hsu: gamut. You mentioned your New York Times piece, maybe we can talk about that a little bit. So an opinion piece and, I know it's got an interesting title and often the title is by the editor, not by the author, so I'm gonna ask you about that. The title of the piece, it's a guest essay in the New York Times opinion section, and the title is, in the Future China will be Dominant. The US will be Irrelevant Now, pretty strong title.

Tell me where that title came from.

Kyle Chan: I can tell you it did not come from me. And, yeah, like I, I have to be careful what I say about that title, but let's just say I was pretty shocked when I saw that and I saw it on the day that the piece came out. So,

Steve Hsu: Did you get blown back specifically because of the title?

Kyle Chan: It's funny, I got like a tsunami of emails and messages and responses to that piece. I mean, I was really surprised 'cause I thought I. Again, this is sort of like me coming from China research world. I was like, this is probably not surprising to a lot of people. Both what I'm saying about China and what I'm saying about the US I mean, I'm not saying anything new.

And when I talked to my fellow China experts and China scholars, they're like, oh yeah, yeah. I mean, that kind of put together a lot of stuff that we've been studying. But yeah, I was blown away by the reception and yeah, I could tell the difference between people who just got stuck on the headline versus people who read the piece. And then even people who read the piece, I'm still like, that's not my best work. My, my real work is something else. Like, read, read my other writing or read my other research. Cause 'cause this is sort of just like an overview. you know, I get into what I consider to be more real analysis elsewhere. But, but yeah, that, that has been, that has been quite a journey.

Steve Hsu: Now since that article came out, are you now officially a CCP shill?

Kyle Chan: Yeah, it's so funny. I got called all sorts of things and you know, I think both of us are pretty active on Twitter and, and or x and, you know, you developed thick skin and yeah. I even, even then after that piece, I still got like, some, some new ones. Yeah. CCP, shill on the one hand, but then also some people who really didn't like, you know, my use of the word like China as a threat or something, or you know, I, I get, I get a lot of flack from both sides, I gotta say. So,it's fine.

Steve Hsu: Get it. You can get it from both sides. So you know, even if you say they're a competitor. You know, then there'll be people who say, and I think with reason, uh, they might ask, well, why, why do you have to think of them as a competitor? Why can't we cooperate? And so you, you're gonna get criticized from both sides usually.

Kyle Chan: Right. Yeah. Yeah. It's just part of the game.

Steve Hsu: So now I feel that piece was totally reality based. So in other words, you, you, you mentioned a lot of things in the piece, which I think for people who are really following the US, China, you know, shall we say competition these are key indicators. Like where are, where is the EV industry, where is renewable energy?

Where is, where are chips? You know, where ship manufacture, ship building, you know, all these things. I think for people who are serious and follow the numbers. Nothing you said in the piece was particularly controversial or even necessarily new to, to people who are deep experts in the field. But I think the average New York Times reader probably in aggregate, it was a shocking article.

Like for the people who actually read it all the way through and thought about it, I would say over 50%, maybe 80% of the people who read it were a little bit shocked by the end.

What, what? What do you think?

Kyle Chan: Yeah. Yeah, yeah. I, I think so, and I, I did mean for it to be a little shocking. I kind of wanted to cut through the noise a bit and connect the dots for people because yeah, I, I think, and, and I was talking to the editor about this that I worked with, like, there are a lot of stories about, there you get a story about BYD or you get a story about Chinese EVs, or you get a story about like deep seek, right?

So you have like the whole deep seek moment and people are focused on like that particular company sometimes. And if you're, if they're more sophisticated, they might kind of see the industry as changing and China moving, moving up the ladder in that particular industry. But I just hadn't heard that many people in sort of like mainstream media talking about.

All of these different sectors, all these different scientific fields as well, sort of backing. And, and that's a, that's a big story, uh, as well that I know you've actually written about. So there's just these bigger trends happening, behind each of these individual sectors. And so, yeah, like, you know, I'm sure some of the people who, who you speak with and, and probably who listen to this, to this podcast, you know, they, they might be, deep in the weeds on like biotech in China and being like.

Wow, there's some big changes happening here. Some of the oncology treatments are, really, taking off. the clinical, the pace of clinical trials is really incredible. But they might not have a view into the shipbuilding world. Or it might be very light. And so to kind of put it all together and just be like, look, this is, this is not just a single sector or a single industry or a single company.

It's not just DeepSeek or BYD or you know, any of these ones that make the headlines. It's a much, much broader pattern. And that's why we need to wake up. Like, you know, we can't just be like, well, you know, we'll, we'll, we'll concede that one. Solar panels. Sure. You know, we who need, you know, you, you can make cheap solar panels, whatever.

but. Like a lot of these, and especially the point about the industries of the future, I know this is sort of like a, a hackney term now these days, but I really believe that the areas that China is really focused on are the building blocks for the next generation of technologies, emerging sectors that are gonna be not only sort of like the economic growth engines, the job engines of the future, but sort of broader progress is gonna be built on that.

You know, AI, robotics, you know, all that, all the stuff that, that we've all heard about. But I want to bring attention to some of those things too. It's not just today, it's like the trend going forward that people need to pay attention to.

Steve Hsu: Yeah, you know, I think a moderately competent government, you don't have to be super duper competent. You're government officials, but if you're moderately competent and you listen to area experts, it's not that hard to formulate a kind of China 2025 plan where, you know, you might compare a few different areas and say, which of these should we put more weight on or more resources behind?

Because they will, as you say, be the building blocks for the future. It's not that hard, I think, for them to like to pick the right ones. What, what I think is unique about China, and this is something I think both you and I have written about, is, is, you know, there are a few factors which affect all of these verticals.

So you have what the surprising thing is, as you said there, you have all these different verticals where if you were able to go into the weeds in each of them and determine what, where is us versus China competitiveness in this particular vertical of like humanoid robotics or self-driving cars or, you know, cancer treatments, whatever it is, you would see them advancing across all of the, almost all of them simultaneously, very fast.

And, and the factors that underlie that are one human capital, they have just, you know, an extremely large population of young people and engineers. Um. They sort of have their act together at the government level. Like I just said, the government can maybe not pick individual winners, but they, but they do know what areas to pump resources into.

And then third, they know enough to get out of the way and let actual market forces determine which companies actually win in the space. So if the, if the companies, if the tion somewhat fair and different regions, different cities have their own individual companies, car manufacturers for instance, that have to actually compete in the, the, the, the, entire domestic marketing globally, then like the, you know, quality will rise to the top.

And so if, if you have those three factors, then you can't advance simultaneously in each of these verticals all at once. But I still think most Americans don't realize that those factors are in place and indeed that they are advancing in each of those verticals at the same time. I think that even that basic idea, which you conveyed well in the piece, is pretty new to most people in America.

Kyle Chan: I think you characterize it really well. Yeah. I think, again, I have to kind of step back and see China through someone through the eyes of someone who doesn't follow China all the time. And that's, that's actually hard for me to do. So like I, I think some of the ideas of how Chinese industrial policy works that are out there are like, either it's like a very top-down system like Beijing leaders or Xi Jinping himself says something and the everyone executes, you know, and it's like, well, I think they tried that maybe in the past.

And the sort of a command economy structure and that didn't work out so well. And also another idea is that you just kind of like. Flood the zone with subsidies. And again, if it were that simple, like at a, at a certain point, you're gonna run out, right? Like, if you're not gonna build competitive industries, if these companies don't upgrade technologically and become, you know, not just domestically competitive, but globally competitive to a certain extent, then you can't just subsidize in front, you're gonna run out.

So like, at a certain point it's about, as you said, kind of like shaping the market, shaping the ecosystem, like, providing, you know, I kind of think about, I, I know, I know you're a, you're a physicist and I, and I love the physics analogies, but I often reach for the biological ones more when I, when it comes to industrial policy.

And it's like, you know, you can't just turn the knob all the way to 11 and expect that to work. It's, you know, you want to kind of like. Add the ingredients that you think will help, you know, if there's some key component or key, you know, aspect that's missing, maybe the government can sort of backfill or, or, or contribute in that area.

And then especially looking at sort of like problems in the supply chain and, and, and trying to, trying to, fill those in so that, that can be very useful. But overall, it's, it's this mix. You're right, it's not just like top down. It's also not just flooding the zone. It's sort of like trying to cultivate this ecosystem.

And AI is like a classic example of that, where it's like DeepSeek itself is not a state owned enterprise, you know, by any means. It's quite the opposite, right? But I would argue that, you know, even though it's not a direct product of industrial policy in the traditional sense. The people who are incredible, like researchers who are doing the cutting edge research at DeepSeek, they came through the Chinese university system, which had investment research grants, a lot of which was targeted to AI. So that all sort of benefited that broader, industrial ecosystem.

Steve Hsu: Yeah. You know, one of the points I made, in the wake of, mark Zuckerberg going out and, and hiring, you know, with exorbitant salaries and comp packages, all of these top, AI researchers from companies like OpenAI and elsewhere is, you know, a big chunk of those people were actually educated in China, at least at the undergraduate level.

And, um, one of the things you, if you're in my position, a professor that is in a quantitative subject, you see that in China because there's so much emphasis on math, um, solid grounding in math, the, even the software developers, people who are coders generally have much better math chops than the coders in the United States.

And so, to do research in AI, there's a certain sort of core set of things you need to understand, like multi-variable calculus, linear algebra, statistics, probability, things like that. Uh, it's. It's very, very likely that a Chinese CS major actually has pretty good mastery of those things. And so they could, if they wanted to do AI research, whereas in the US you're already cutting it down by maybe like only a quarter of CS grads in the US are strong enough in those areas that they can pick up a, they can download a paper from Archive on AI and actually read it on large language models and actually read and understand the matrix notation and stuff like that. So tensor notation. So, even that sort of really basic commitment to just education STEM

education, it math itself in the pool from which, they, for example, AI is able to draw from. And that would be the same thing if you were doing, you know, trying to use computational methods in genomics or biotech or something like that.

Just a lot more people in China are able to sit down and actually read papers in that area and actually apply what they read in those papers. Whereas in the US it's much, much less common for a biologist to actually have that math

background.

Kyle Chan: Yeah. Yeah, that's, that's really interesting to have that kind of scientific foundation. Yeah. You know, it's funny here in Princeton. I like to kind of just pop into random seminars and fields that are not my own. So like, and, and I, I have been to more than one CS seminar or CS talk, where I look around the entire auditorium like almost all Chinese people do.

The speaker is Chinese. And I'm like, you literally could just have this whole talk in Chinese right now. Right now, like in theory. But, yeah, but it's just sort of like an assessment , kind of like the strength of the talent pool, the depth of the talent pool. And you can, you can see it even out here in Princeton at the graduate level.

I mean, it's very, very. Obvious. So, yeah, when I, when I saw that, you know, this sort of like, Damien Ma MacroPolo had, sort of like the tracker of AI talent and you know, over 40%, almost 50% comes from China. And then now, yeah, with a sort of metas, super intelligence team, like seeing how many people came from China, I was just like, on the one hand I was, I was shocked.

I was like, wow, that is, I mean, these, those guys, their lives are like, you can imagine their parents might have come from a generation that was like dirt poor and they're gonna get a hundred million dollars pay packages. I mean, what a, what a change. But overall it was just, you know, it was surprising, but also not, not that surprising.

I mean, it was like, yeah, I, I, if you look at the nps, co-authors on like, virtually like, you know, a huge chunk of the papers, it's a lot of, It's a lot of Chinese researchers, or, or Chinese origin researchers. So yeah, it's, it's just that kind of depth, I think again, really hasn't been emphasized enough, in, in trying to understand, you know, the latest, you know, you know, where's the next DeepSeek coming from?

Like this, this sort of like a deeper level, yeah.

Steve Hsu: I'm curious, in the sociology world, are people hostile to, you know, if, if you break out human capital as a separate factor for productivity or, or, you know, development, are people hostile to that? I mean, certainly if, certainly you couldn't link it. I mean, you couldn't link it to anything like genetics, but just even saying that.

This system produces more 22 year olds that can do linear algebra per capita than this other system. Like, is that an okay thing to do in sociology, or do people get mad at you over that?

Oh, I think, education, human capital, the pipeline. I don't think those are controversial things. In fact, I think there are demographers who like to specialize in these sorts of questions. Okay, good. Because I, for example, like, uh, there's a guy called Garrett Jones, I don't know if you know him. He's at George Mason University. He's an economist and, and he, he's done studies linking, like he tries to estimate the, the, the sort of average IQ of a whole country, which, you know, is equivalent to just saying like, how well, how well they do on PSA scores or TIM scores or something like this.

And then he relates that to their economic growth rates and things like that. And, and, and of course he doesn't mention. DNA or genetics or anything like that. It's just country by country. There's an education system. It produces these outcomes and maybe that is an important input factor into productivity, growth or whatever.

But even that, like I think he's found it somewhat controversial. Like he's gotten some pushback that like, what shouldn't be looking too closely at these things. I, I don't know, apparently maybe you have not felt that.

Kyle Chan: Yeah, I don't know. And also, the funny thing is like in the business world, you talk about talent all the time. I mean, that's like the word that you use. So, yeah, maybe that encompasses maybe a broader set of performance metrics or something like that. But whatever it is, like there's a sense out there that this matters a lot.

And if China has a lot of it, that also matters a lot.

Steve Hsu: Yeah, I think, I think people still, even today, underrate how big the gap is because in almost every metric, if you look at in, in terms of human capital, if you look at China, it, it tends to dwarf the US and it tends to be more like China is roughly equivalent to rest of world

combined. And,

Kyle Chan: Yeah. Yeah. Yeah.

Steve Hsu: You know, and so I think still people can't really quite absorb

that, that fact, even though it's in the numbers.

Kyle Chan: Also, I, I, I think people don't realize that like factories, you know, like, again, like a battery plant, you, you take kind of like a mid-level manager, you the level of technical capabilities you need to have, and then you multiply that by. Across an entire factory, and then times all the factories in that industry, and then times all the industries.

I mean, you really need, you know, I, I was, I was talking to someone about industrial robots and industrial automation and just deploying them. So it's not, you can't just sort of like install ' em and let them run. You need people who have the technical skills to be able to do that, to understand some of the mechanics, how to operate them, how to repair them, how to deploy them properly.

And you need a lot of those people. I mean, the scale of what's happening in China in terms of manufacturing, in terms of the number of new factories being opened, or factories where, you know, these automation systems have to be serviced or managed. Yeah. You just think about the number of people who are needed, who are at that level.

And that's, that's something that I think a lot of people don't realize,

Yeah, I mean, it's a great example. So you've probably seen these videos of EV factories where they're turning out a car like roughly every

minute, and, and there are these robot arms that are simultaneously welding and moving a plate into place and doing. And if you just think of the spatial reasoning required to install, like even if the robots all came from Germany, which is not true anymore, actually, mostly they're built in China, the robots.

But you get the robots like placing them and solving these problems. These, so that all the arms can work properly in the cars placed in a certain position. Like all of that is actually very, very g to use the verbot term. Very g loaded. Like a lot just can't, that like if I said This is your job and figure out where to place these robots so all of these arms can operate simultaneously without screwing each other up.

And, and they can do, get into the part of the car where they need to get into and you know, we need to put it in this corner of the factory on this conveyor belt. A lot of people just couldn't solve that problem. It's actually a pretty difficult problem to solve. And the way you solve it affects the overall efficiency of your factory. Like how many squares yeah. Yeah. of factories do you need to actually do what you want to do, right? So, so all these things are just, require a ton of human

capability.

Totally. Yeah, absolutely.

Steve Hsu: Yeah. Okay, so, maybe we can move on from the New York Times piece to the Substack post. I, it wasn't just you, I guess there were multiple authors right on that, and so we maybe should give credit to your co-authors, but I, I was interested in the philosophy of this gentleman, Lou Fung, who's at, I think Beijing University. Maybe just tell us a little bit about the article and, and this individual.

Kyle Chan: Yeah. Yeah. So Luang, he is a very prominent, I don't know how to categorize him now, a political economist. Economist. And he has done a lot of work on industrial development, in China in particular. in particular this, this piece. So, it wasThomas at Sindification, who actually asked me.

He found the piece, so he should get credit. and he asked me to write an introduction to it, given that it was related to what I work on. And, This piece talked about what I called industrial maximalism. That is all the arguments that you hear of like from the west, the critiques, the criticisms about over capacity, about how China should shift to, away from sort of more industrial development and towards, you know, consumption or less production. And, you know, all these things that his piece was saying, almost the exact opposite. He was saying China should double down on manufacturing, on industrial capacity. He even had a whole part where even steel, he argued, was not at a state of overcapacity. I thought that was so bold. You know, so, so if anything, I just found the piece almost like a, like a very erudite troll like a lot of the Western critiques, but I also found it to be really fascinating, like point by point.

And I, and I wish that yeah, maybe I could find a way to, to get out the whole version in English, because it's just, it's, it's very, very long. It's this very long interview where he's talking about, you know, what he thinks are the sources of strength for China's industrial development. Why does he think the US can't do that?

Why other countries have struggled. He has a lot of actually very interesting cases where he looks at specific companies. Some of these semiconductor companies today in China, had their roots going back even to the you know, to the pre-form years. So, I find all this historical stuff very interesting as well.

But, yeah, overall this argument that China needs more of this, that actually this is, this is the direction that China should, should be focused on. I just find it so fascinating. And yeah, he has a few little jabs too at, you know, JD Vans, with the peasants comment and things like that.

So, you know, I, I, I think he had fun with this, Q and A in, in addition to sort of making a serious argument.

Steve Hsu: Right. So the original document, it, was a Q and A, where is it published? Like, what's the Chinese audience that it's actually aimed at?

Oh yeah, so I think this was Guancha, which is, I don't know how to characterize it. so it is a state affiliate media. It is very, it's sort of a very prominent platform, for a lot of, you know, researchers will, will, will write pieces on there. You know, I wouldn't quite call it the New York Times of China, but it has that kind of intellectual sort of exchange. So the readers would be other economists, academics, but also policy people,

business people?

Kyle Chan: Yeah, yeah. I, I think, I think broadly speaking, yeah.

Steve Hsu: Yeah. Okay. And, and, and the, the piece that I read in translation was already pretty long. And you're saying the, the whole thing is actually even

longer.

Kyle Chan: Yeah. Yeah, that's right. That's right. He goes into like sort of history lessons and I mean, it's very interesting. Again, like he, he really pulls out a lot of, a lot of case examples, and, and sort of dives into them. But, but yeah, I, I think one of the biggest takeaways actually was just thinking about like, industrial capacity is not just something about building physical factories, but as sort of this, you know, and this, this relates to my work, you know, the sort of organizational capacity that a society has to be able to do this, you know, tying back to what we're talking about with talent, tying back to what we're talking about with, you know, having a, at least sort of like decently operating, government that can implement some of these things.

So yeah, that's, that's where I think I found it very useful that he was sort of trying to cut through some of the previous ideas of like, oh, you know, a lot of the stuff just got outsourced to China. Well, if it were so easy, you know. Why isn't every other country doing that? Right? Things aren't just outsourced to China.

Like China created a system to make it attractive for other companies to set up their operations there. They did a lot of, you know, industrial policy to not only do that, but to make sure that the learnings from those companies got transferred over in some, some way and absorbed. And again, all these terms, I mean even the word tech transfer unfortunately just like it is not about handing you or handing someone a piece of knowledge, right?

It is. I I like the term actually sort of like absorption capacity, in a, in a way more because, and this comes back to sort of China's scientific foundations that you, it's one thing other countries have tried to lure foreign companies to build plants, in their jurisdictions and train their workers, and then it sort of just ends there.

And the question is, why has China been able to get that technology and understanding and sort of technical knowledge and have a diffuse throughout society more broadly and then create homegrown competitors that can not only catch up, but then actually, you know, move beyond, beyond that. So, that's sort of, to me, the most interesting question. So I like lu fun's sort of readjustment of how to see some of these issues,

Steve Hsu: Yeah, I, I think people in China are aware of this unique aspect that, that, you know, the foreign company, apple comes in there, is knowledge absorption, but the Chinese are able to both contribute to, Apple's innovation in, in trying to localize the process, but then also potentially go beyond it and create competitors. And, just to pimp a recent episode of Manifold, I interviewed the author of the book, Apple In China, which is a really, really exhaustive look at the operations of Apple in China, and then how that gave rise to domestic smartphone manufacturers and how the, actually today we discuss this at length, like the, most of the, the leading innovation is actually happening in China when it comes to smartphones. And in a way, Apple is actually kind of a laggard now in terms of, you know, the features and the capability. So, so that's a

Kyle Chan: with Tesla,

Steve Hsu: And if you want a very specific, in depth discussion, that that episode of Manifold, I think is very good that I forgot the author's name, but

Kyle Chan: Patrick McGee. Yeah, he was really great. I actually interviewed him too for a different podcast.

Steve Hsu: So he spent, so he spent a huge amount of time writing this book. So let's maybe steelman both sides of this. So, from the western side where this term overcapacity comes from, the idea is something like, wow, these autistic Chinese guys just keep building more factories and they just kind of keep wanting, making, wanting to make more stuff.

And they don't have good profit margins. They're just driving. It's a race to the bottom. They're just making so much of this stuff and flooding the world with cheap, you know, maybe high quality. Now we grudgingly admit it's high quality, but cheap stuff, and they should just calm down and relax like people in Europe or California and take more time off and do other stuff. And, it's just, it's just, a mistake. For them to be so frantic focused on industrial production, technology, you know, all this stuff. and maybe now you could say like, if, if, if someone said that to Lu Fang, what would his response be?

Kyle Chan: Yeah, I mean, I think his response would be that this is something that is helping consumers in the rest of the world, and this is a product of China's investment and trying to develop like this very sophisticated manufacturing platform, right? This whole system of industrial clusters of, you know, technological upgrading that makes it Yeah. Uniquely capable of producing relatively affordable, high quality products. Yeah.

Steve Hsu: You know, one, one of the things I, I wrote a whole thread on this in, on X 'cause I, I really did enjoy the, you know, your art, your translation of the article. One of the things to consider is that throughout most of the 20th century there were, or at least during the history of the People's Republic, there were American sanctions, western sanctions on China. So they could not get Western technology until that opening period under dunk. And so. You can't blame them for wanting to, you know, surmount those technological hurdles as fast as possible because previously they were denied the ability to do so. So if they're a little bit overly indexed or overly focused on technology catch up, you know, they have things like the Opium War to look back on, right?

So, I think they still feel like they're playing catch up. Even if they've now caught up, they still feel like they're playing catch up. People tend to, you know, pendulums tend to overs swing, right? So, they feel like they still wanna overindex on this stuff and they're gonna keep pushing and there's a little bit of paranoia, which for them is justified.

Like, like, you know, they remember during the Korean War, the US just flatly threatened to use nuclear weapons against them, right? Like we forgot that most Americans can't remember that one. Like that. No, that never happened, did it? And then you look it up and you realize, yeah, in fact it did. And they never forgot that, right? That there was a period where they were under technological embargo. you know, which is why the chip sanction thing pisses them off so much. But they were under a technological embargo. And at the same time, MacArthur was urging, you know, the US to just drop nuclear weapons on Maturia, nuclear bombs on Maturia. So they remember that. And that is definitely gonna color their strategy, right? They're gonna be more risk averse. About ever falling behind in any one critical area because they used to be behind and it caught, they, they paid dearly for that. Right? So I think like, just in people to put yourself in the psychology of the other side, even if you don't disagree, even if you don't agree with it, even if you just say like, oh, you know that anti-China Cold War hysteria, that was an aberration.

We'll never see that again. Americans are so diverse and nice now that could never happen again. You still have to allow for the fact that the other side doesn't believe you. Like they, they experienced all this stuff

Kyle Chan: Yeah.

Steve Hsu: victim of it. So they're, they're not gonna give up on that right away. They might give up on it in 10 or 20 years, like in 10 or 20 years. Chinese kids might be lazy and, you know, like lying around all day, like, you know, but, but it, it won't happen for a while. So right now they're, they're still like very, very focused on working hard to get ahead.

Kyle Chan: Yeah. Yeah. And, and in fact, if anything, I think going back to that Lu Fang piece. He sees it as sort of a natural kind of progression of de-industrializing. That this is just kind of what happens over the course of time for a society, unless someone steps in and tries to alter, especially the financial system, to try to allocate resources, in such a way that incentivizes that industrial capacity to remain.

But, yeah, the broader point about self-reliance, this paranoia being in a world that is very insecure, where someone can cut you off, you know, later on it was the Soviets that were the big threat. and then. Now, most recently, like ever, all these fears, like imagine you have been afraid of something your entire life and people have been telling you, don't worry about it, don't worry about it. It's okay, it's okay. And then it turns, you know, you're afraid of a robber, you know, burglar entering your house. And people are like, it's not gonna happen. And then a burglar actually enters your house and then you're like, confirmed all of my like, safe room, you know,investments were worth it.

And in fact, I'm gonna double down now. So that's actually what's happening now. I think with the export controls, with the Huawei sanctions, with all of these, you know, it's sort of like, I, I joke that, there's sort of like a rule of three. A when a Chinese company gets mentioned three times in international media, like it will take, within three years, it'll be an entity listed. And you know, the sense that like, at any moment we could be, be cut off, you know? And so Xiaomi Lei Jun has the whole story about how he thought, Xiaomi was gonna get cut off. And that was partly the reason why they pivoted EVs. So this, this sort of fear. Both at the policymaker level, but also at the company level. I think it has only sort of gone further into overdrive, with the actual sanctions that sort of validated all these concerns. So yeah, this, this, this feeling of like, could we be cut off and what do we do? How do we prepare now that I think is, is really important for, for, for people to understand who are not, you know, feeling that kind of sense of insecurity all the time.

Steve Hsu: Yeah, I think that. You know, it will be a while. You know, I would guess the timescale is decades before they feel comfortable enough that they can take their foot off the gas pedal. I think they're gonna have their foot on the gas pedal for another decade. And if, if they've actually, so in, you've probably seen me post on x, these skull graphs

Kyle Chan: Oh yeah.

Steve Hsu: You know, like you look at US electricity production, it's kind of flat. And then you look at Chinese electricity productions, this, and then during, during this period of time where American America is really dominant and they're the hedge mine and the Chinese curve is way down there. It's like, the US is like, we're number one. We're number one. And then the curves cross. And when the curve crosses and the American guy's like, oh shit. And then like, but they, you know, because the first derivative is so large, this just keeps shooting up. And eventually you end up in this era where, where the Chinese appear and the Americans are down here, and then there's like a little skull.

Kyle Chan: Yeah,

Steve Hsu: And I just don't see that, I don't see that letting up in most of these metrics KPIs. You know, for another decade or something.

Kyle Chan: Yeah, yeah. And they're making a big bet. I mean, they're betting that more investment in these scientific fields, more investment in these industries is the long run right thing to do. And so it'll be very interesting. I mean, we will watch the experiment unfold before our eyes. Like we, we will see, you know, like these are huge, huge bets. In many ways I think the US should be doing something similar. Like

Steve Hsu: Yeah. Well, I

Kyle Chan: I. Same way. Exactly. But there are certain industries that, yeah, I, I think we realize now it would be, it would be a shame to kind of like, sit back and hope that it all works out and, and not have, you know, not have tried to do something in the process.

Steve Hsu: You know, I, I recently tweeted there was a, there was a, an article in MIT technology review and they interviewed this economist David Autor.

Kyle Chan: Yeah.

Steve Hsu: You know, he, I think he may have coined the term China shock. And now they're talking about China Shock 2.0 where instead of shock on legacy industries like making cheap furniture or dolls or t-shirts and, and of course that did affect American employment.

Now it's Shock 2.0 where it's the advanced industries where they're threatening to take away the advanced industries from us. I wasn't expecting that tweet that posted on X actually got a lot of attention, but it did because I, I think people didn't realize this was happening and when they read David Autor saying it, they suddenly thought, okay, this has credibility. Like this guy has been saying this for three, five years, but we just ignore him. But this guy, David Otto said it very explicitly and so now, now we

Kyle Chan: He's the China shock guy. Yeah.

Steve Hsu: Yeah, because he's China Sharp guy. But, I think to me the obvious prescription, I think Professor Otto would agree with me is that the, the, the main thing is America has to actually strengthen itself and, and compete in those

Kyle Chan: Yes.

Steve Hsu: that it wants to, has to reach a level of excellence where it can retain. Right control of those industries or leadership in those industries. And you can't do that without all the usual things like industrial policy and R and D investment and production of human capital here at home, or at least stealing it from abroad or however we get it. So, I mean, I think all these things are pretty clear.

Kyle Chan: Yeah, definitely. I mean this is, this is going back to like part of what got me so interested in these topics in the first place. I mean, how to understand them at a broader sort of system level. View rather than seeing like this particular company that's maybe not profitable or this particular industry where the economics just don't seem to work out.

You see that some of these are links in this chain or are nodes in this broader network. And if you break that critical node, you're gonna have huge spiller effects onto all the other dependent industries and upstream and downstream linkages. You know, these are all interconnected. And so, yeah, like to me, in some ways the, the puzzling thing is not what China does, but what the US.

Like does or doesn't do when it comes to these things where it seems so glaringly obvious that, you know, here here's a coordination problem that is sitting right before our eyes and we are all like, kind of nodding our heads being like, yeah, boy, that, that would be, that would be bad if, you know, we got cut off from this, you know, critical part of the supply chain.

And then it happens and we're like, oh, that's, that was bad. But we still don't really step in and, yeah, the question is, you know, can we, can we kind of overcome some of our very strong priors about what, you know, policy can do? And, and to be fair, you know, I, I share like a, a strong sort of skepticism of, state intervention.

There are many cases where it does do a lot more harm than good. So, you know, this is not just wholesale, like, oh yeah, we should take, you know, nationalize everything. And, but to think carefully about this, the government could step in and at least support private industry in a way that wasn't just totally up to the markets and up to, you know, global price, prices, which can be volatile or unpredictable.

Yeah. That is something that I hope people think about more. Mm-hmm.

Steve Hsu: To take a specific example, like the Biden Chips Act, I was not very optimistic about what it would accomplish or that it would be conducted efficiently, but directionally, I was not against it. I was not against the government saying, Hey, chips is an area we cannot lose. We have to support some domestic activity in this area.

But of course the devil is in the details, like, can they actually execute properly and make it work? I think for people who are really actually, so, if you've done cutting, if you've done frontier level research in an academic lab and you've built a company, so you've seen both the application in a marketplace and also the like, where do the original ideas come from?

If you see those two things together, then you can understand this ecosystem that's required for America to compete. But if you've only done one or the other, you could be wrong. The university guy doesn't necessarily understand, like most of the stuff people do at the university isn't that valuable and, and can't be commercialized.

And, and, and the applied guide in Silicon Valley might not realize like, oh, this, this thing that I'm, oh, I can now think of a hundred different uses of this thing. That came from some idea 20 years ago that at the time it was a crazy, considered a crazy idea and only DARPA or NSF would fund it. Nobody else would fund it.

Kyle Chan: Yeah. Yeah.

Steve Hsu: So it's important to have both of those connections and not a lot of people actually have seen both of those, different, parts of the ecosystem.

Kyle Chan: Yeah. If those people are out there, they should get to work because we need that kind of, like that intersection of, of experiencing and, and capabilities.

Steve Hsu: Yeah, I'm very worried because right now we're doing everything wrong. You know, we're, we're, you know, making it harder for foreign students to come here and pursue their education. Here we're cutting funding to large chunks of science. You know, it's, it's all gonna be, I mean, of course, maybe it's just gonna be a transient hiccup that just lasts for a few years, but it could have really negative long-term consequences for us competitiveness as well.

Kyle Chan: Yeah. Yeah. And that's the thing that I was, I don't know if I pointed it out in that piece, but I, the thing is the, the like breakthrough treatments, medical treatments that are gonna come out, you know, in 20, 30 years, those are being researched. You know, to your point, those are being researched to date.

There are labs who are testing, you know, new sorts of areas of research, and exploring that now, who are getting their funding cut off. And there are a lot of, sort of these forking paths that are just never gonna happen because, you know, at this stage they got, they got sort of nipped in the bud.

Steve Hsu: But it, you know, it is very hard, like if, if, if even some engineers who are more, like, they've only worked in industry, they've only worked very close to like the market and the customer shipping a product, even they have a hard time understanding like, what are you physics guys doing with, superconductor?

Like, what's the point of that? You know? You know, I, I've heard both sides. Like there are people in Silicon Valley who think all these academics are useless, and, oh, if you just give some smart agentic engineer the problem, they can solve it. And they just lack a historical understanding of the innovations that they currently work with. Where were those innovations 20 years ago? Right.

Kyle Chan: Yeah.

Steve Hsu: It's very easy to get this thing wrong.

Kyle Chan: Huawei's Ren Zhengfe, Huawei's founder, he actually recently had a piece talking about this. It almost sounded like he was trying to convince people in China too, of the value of sort of foundational science of, you know, even like the way he characterized. He was like, yeah, some of them are kind of weird characters and we have to we have to sort of like embrace them in our society because they're so valuable and they might be kind of like a bit odd and they might have like slightly odd ideas, but actually those are the seeds that, you know, are so later on that become basically the next, you know, Huawei or Nvidia or, you know, whatever it may be. So, you know,

Steve Hsu: Yeah, I think I read at least in translation some of that. And yeah, I mean he was, you know, Huawei itself actually hires a lot of people who are doing sort of very mathematical or theoretical work.

Kyle Chan: Right.

Steve Hsu: And you know, a little bit surprising. And I think he was urging other companies to do the same.

Kyle Chan: Yeah. Yeah, exactly. Yeah.

Steve Hsu: Looking further forward in the future, I, for you, Kyle, the sort of US China competition or, or China, the, you know, the, the, the biggest story of the last a hundred years, which is probably China development or last 70 years at least, is, is that where you intend to focus most of your energy going

forward?

Kyle Chan: Yeah. Yeah. And to continue to focus on, yeah, developing in, in, in Chinese technology and science and industrial policy. So those are the main areas. So yeah, like I, I kind of like to position myself as someone. I'm not gonna be like a leading chip expert. There are people who are deep in the weeds on specific photo resists and can tell you, you know, why, you know, this versus that.

That's not what I'm trying to do. But I'm trying to see again, you know, maybe the New York Times piece is actually useful as that characterization where I'm trying to connect the dots here and see what are the broader, systemic factors that are contributing to this. And, uh, how is that shaping? Yeah. Not just the US China relationship, but sort of the China global relationship. And yeah, I actually look at a lot of stuff too, uh, about China's, uh, interactions with, you know, global south countries or China and Europe, which I think is like, um, you know, maybe Americans don't care as much, but Europeans really care.

And, there are a lot of people who care about, you know, the rest of the world. So, those are all things that are, I think, very dynamic and there's a lot of things happening. So those are, those are areas I'm following very closely.

Steve Hsu: Let me, let me throw out a couple speculative things. So, there's one meme that's around on the internet that says China has already won.

Kyle Chan: Hmm.

Steve Hsu: So if I were to steelman that, I would say something like, well, you look at any of these verticals and you know, if they're not past, if they haven't surpassed us, they're, they're coming close and their, their first derivative is clearly much higher than ours. And this is happening across, you know, the 20 most important areas. and we are not gonna get our act together. And actually our, the rigor of our education system is, is actually going slightly down as opposed to up. So how could we possibly win out? Now the, the one scenario, which I think if, if you go around Silicon Valley and talk about this, some people will just say like, they'll just deny that the what I just said was true. So that's just a different, I think, perception of reality. But for the people who accept that, that's in terms of the metrics, that is actually a fair characterization of what's going on. The usual place where the Silicon Valley people can still pin their hopes is the following. Fast takeoff in AGI led by Americans. And that's the one thing that saves our asses, right?

Kyle Chan: Yeah.

Steve Hsu: That's literally what people, so, so again, we could meet people who just don't believe me that the cur, you know, the skull curve predominates, right. Like they would argue, oh, they're not really ahead in rocketry, or they're not really ahead in this, or, you know, they're not really close to us in jet engines. You know, they could argue stuff like that. So that's a different category of person. But for the people who accept these numbers, and I think the numbers actually are, as I present them, are correct. Usually if they're a technologist, the thing that they're gonna retreat to is we're gonna win the AGI battle, and that's what's gonna keep

Kyle Chan: That'll be like the Trump card. That'll, yeah. yeah.

Steve Hsu: You know, number one. And therefore like, please, please nationalize my company and write me a check for, you know, for $10 billion. So, but I'm curious, what, what are your thoughts on this?

Kyle Chan: Yeah. Well, so one thing that's very interesting that I follow is the debate in China about the zero to one versus one to a hundred sort of innovation, scale up thing. And, and it's still an ongoing debate. I mean, I think, you know, I think people still debate like, what, what was Deep Seek? Was it a zero to one innovation or was it another fast following thing?

And, you know, you can go back to like mRNA vaccines or sort of other, other areas where I think there is still a sense of like, yeah, I, I don't know. Like, could China have done this? What counts as an innovation? there are other people who argue that like drone technology or like lidar systems today, are, you know, are essentially a Chinese invention now because yes, like parts of that have been out there for a while, but to really become a real thing in the world, is, is something that, that China did basically.

So that's all, all over the map. And I guess one thing I I will say on, on the US side is, There, there are possibly, I mean, maybe this applies equally well to the training side, but I, I would just, um, have a big dose of humility for like, I don't know what the next sort of ChatGPT might be. 'Cause to me, ChatGPT was itself so huge.

And I guess, you know, if I had been following a little more closely, I would've known that like, okay, you know, GPT-3 0.5 was already out, and that's actually not, it was actually not even the, you know, the hottest model. But still, I, you know, there's just. I feel like we're just living in a time of very large frequent paradigm shifts in technology.

You know, and then you can, you can zero in on specific, specific industries you can look at like batteries. I mean, it was just yesterday I felt like when NMC batteries were seen as the standard, and then you have the shift to LFP and then maybe solid state or maybe sodium ion, so sodium,will, will be next. Like, who knows? I now feel like we're in this, this state where it's not like a single race or it's not like a one and done, you know, type deal. So, yeah, I, I'll be curious. Like this is sort of hedging, sort of like punting on the question, but I, I actually think it's just sort of exciting. I mean, I hope that both countries continue to push the scientific and technological frontiers for humanity's sake in terms of coming up with better, better medical treatments, coming up with sort of like better things to help with our material lives. I think, you know what, what's not to love about that.

Steve Hsu: You know, I like this optimistic perspective. It's so seldom articulated. Probably, you know, that China talk host Jordan Schneider.

Yeah. So I, I was talking to him, a year or two ago about this stuff and. I took your line at the time. I just said, look, isn't this great? Like, all these technologists in China, you know, they're selling you their stuff at rock bottom prices.

Like they're making stuff more efficiently. They're selling it to you, and they're actually starting to make real fundamental breakthroughs like, you know, new battery technologies or, or you know, new, new AI models. You know, so, so what's not to love? if, if suddenly, like a huge number of scientists and technologies and factories appear in one corner of the planet, but they're sharing, they're selling everything to the rest of the world at rock bottom prices, what's not to love?

Like, that's just a great, optimistic future. And he just couldn't believe I was saying that. He was like, yeah, but aren't you worried about what happens if they become the hegemon and stuff like that? And I'm like, no, I'm not really that worried about it. But yes, I think that the optimistic future we should aim for is one where the US and China can coexist and mutually benefit each other. Right. It sounds like a cliche, like win-win is a cliche, but you know, I, I, I still think it's possible

Kyle Chan: Yeah. Yeah. I mean, I just think like, just from a pure sort of science standpoint, like yeah, it, it would be so sad to have so much scientific capability, bifurcated and isolated from each other, when in fact that's really, you know, like, I don't know what the term is, but like, you, you want that kind of multiplier effect where like each group alone can do a lot, but you want that kind of, you want all of it to be, you know, sort of, intermingling and fusing. Yeah.

Steve Hsu: Short of an actual, like hot war, I, I don't think the site science and technology ecosystems can fully decouple between the two countries because, you know, you still have flow of human capital back and forth. And now with. AI and, and people publishing papers and stuff, that the information spread, the knowledge spreads so fast, faster than it ever has before.

So, I don't think, you know, if, sure, if we have a hot war, who knows? Like maybe the whole planet will be destroyed. But I think in general, we won't, we won't, we, we will benefit from all, any investment, whether it's made in China or the United States in fundamental research, the whole world actually does eventually benefit from that.

Kyle Chan: Yeah. I would point to GitHub as like, just spend time on GitHub. Just see who's, who's, who's contributing, who's, you know you can see like all the different names, different people around the world. I mean, it's just amazing people who have never met each other, people who may never meet each other and building complex software that is usable, that is actually, you know, companies use that as the foundations for some of their commercial services.

Yeah. I just think that's so incredible. Maybe, you know, I, I can see yes, there could be like a dark cloud in the horizon in terms of yeah, some of these issues flaring up. But yeah, I mean, in, in some ways, again, I, I, I just think we're also so interconnected. I, I think you're right that, um, some of these kinds of knowledge flows. It would be very, like, to cut them off would mean quite severe, um, I don't know, like a state of repression for yourself, essentially. You know, much less for the other guy. So, yeah.

Steve Hsu: Your example of GitHub is really one of the best because that, that even goes beyond the US China thing. Because if you look on GitHub, the barrier entry to write code to to develop software is, is, is so low. And the cost of hardware is so low now that, you know, you see names Eastern, lots of really strong developers come from Eastern Europe, from

Kyle Chan: Yeah.

Steve Hsu: from Iran, from you know, central Asia. It's like everybody in the world is able to count. GitHub is probably a very unique place in which everybody in the world can contribute. Unlike in say, molecular biology or physics. You need to have access to a very high quality home institution to be able to participate at a frontier on GitHub. Almost everybody can contribute. Some poor kid in India who just has a shitty internet connection

Kyle Chan: Yeah, yeah,

Steve Hsu: on those. They're coding on their phone or some hundred dollar laptop they can contribute on GitHub. So that's probably the best example actually .

Kyle Chan: yeah, yeah, yeah. So that's, that's an area where I just, if you see that world, like, I feel like a lot of people who are not in the software or coding world don't know about things like that. And if, if you do know about things you, you realize like, wow, there can be incredible collaboration worldwide.

And, you know, talk about democratizing the process. I mean, that is really like Yeah pretty mind blowing what you can contribute. You don't even know who, who likes, might be able to solve some bug or some puzzle that you've, you've been able to, unable to crack. And it could be someone from anywhere, at any level of education.

Steve Hsu: Yeah. Well that's a nice, optimistic place. Maybe we can end our conversation. So Kyle, thanks a lot for being on the podcast, and I wish you all the best.

Kyle Chan: Thank you. It was my pleasure.