Industrial Maximalism and Its Discontents: Dan Wang on US-China Competition – # 104
So I think that we have moved on from this idea that the Chinese cannot innovate. I think that idea is now decisively buried, and I am glad that we have buried that idea.
We’ve moved on to another idea, which is that, okay, the Chinese are much better at scaling, going from 1 to 100, whereas the Americans are still good at going from 0 to 1. And I want to suggest, no, let’s bury this idea, too, that the Chinese are able to both innovate as well as scale, that numbers are continuous. And so I think this idea of going from 0 to 1 as just exclusively the remit of the West is not empirical at this point.
And there’s also an idea of, you know, what does it matter if you go from 0 to 1 if you cannot go from 1 to 100?
Welcome to Manifold. It’s my pleasure to be here with Dan Wang at his home institution. We’re here at the Hoover Institution.
Dan, as you know, is the author of Breakneck: China’s Quest to Engineer the Future. This was a huge success. The book was listed as one of the best books of 2025 by the Financial Times.
I think it will become one of the go-to books for everyone in the United States or in the English-speaking world that wants to learn more about China and the competition between the U.S. and China.
Dan, welcome to the podcast.
Steve, welcome to the Hoover Institution. We are sitting in a seminar room at the library just to showcase our intellectual content. I’ve placed behind me a copy of the U.S. Industrial Outlook from 1991. This is the intellectual caliber you’ve reached, Steve.
A classic, classic. So I want to congratulate you for the success of your book.
Thank you.
Now, for the audience, I’m not going to try to do justice to the book because it’s quite a lengthy book, and Dan has been interviewed something like 70 times about his book.
Is that the right number?
Yes.
So I won’t try to rehash all the details of his book. I’m going to drill down on certain topics that I think my audience, the manifold audience, which is very interested in the U.S.-China competition and the development of China in the last generation — I think that audience already has a fair amount of background on this.
Dan’s book does a great job of introducing someone who isn’t an expert in this topic to the situation. But we’re going to try to drill down on a few topics that I think my audience is, and I myself am, particularly interested in. So I hope that’s okay, Dan.
Sounds great.
Okay, so one of the things, one of the themes in your book is contrasting the so-called engineering state of China versus the lawyerly society of the United States. And I think that’s a brilliant formulation, and it’s gone viral.
So just the other day, I was listening to a keynote talk that Adam Tooze, the historian, was giving for the London Review of Books, and there was a pretty big audience. It was a keynote address, and he spent a fair amount of time discussing — I don’t know if he mentioned you, but he used that exact terminology.
So you’re already affecting some of the leading thinkers in our society.
Affecting or infecting, Steve?
Affecting.
I would like to hear a little about your reflections on the book tour. So what were the things that surprised you? Were there any particular questions where people challenged you or changed your thinking on the topic?
I think the first thing I should acknowledge is that this idea of lawyers and engineers has become a memetic idea, but it hasn’t exactly been original to me. And I want to first acknowledge that this is sort of an idea that’s been more or less in the air for quite a long time.
We’ve had Bill Clinton quip in something like 1996 when he was in China to say,
“Well, we are governed by so many lawyers, you’re governed by so many engineers, we should have a swap.”
So I think what I’ve really tried to do was to take this concept and weaponize it and really try to create a bit of a framework in terms of thinking through China and write it in a way that fits my strengths as a writer, which is to write these very long essays in the form of annual letters, really to try to explain what works and doesn’t work in China.
And I think one of the things I’ve really wanted to do is to try to capture this fact that I would say that China, modern China today, is defined by two central facts, two central trends:
- The first is that it has lifted hundreds of millions of people out of poverty, a few of those people into the global elites, as well as the broader middle class. And that must be acknowledged that there is very real economic progress there.
- At the same time, the Communist Party has been repressive in novel ways, being an authoritarian power in the 21st century with all of the tools that it has, in ways that I think repress quite a lot of people and suppress their own human flourishing.
And so I think that is sort of the two facts I really want to get across. And as part of this book tour, I think what surprised me is the extent to which people are now quite curious about China. I think China is in the water, so to speak. Everybody has to have a view about China, whether that is something very social media driven like home prices in Chongqing—that’s been a weird mimetic trend to take off. You know, whatever aspect of weird industrial chemicals, that was a TikTok meme very briefly.
So the youths are on to this China thing. At the same time, the older elites are also very on to this thing. And so I think what I’ve been heartened by is that my book isn’t just being read by folks in the Bay Area where we’re speaking, not just folks in Washington, D.C., not just folks in New York, but also folks in Michigan, Ohio, Indiana, the broader Midwest, as well as the rest of the world. Because I think it is really important for all of us to be at least more curious about China.
I think Trump really did you a favor by starting this trade war and ensuring that every day on the front page of The New York Times and Wall Street Journal, there was some headline about China or U.S.-China competition. So definitely the timing of your book was exquisite.
Perhaps. And I think that one might have to say, “thank you, President Trump.” And maybe more of us will be compelled to say, thank you, President Trump, more on an ongoing basis now.
But I think there was something about the year 2025 in which we started the year with news about DeepSeek, moved on to news about electric vehicles, and then had more of the trade war, the 14th five-year plan. But I think that, you know, every day now there is going to be more and more China news. And frankly, that’s probably a good thing because
What is more important to the United States now than these two big trends, namely the rise of China in a more adversarial relationship, as well as the rise of technology, which has been something that I’ve been thinking about for the past decade.
One of the things I maybe not everybody in the audience is aware of is that your life really prepared you exquisitely to write this book and even prepared you for that specific formulation of lawyers versus engineers.
So when you were working in China, was it roughly 2017 to 2023?
- Correct.
You were working for Gavekal, which is an investment analysis firm. I mean, you were often analyzing companies in the chip industry, semiconductor industry. So you were very familiar with the engineering state, the technological development of China, the competitiveness of the products and companies.
But then I think when you were writing the book, you were in residence at Yale Law School.
- Yes.
So you were surrounded by the top legal minds in our country. And so you had both juxtaposed right before you, and that must have helped you formulate the ideas in the book.
- Yes. And I would furthermore add that my mother was a radio news anchor as well as a TV news anchor in Yunnan for the Yunnan Broadcasting Network. And so she has also prepared me exquisitely to speak to you today, Steve.
So I think that that is absolutely right, that I’ve been thinking about China and technology, working for Gavekal Economics, working for my rabbi, Arthur Kroeber, thinking very extensively about China’s developments in
- semiconductors,
- clean technology,
- manufacturing broadly,
living in Beijing, as well as Hong Kong, as well as Shanghai, throughout the entirety of zero COVID as well.
And after zero COVID fell apart in China, I moved to the Yale Law School where I was a fellow. And that was really the contrast that set everything up, that I lived through zero COVID, in which the numbers are right there in the name, no ambiguity about what zero COVID could possibly mean.
Thinking through the history of the one-child policy, which was in part heavily influenced by a missile engineer who was one of China’s top cybernetics experts, mathematicians at the time.
And then sitting at the Yale Law School among what was, I think, the self-consciously grooming America’s ruling class at a time when this was in 23, 24, when the Biden administration recruited very heavily from the Yale Law School. We had folks like
- Jake Sullivan,
- Gina Raimondo,
- Brian Dees,
who were all graduates of Yale Law. These were people who, I think, suborned the economists and tried to really assert their influence as lawyers to run a lot of policy.
And I think all of that came together very, very well to think that, well, you know, actually, the engineers and the lawyers, though it collapses quite a lot, no question, that is actually a pretty decent framework.
No less bad than socialist, capitalist, democratic, authoritarian to have another lens to think about this important relationship. Now, setting aside China entirely, as somebody who spends all his time talking to other scientists and talking to technologists in Silicon Valley, the idea that our nation should be run by lawyers is, to us, a shockingly bad outcome.
And, I mean, it’s something that people decry constantly in the circles that I move in. And, I don’t know, perhaps you disagree with me, but I’m curious how you think about that.
Yeah, well, there’s a brilliant politician who offered this quote that:
“Power isn’t something that is ever given to you. Power is something that you have to seize.”
Now, who said that? Was that Mussolini? No, that was someone far more ruthless. I mean, of course, Speaker Nancy Pelosi.
So, I think that, you know, if you are going to want to have power, no one is going to give it to you on a silver platter. And rather than, you know, decrying the lawyers for having seized all that power for themselves, I would say, why don’t we point the finger at the economists, let’s say, who had power tenuously, but then gave it up and they were really, it was really pried out of their fingers by the lawyers, I would say.
Why don’t we say to the engineers and scientists, can you not get better at organizing society? Can you not make a case better to the population and speak in a little bit more of a coherent and appealing manner than you presently do and try to get that power away from them?
So, that’s my challenge to the scientists. We can’t expect the lawyers to let go. You’ve got to seize it from them.
Yeah. It’s very interesting because we may have just entered an age where, if you think about it, so Trump is more of a business person, entrepreneur, although he certainly has had a lot of experience with the legal system. He’s not a legal mind himself.
The other people who are contending for power, like Elon Musk, they often control these huge platforms and can subtly influence the messages that are promulgated on those platforms.
So, we could be entering an era where vast wealth and control over these media platforms is what propels people to power, not Yale Law School.
Yeah. Well, I think the first thing to acknowledge is that Trump is an excellent, splendid product of the Lawyerly Society.
Yeah. He has abducted Maduro from Venezuela using, because this fellow is designated a narco-terrorist, charged by… And a machine gun owner.
And a machine gun owner.
And I saw some tweet about this. I’m not sure if it was actually true, but one of the charges against Maduro now is that he has been dispossessing native peoples, indigenous peoples, from land in the course of oil drilling.
So, you know, there is something, there is a legal pretense, a legal fiction upon a lot of things that Donald Trump tries to do.
I think a lot about some of these quotes. There was someone, there was a Latin American ruler who once said,
“Anything for my friends, for my enemies, the law.”
So, you know, law can be terrifying. And Donald Trump knows what lawfare is. He was schooled by Roy Cohn, who practiced lawfare extensively.
This is a man who, for whom lawsuits are absolutely central to his business career. He has sued totally everyone. He keeps suing people three times a day before breakfast. And he is very intent on, you know, flinging accusations left and right, intimidating people, and trying to establish guilt in the court of public opinion.
So the comment you made about the sort of trumped-up charges against Maduro is illustrative of why people who come from more of a science or techno-entrepreneurial background really hate the ideas, the idea of lawyers running society.
It’s because, for us, the number one thing is uncovering truth, uncovering how the world works.
So in the case of a scientist:
- What are the natural laws?
- What are the scientific or technical mechanisms by which things operate?
- How can I make a better transistor?
And for an entrepreneur, the truth is discovering:
- What does the market really want?
- What thing can I invent and scale and show the market that they want?
So for us, it’s mainly that concern, not making things up or making arguments regardless of the true reality of the situation, which is the way that we think about lawyers. And so it just seems very repulsive for me.
The most jarring thing that has occurred to me in dealing with people in positions of high power in our government and other governments who are trained as lawyers is, as a scientist, I see they’re often just not very anchored in the reality of what is true or not true in the world. And they’re just making arguments on whatever side they want to make the argument in favor of.
They’re free to make that argument and often free of the actual kind of scientific or rigorous thinking that we’re used to.
Yeah, I get where you’re coming from. And I think it is a position that is easy to be sympathetic to. But I want to rebut it somewhat with this idea that what is truth? And if we can discover the truth, what follows is not necessarily something that scientists are able to comment on very well. You say that entrepreneurs are trying to figure out what consumers want.
Well, is that a truth of what consumers want? That sort of process of trying to determine all of that does not seem quite as straightforward as measuring gravity or something, right? That is, consumers are fickle. The economy is a web of relations. There’s no commandments dictating what people want to do. And so that process immediately becomes very, very complicated.
Now, what we could have are scientists asserting themselves to say that global warming is true. Okay. And I think that is a statement that we can more or less say is a factual statement now. But what we ought to do with it, I think that doesn’t necessarily follow.
Does that mean that we need to cease all economic activity and lock people up so that no one emits any carbon? That is something that lawyers, as well as economists and many other folks, as well as all sorts of humanists, have to get involved in to figure out how do we resolve these normative questions? How do we handle disputes within society?
Because the Chinese are very capable of following science to its logical conclusion, which in the case of zero COVID was to essentially lock up the residents of its biggest city of about 25 million people, or to say that, well, we have this population crisis. The solution is extremely simple, one child per couple, which ended up being enormously disruptive to everyone.
So I would say that what we want is pluralism. What we need to have are some scientists and the ruling elites. We need to have some economists, lawyers, humanists, et cetera.
Yeah. I think that in navigating a rules-based system, principled system for determining how society should react to some discovered conclusion about reality, I mean, that is where you need a sophisticated legal system, a sophisticated system, civil society.
I think all those things are quite necessary and beneficial. It’s the commitment to our knowledge of how the universe is uncertain and we have to be disciplined in practicing our discovery of what turns out to be true or what doesn’t turn out to be true.
I think good business people are like scientists in this way that they can’t come in and say,
“I am sure people want to buy X. I am not going to give them X.”
They have to sort of look at,
- What features does X need?
- How can I change it to make it more popular?
So there’s a process of discovery that they’re committed to, which is an empirical process of observing what’s true in the world and reasoning based on that.
And I just find a lawyer that is practiced in debate and will adopt either position and throw themselves fully into that position without concern about whether the position is actually fundamentally true or false. That to us is very disturbing. That finding of truth is our sort of most sacred activity.
Yeah, well, what you call arguing out of both sides of your mouth, I can call empathy and understanding what the other side thinks. This is something that I think the Communist Party, for example, is not very good at doing. They do not have a good sense of how other people think.
And, you know, the Lawyerly Society has created some astonishing companies. The West Coast is the only region in the world that has created several companies worth trillions of dollars.
And maybe that’s a weird thing. Maybe things are overvalued. In fact, we can be sure that they’re overvalued. But I think that, you know, it doesn’t seem to me that the Lawyerly Society has been awful at creating companies, creating products, even though lawyers can argue any side.
Yeah, I think characterizing the U.S. as the Lawyerly Society makes sense in contrast to China. But as far as the way that the Google founders had to operate or the way that Jensen Huang had to operate, the law is part of the system that they’re in. I don’t know if it’s their primary concern or the thing that made their company valuable and Yahoo not so valuable.
It wasn’t necessarily that,
“Oh, this is a Lawyerly Society and that dictated the outcome.”
But certainly, I think property rights, well-defined rules for how businesses have to operate in society — all those things I think are important and maybe lacking in China at this point.
I think that’s a fair point.
Yeah. So I want to turn to a concept called industrial maximalism, which is promulgated by something sometimes called the Industrial Party in China. And in your book, you spend a little bit of time talking about this, and I think you actually investigated, I think you went and read some of the original documents or essays that started this movement all the way back in the early 2010s and mainly online. It was an online movement. It wasn’t started by the central government. It somehow made its way into the halls of power, but started really as an Internet phenomenon.
Some people who are watching the tech competition with China very closely, both from the Silicon Valley side and from the Chinese side, might say that this industrial maximalism idea has actually won out in the Chinese government. So the Chinese government, some people would argue, is actually behaving as if they’ve now embraced this idea of industrial maximalism. And so I just want to discuss that a little bit with you.
So maybe just say a little bit for the audience about what you wrote in the book about the Industrial Party and its history.
Yeah, the Industrial Party is not a legitimate political party because there are only, aside from the Communist Party, I believe there are eight other tolerated political parties in China, but all of them must be subservient and loyal and obedient to the Communist Party. So the Industrial Party is more of an online meme movement. Maybe this is one of these early memes from the Chinese Internet when it was a relatively free space before the Great Firewall really managed to slam down.
These were people who were essentially advocating for technocratic rule in order to pursue science and technology. These were people who have understood that China’s weaknesses stem almost entirely from backwardness in science and technology, that it was invaded brutally by Japan as well as partially colonized by Western powers. And the solution is to pursue not just the bomb, not just the satellite, but all sorts of important science and technology, and really to organize society, the entirety of society, to pursue these sort of things.
Now, there are some important canonical texts about this. One of them is something like, the title is something like, study or wash dishes. One of the most interesting ones is something called the Morning Star of Linggao, in which there is this online fan fiction community that imagined sending a lot of people from the present into something like the Ming Dynasty to industrialize the island of Hainan. And so, you know, it’s a kind of an interesting little read there.
At various points, the Industrial Party has been censored. And so, some of these people have been quite interesting. They were empowered to be major voices on the Internet. And I think these creative people did not always follow the dictates of the party. And when they spoke up, sometimes the Morning Star of Linggao was interrupted based on that fact.
I think that actually the canonical text of the Industrial Party was The Three-Body Problem, which is one of China’s best-known cultural exports now. And I love The Three-Body Problem. I think that it is a remarkable piece of science fiction. One of the underlying themes was that Liu Cixin said that it created a scenario in which humanity bands together under technocratic rule, builds these enormous spaceships controlled by the Navy in order to confront an extraterrestrial threat. And that feels like the sort of prescription that the industrial maximalists are really interested in.
So, for our non-Chinese listeners, I think it’s important to give some cultural context here, which is that for thousands of years, China regarded itself as at least the preeminent civilization that it knew of on the planet. And the reading of the past couple hundred years is the idea that some barbarians with superior science and technology came and inflicted terrible humiliations on Chinese civilization. And we are now recovering from that period, and China is reassuming its rightful place as one of the leading, perhaps the leading nation in the world. That’s sort of the background to all this.
So, I think if you’re a scientist or engineer in China, it’s natural for you to attach to science and technology as the thing which the barbarians used to beat us a couple hundred years ago, and the thing which we have to perfect now to restore our civilization to its rightful place.
Do you think that’s fair? That’s in the mind of like almost every Chinese person on the planet?
Yes, I think that is a fair reading. And I think that that is instilled into the minds of many. But I would offer two remarks. By the way, I’m not endorsing it. I’m just saying that that is a story that every Chinese person is familiar with.
Yes, that is a story that even before the Communist Party, the Chinese rulers have created this story. And the Communist Party has indeed put it into the heads of everyone. And I would want to offer a little bit of nuance here. First, if these were barbarians, how do they get all this great science and technology? So maybe there is at least a little bit something else here going on that these British, Dutch, red-bearded people were able to invent a lot of great things.
So was there something deficient in China that it wasn’t able to do? Maybe, maybe not. And the other part that I would offer is that, yes, China fell behind in what it referred to as the century of humiliation, in which these Western barbarians came over and seized major parts of China.
It made the people and the government suffer various indignities. And much more serious was the Japanese invasion, which was a brutal invasion by a fascist power that really ravaged a lot of the country and pushed the state into these pretty interior remote areas in order to carry on the fight.
And what I would offer is, you know, that was the century of humiliation. Yeah, pretty bad. I agree.
But what about the quarter century of self-humiliation that the Communist Party inflicted right afterwards, after 1949, when tens of millions of people perished after various landlord struggles, after the Great Leap Forward in which a famine ravaged the land? We had the Cultural Revolution.
Afterwards, there was this other spasm of the Wanchau policy. And then there’s some things that we don’t even speak so much about. You know, the strike hard against crime campaign in the 1980s. There were various issues with the Inner Mongolia Separatist Party that we don’t even think about. That created incredible ravages.
And that China was a society, after the Communist Party took over, that suffered these extraordinarily violent convulsions throughout society.
How do we explain something like that? Was that something that the barbarians inflicted upon the Chinese? No, I would say that that was something that emerged organically from them.
And so it is fine and good to think about the century of humiliation. Let’s also consider a little bit about the quarter century of self-humiliation.
Yeah, I mean, I think that the apologists for the current Chinese government, for the Communist Party, you know, they would say,
“We had to break a bunch of eggs and we made a bunch of mistakes to get us to where we are now. But now we finally emerged in our rightful place among the leading nations.”
I think one should not minimize the mistakes that were made, the terrible suffering, all that sort of thing. But I think from a nationalist perspective or even an ethno-nationalist perspective, you know, mistakes made by your own people, which inflict suffering on your own people, somehow—
I’m not justifying this, but I think this is the perspective—are more tolerable than things inflicted on you by some alien group of people. Right? I think that’s the psychology that governs this.
So they’re willing to overlook:
- famines and great leaps forward,
- all kinds of terrible things,
- one-child policies,
- things like this.
A miracle did happen that in a period of one, between one and two generations, it went from absolute levels of poverty, similar to the poorest countries in the world, to being possibly the most advanced technological civilization on the planet.
And so they’re willing, these people who are nationalists, are willing to forgive all of that based on where the country has now arrived.
Yes, and I think they are able to justify this and self-justify this. And I think that the rest of us don’t have to give in to some of that. And we can question whether self-inflicted harms is somehow morally better than other inflicted harms. And do the ends justify the means, whatever, an ancient philosophical question.
They would say yes, but that is something that we can also be—we can also interrogate and be critical about.
Yeah, I think from—I mean, you and I were both raised in—
And so from our perspective, the idea is that the greater country is one where one can critique what the government has done in the past, criticize it, try to open it up for consideration, even if it is embarrassing for the current regime or current leaders. And that’s a sign of strength of the country that that can happen.
You can have dissidents. The dissidents are allowed to air their views and people consider those views. And the country is internally stronger because of that.
I think one argument that people would make on the other side is that we are rising, speaking as if I were an industrial maximalist or Chinese nationalist or something,
“We are rising from the ashes. These guys came, they fed us opium. Japanese came and killed so many people. MacArthur wanted to use nukes against Chinese cities during the Korean War. We had to overcome all that. And we’re still overcoming.”
The moment we showed some ability to move up the value chain, the Americans tried to completely crush us.
And so when this is all over and we are on top and we’re no longer threatened, Threatened by the West, then we can relax and create this beautiful vision that you have for how society should be.
But right now we’re in the middle of a war and you don’t necessarily see the war. The war is being conducted in a very serious way, not for the average person to perceive, but for people who are watching carefully, this is a war.
This is the third world war, but it’s being conducted in a more subtle way. And I think that is what some of my listeners who belong to this party, the industrial party, that’s how I think they would articulate their position. That is a view that I wonder whether they can allow for any sort of relaxation afterwards.
Do you want to believe that you are in a war and a sort of a silent war in which the West is trying to suppress you? I think that is a reasonable reading of the evidence so far, especially over the past decade when the U.S. government really weaponized China’s dependence on semiconductors.
I was covering all of these twists and turns of U.S. export controls. I’ve been on net fairly critical of U.S. export controls. I’ve written several essays in the New York Times and in Foreign Affairs saying that this will, in fact, over the longer run, stimulate China’s technological self-sufficiency.
But what I would like is for a little bit more self-confidence from the regime. If China is, as you posit, the world’s leading technological civilization, a claim that I would mostly sign on to, they’re just starting to enter that phase. They’re not fully in that phase, but they’re just starting to.
There’s some questions about:
- When they can achieve semiconductor self-sufficiency or supremacy, whatever it is.
- Whether they can actually replicate all the capabilities of Airbus and Boeing.
But I think that it is a reasonable claim to say that China is, in many ways, technologically more self-sophisticated.
I just had an op-ed come out in the New York Times saying that China is winning the electrical age while Donald Trump is invading Venezuela for oil that the U.S. doesn’t need.
But, you know, in that case, why can it not gain a little bit more self-confidence and be a little bit more, let’s say, beneficent in a Confucian sense against the rest of the world?
If there is kind of this siege mentality, one of the things I think about as someone who thinks about not just authoritarian systems but Leninist systems in particular, in which I think that the Communist Party and its Leninist heritage—the leaders wake up every day feeling like this is a life-and-death struggle against the Western colonialists, capitalists, whatever it is.
And I wonder if they will ever feel like they can relax. I suspect not.
Let’s say that somehow the West is in disarray and fairly weak, that China can close the gates of the Celestial Empire and just be serenely untroubled. Do you really think that this Leninist regime can ever relax?
No, I think that these people are highly, highly paranoid. They’re trained to be paranoid. And what I want to do is to start relaxing now rather than some mythical end state that they will never feel comfortable with.
Yeah, I think this is the essential point. When China does achieve its full potential as a civilization—I mean, its economy, its technology—will they be able to transition to a different kind of government, a different sort of set of values for what’s allowed in society or not allowed in society?
You could have different views about how likely that is. Will they be so locked into this paranoid worldview that they won’t be able to get out?
And I want to say that this, by the way, this NATSEC (national security) perspective exists on both sides.
In the United States, we had this McCarthyist era, we were very worried about communism, we had the Sputnik scare.
So the question is, at what point do you get confident enough that you can sort of relax and stop locking up your dissidents and not think every visitor from this competing country is a spy?
Can they:
- Be locked into this paranoid mindset forever?
- Or could the world look totally different to whoever the new set of leaders are 30 years from now?
I think nobody really knows the answer to that. I think some people just assume China will fully recover from its bad few hundred years and then it can become just a normal country again, even though it is starting from what is currently a communist authoritarian system.
Other people would think, no. It’s very different. Like I was just listening to Dario Amadei at Davos talking with Demis Hassabis and saying:
“No, we cannot possibly give them NVIDIA chips because these people are so horrible. If they get AI first, it’s game over for the whole planet,” et cetera, et cetera. So I don’t claim to know the answer for which of those two scenarios is more plausible, but I think it’s the one, it’s what people should really be focused on, the future of a world in which China becomes possibly the most powerful country.
Yes. And I would say that, in general, the track record of America is that there have been these paranoid spasms. The Sputnik moment was a stimulus towards science in the US, but there’s been parts of the American berserk, in Philip Roth’s terms, that really deranged this country.
On the other hand, this is also a country that I think tried to build up Europe, tried to be kind or utilitarian towards Japan and Germany after it vanquished them.
That for the most part embraced, what we call the liberal international order by being open to globalization, hoping—now we say foolishly—that China will also grow into its own image, which was a bet that proved wrong.
I think we can say that now, but there, the episodes of paranoia also coexist with episodes of generosity in ways that I think the Trump administration calls “foolish generosity” at this point, given how globalization, in their view, has not worked out very well.
So coming to the current Chinese government and the nature of life in China, you’ve commented on the parity that’s experienced even by the elites.
So I guess the way I would describe it is: I think the average person in China is pretty happy with their government just because of the very strong growth and development that’s happened in the last generation.
I think if you question them carefully, they might say something like,
“Oh, there are certain topics it’s best not to talk about. We could get in trouble. Let’s not talk about that,”
and just change the subject. They have that awareness.
But in general, I think they think of their government as a good government and they have some confidence in it. I think I could be wrong about this, but this is my impression for sort of ordinary people.
I agree with that impression.
But then I think you make this insightful comment, both in some of your interviews and also in the book:
- For the elites, there is still the sense of participation.
- Because you might be doing well in your business because of some guanxi or patronage network.
- The person at the very top of that network, who’s some high party official, might get axed for corruption.
- Then suddenly your whole safety net or your whole system of power connections vanishes.
And that’s why you send your kids to Western universities, try to get as much of your capital out, buy a bunch of property in Vancouver and L.A.
So talk a little bit about what you think the world looks like for these relatively successful elites in China.
Do they not feel as comfortable about their future or secure in their future as, say, a wealthy person in San Francisco?
Who are Chinese elites?
Well, I think that in the state party context, these are people who are relatively high up within the Communist Party. These are relatively high up in the People’s Liberation Army.
Let’s check in: How are they feeling?
You know, right now, the Politburo has, I believe, now formally 23 members. The one spot traditionally reserved for a woman has been axed.
One of the two People’s Liberation Army Politburo members has been given the sack.
Potentially another Politburo member who is in charge of the weapons state-owned enterprises may also have fallen.
There’s this journal now that has tracked that the Communist Party has disciplined around a million officials last year, and this keeps going higher.
And, you know, the discipline process for a Communist Party official is extra legal. They have their own party system. It’s like being court-martialed in the military system here.
That’s kind of a grim fate.
And so, you know, that’s a million party officials disciplined out of a population of about 100 million party officials. That feels fairly substantial, both tigers as well as flies.
So, where do Chinese elites work outside of the party state?
Well, much like the U.S., some of them are working in:
- The financial industry
- The semiconductor industry
- The broader tech industry
How are they doing?
- A lot of the financial industry has been smacked around.
- They’ve had pay ceilings imposed on how much they can earn, about $300,000 if you’re working for a big state-owned bank, which on Wall Street, is what a second-year associate makes at some of these bigger banks.
- A lot of the tech sector has obviously been smacked around, especially if you’re working in consumer internet about five years ago.
- In the semiconductor industry, there’s been an enormous amount of graft that the party state has been keen to crack down on. And so, this is where I feel like much of the party state doesn’t feel very safe in Xi Jinping’s China. Xi Jinping himself sent his daughter to study at Harvard University. I did not go to Harvard, but his daughter graduated college in the same year that I did.
Among my friends who did attend Harvard, some of them knew her. So, if Harvard is not too good for Xi’s daughter, why should anyone else not send their kids to Stanford or Berkeley or Michigan or wherever else?
When people have the ability to acquire an American status, passports, university education, they generally tend to seize it. It is exceedingly rare for Americans to want to feel like they have to do the same thing.
I think it is still relatively rare among people in developing countries, given a choice between spending much more of your time in China or the U.S., that they still, on net, a little bit more, want to spend their lives in the U.S. That sentiment, I think, is easy to find.
If I were talking to some people who grew up in China but are working in the United States and I stated the case you just made, I think all of them would be familiar with it. I’m not sure how many of them would really strongly feel that they are better off, safer, or more secure in the United States than in China.
I know a lot of people who have the opposite view and feel that now is the time actually to go back to China because the companies there really are doing cutting-edge stuff. They wouldn’t sacrifice their intellectual development by being at Tsinghua versus being at Harvard or something like this.
I think it’s quite complicated. I don’t actually know the answer myself, but you could easily find people who don’t necessarily have the psychology you described, but there are many people who do have that psychology. I think that’s a key variable in what’s going to happen to the country.
- Are they really going to lose 50% of their top people to this?
- If they only lose 10% of their top people to this, then they still have an overwhelming, potentially, advantage in human capital over the United States.
So I think it’s something to monitor very carefully—the mood of the most able, dynamic, entrepreneurial people in China. Like, are they better off staying there and building that economy, or are they better off trying to get out?
Yeah, I think that’s a great observation. It’s hard to monitor them. We can find plenty of examples of people who’ve decided to move to China over the last few years.
You can also cite that, I believe it was Matt Sheehan who did this work, that 90% of the major AI researchers who published in big papers over the last couple of years are still in the U.S. and have declined to move elsewhere.
We can find numbers for both of these. But maybe, Steve, you should be the sharp tip of the spear in trafficking.
I try to spend a fair amount of time in China. Then inevitably, when I’m at one of the big labs here, I bump into a bunch of people who have China backgrounds but maybe got their PhDs here or something. I’m constantly interrogating this question, trying to see how people feel about things.
It may vary by age. It may vary by the family background of the individual.
A funny story I think I told before on this podcast is that I was visiting DeepMind. My host, who’s actually, I think, a Google fellow or has some pretty fancy title, was showing me around. I was there to talk about the use of AI in physics, using the current best AIs to help physicists do their research.
This guy was showing me around, and we encountered a group of three or four obviously Chinese AI researchers, very young people who were walking through the pavilion down at Google where they’ve put all the AI people in Palo Alto and Mountain View. Everybody’s young, and you hear a lot of Mandarin at all these labs.
My host, who is not Chinese, introduced me to these young people and said,
“Oh, this is Professor Xu. He’s helping us make our AIs more useful for physics.”
I was shocked because, first of all, I’d encountered this in Japan and Korea, but not in China. All of them bowed to me slightly, which I thought, Is that coming back? Like, students bow to teachers? Because I had not experienced that in China in the past when I visited.
But anyway, they all kind of bowed to me, and then one of them said very sincerely and naively, like a young nerd:
“Professor Xu, if you find any problem with our model, we will make it better.”
It was so naive and sincere, the reaction of this kid. So anyway, but it was like a cluster of Chinese. And these guys were involved in the most hardcore model training, pre-training of the models at DeepMind and Google.
So, like, I would love to pull them aside and say,
“Well, why are you here? Why are you not at Tsinghua? Like, what are your classmates doing?”
You know, so I’m trying to monitor that kind of situation as much as I can.
Well, I would love it if more of these, your students at Michigan State bowed to you much more actively, Steve.
What a nice idea.
So give us a sense of the buy check of the moment.
How many of these talented Chinese origin researchers that you’ve been speaking to on your big tour of Silicon Valley, how many of them want to move to Tsinghua?
You know, my best spy when it comes to this is a Chinese-American kid who grew up in California, but unlike me, actually mastered the Chinese language.
So he’s fully bilingual and after doing his undergrad degree in the UC system was asking me where he should do his PhD in AI. And so we were talking about the strengths and weaknesses of various programs, and he ended up going to Tsinghua.
So he’s actually a grad student in Tsinghua now, and he regularly reports to me about what the scene is like for these huge numbers of kids going through Beida and Tsinghua and other universities there and working with the companies.
So here it’s a little funny because the big labs aren’t that well integrated with academia, and academia is a little bit isolated right now from the frontier developments. But I think that’s less true in China.
So a lot of the academic groups are collaborating tightly with ByteDance and some of the other companies that are building frontier models.
So he reports to me that the scene in, partially for this reason, because he’s on the academic side, he’s doing his PhD, the scene in Beijing is more vibrant than the scene here on the Stanford campus.
Because most of the Stanford people are locked out. They’re not actually able to do this frontier-level work at Google or at OpenAI or at Anthropic.
- The companies are doing their thing, and
- The academics are doing their thing.
But in Beijing, there’s a very fertile mix between these groups.
And so he keeps thanking me that I didn’t encourage him to go to Tsinghua, but I gave him what I thought was a realistic view of what he would find there. And he’s quite happy there.
So it’s only one data point. It’s a very important thing to just keep your finger on, to understand where is the talent.
Yeah, and the people that you’re speaking to here, how many of them want to keep actively going to China?
Yeah, it’s very interesting. Will they work for ByteDance?
Yeah, so after my stay here in the Bay Area ends, I’m actually going, I will be visiting Tsinghua for a while.
And my host is a young, new assistant professor at Tsinghua, whose prior position was as a postdoc in the U.S. at one of the top universities. And he’s quite happy to go back.
I don’t know. I mean, I think, of course, he’s super nerdy, so he probably doesn’t think at all about precarity or getting on the wrong side of the Communist Party.
He’s just thinking about,
- “How many smart students can I get?”
- “How much research money do I have?”
- “What companies can I collaborate with?”
But I don’t think it’s a completely zero-one kind of, you know, one-sided decision. Some people could decide in one way, and some people could decide in other ways, maybe even because of food.
Some people might just be like, “Wow, the food is so much better in Beijing than here,” or something.
So, and the research stuff is kind of comparable.
So, who knows? But I think it’s something we should all keep tabs on.
Let’s switch gears now.
I have a bunch of quotes, which I loved, that came from your, I think the ones I wrote down came from your annual letter.
And so, for my audience, maybe, I’d love to hear a comment on this.
Maybe one of the reasons Dan’s book was so successful is that he already had a huge following among all the people who seriously think about China and U.S.-China competition through the annual letters that he wrote.
And you always write these at the end of the year during the holidays, right?
And how many years were you writing that letter?
“Eight now.”
So, I mean, I’m sure at the beginning, maybe not that many people were reading it.
Only my mother.
Okay, only their mother.
But by the end, I know you skipped 2024, right?
But by, say, 2023 or 2025, like, everybody who is seriously thinking in this space is reading his annual letter.
They’re beautifully written.
They’re not short.
These are long, kind of discursive essays.
And one in 2025 I really enjoyed.
Do you think that helped your book when it came out?
People already knew who you were and already were familiar with your thinking?
I think that it was a slightly higher stakes letter to write because this was a, I have a book audience that is onto me now.
Yeah.
And so I’m speaking to a broader array of people who are not necessarily thinking about China.
And what I really tried to do was, for my book, was to reach the lawyer in Ohio or Indiana.
And so, now that these people, if these people are interested in my work, I wanted to hit them with a really big annual letter.
This is also an annual letter that is more driven by Silicon Valley, where I spend a lot of time now, and not just China.
Yeah.
So let me quote, I think these quotes, I think I got them from you, not from the book, but from the annual letter.
And I just want you, just feel free to riff on, you know, extend what you said in the quote, or make any comment that you think is appropriate.
So here’s a quote:
“Probably the most underrated part of the Chinese system is the ferocity of market competition. It’s excusable not to see that, given that the party espouses so much Marxism. I would argue that China embodies both greater capitalist competition and greater capitalist excess than America does today.”
So, beautiful.
Yeah, I think that, you know, there’s a lot of ways in which the tech companies in Silicon Valley feel they are very, it’s like a gentleman’s club of, you know, playing with, against each other.
Where, you know, Google’s got search, and then Amazon’s got e-commerce, and Facebook’s got everything else.
And so, they don’t really tread that hard against each other’s toes, whereas in China, everyone is fighting everyone over everything.
And you have these, I like this chart that Kyle Chan of High Capacity has created.
We got these big interlocking circles that are highly complicated.
And I think that just really shows how everybody is up in each other’s business all the time.
Whereas in the U.S., at least among the tech companies, it is much more gentle.
You know, we’ve got union protections here, where you don’t have unions, really, in China.
They arrest the Marxist organizers.
And so, you know, what sort of socialism is this?
So, you know, Peter Thiel was famous for saying that competition is for suckers, for losers.
“What you should try to do is get a monopoly, and that’s how you really create a trillion-dollar company.”
And one critique of our lawyerly society here, or our political system, is that business can co-opt government to the point where they can actually end up with a monopoly, and government doesn’t do anything about it because they have enough influence in Washington.
And whereas in China, that’s not going to happen, the government likes to see these companies beat each other up and, of course, drive the profit margins to zero, maybe to the benefit of the consumer.
Do you think that’s just a caricature, or do you think that’s a fair picture of what’s going on?
I think that’s pretty fair.
If we take a look at the Chinese equity market, I mean, that’s – this is not the whole reason, but part of the reason that, you know, the Shanghai Composite Index has mostly trended sideways for 20 years.
To my dismay.
To your dismay.
Well, I’m sorry about your losses.
But, you know, maybe it should have been –
Or lack of gains.
Investing in America, Steve.
I do both, but –
You do both.
So, I think that is part of the reason that, you know, profits are just much lower in China because it’s much more competitive and a bunch of losers, maybe.
But I think it’s great to be a consumer there.
Yeah.
That’s my – well, for Americans who’ve never lived in China or spent time there, buying power you have there is just unbelievable.
Yes.
Like, dollar for dollar, what you can buy, not just food, but electronic gadgets, cars, whatever.
Services.
Yeah.
It’s just insane.
It’s insane.
Which is another reason, actually, I won’t say his name, but one very prominent person who is in Beijing, big venture investor now, was a very prominent technologist in the United States for one of the big tech companies.
He just says, like,
“I can’t have this quality of life in the United States, even though I’m a billionaire — he’s a billionaire — even though I’m super wealthy, I can’t have this quality of life in the U.S. because I get on my phone and do this, and some really delicious food appears 20 minutes later delivered to my office.”
And, yeah, he says, I can’t have that.
Yeah.
Lower labor costs is a real thing.
Yeah.
Next quote:
“Beijing has been preparing for Cold War without eagerness for waging it, while the U.S. wants to wage a Cold War without preparing for it.”
Let me reply to your quote, or rather my quote, with another quote.
This is an apocryphal quote attributed to Viennese satirist Karl Krauss:
“In Berlin, the situation is serious but not hopeless. Whereas in Vienna, the situation is hopeless, but it is not serious.” And this is where I think a lot of the U.S. is, where they are just not a serious country in a lot of different ways.
Now, is China eager to wage Cold War?
I can say maybe they are. But I would also say that the balance of evidence has been that China grew very, very rich under an international system in which it became a major trading power without a large navy. That might be unprecedented.
You know, you didn’t, the U.K. was a major trading power within a large navy. So was Portugal. So was the Netherlands. So was the United States. China didn’t have to build a big navy in order to become a big maritime power.
The U.S. gave it to them for free, right? And so it prospered by any measure under the system. And that system has kind of turned against it. And so it is reacting. That’s all fine.
But China has mostly prospered under this system. And so, but it has also been quite cautious and preparing for energy self-sufficiency, food self-sufficiency. I spent some time in my book examining something called the Mayor’s Vegetable Basket Program, in which there’s incredible amounts of farmland around big cities like Beijing and Shanghai.
As soon as you take the high-speed rail out, you get to hit farmland really quickly. And that’s in part because the mayors, not the party secretaries, but the mayors, people responsible for logistics and operations, have to manage grains self-sufficiency, as well as various parts of vegetables and meat self-sufficiency as well.
And there’s obviously now a drive towards technological self-sufficiency.
Yes, the U.S. right now is obviously floundering to escape a system that it has very substantially built, that it is trying to figure out how to subdue China in various ways. Yeah, but it is unwilling to pay a lot of costs in order to get to a much better state.
And so, that’s why I think there’s parts of America that want to wage a Cold War without really doing enough to prepare for it.
Yeah, when I mentioned that a lot of my listeners are, in a way, members of the Industrial Party, I didn’t mean the Industrial Party of China. I meant they’re members of the Industrial Party of America. So, lots of entrepreneurs, Silicon Valley people, are astonished that we can’t seem to get our shit together.
We talk about reshoring manufacturing to the United States. The hard work of really doing it, the government doesn’t seem, or society doesn’t seem serious enough, really, to undertake that heavy lift. And it might take decades for us to get there. It’ll take decades of sustained effort. And I’m not sure, I’m not, as an American, as an American who grew up in the Midwest, I’m not sure that we still have it in us to do that. I hope we do.
Yes. Let me read a few things about innovation and manufacturing in China from you. I’m going to read a few of them and then you comment.
- “By failing to recognize China’s real strength, the industrial ecosystem’s pulsating with process knowledge, the U.S. is only cheating itself.”
- “Chinese workers innovate every day on the factory floor.”
- “Western elites keep holding on to a distinction between innovation, quotes, which is mostly the remit of the West, and, quote, scaling, which they accept that China can do.”
- “I want to dissolve that distinction.”
So, I think that we have moved on from this idea that the Chinese cannot innovate. I think that idea is now decisively buried. And I’m glad that we have buried that idea.
We’ve moved on to another idea, which is that, okay, the Chinese are much better at scaling, going from 1 to 100, where the Americans are still good at going from 0 to 1. And I want to suggest, you know, no, let’s bury this idea, too, that the Chinese are able to both innovate as well as scale, that, you know, numbers are continuous.
And so, I think, you know, this idea of going from 0 to 1 as just exclusively the remit of the West, I think, is not empirical at this point.
And there’s also an idea of, you know, what does it matter if you go from 0 to 1 if you cannot go from 1 to 100?
So, I cite the example of, you know, who was it who, you know, which American lab invented solar panels? Bell Labs, 1957 in New Jersey.
Where is Bell Labs today? Nowhere, because this thing is gone.
Who owns the entire solar industry today? It is the Chinese. And they own not only the final assembly of the modules, but also the polysilicon processing, as well as the equipment to produce all of the solar wafers and modules.
So, you know, this is something that I think that, you know, let’s not cheat ourselves out of the idea that Chinese cannot do this or cannot do that, that they’re fundamentally constrained because of the nature of their political system.
You know, the political system has gotten pretty far in terms of producing a lot of technology.
I believe that authoritarian powers can do incredible science and technology. We saw this with the Soviet Union. We saw this with Nazi Germany, in which totalitarian systems produce miracles in, let’s say, spaceflight, as well as many other things.
So, I think that the more that we can get our discourse and our level of competition to the level of sustained capacity building, which is what the Chinese have done, rather than thinking, “oh, well, they can do that, which is not that difficult, and we can do this, which is much harder,” let’s acknowledge that they can do a lot of difficult, important things, and that the United States should also get in this game in a much more active way.
If you take the normie boomer, sort of outdated view on China, okay, we could have different stages:
- One stage is, they only make crappy low-end stuff.
- They can’t do high-quality, cutting-edge manufacturing.
I think by now, people, even most normie boomers, have kind of set that one aside, and they say, “oh, well, they can do advanced manufacturing.” But the core idea, the zero-to-one, brilliant, genius breakthrough, that still has to come from the West.
And I think you’re saying, okay, even that misconception should be set aside. I’m curious, in your 70-odd interviews that you did, or addresses to big audiences of serious people, what fraction of the population is ready for that step? Are they already there in understanding that, or are they still resisting that step, or clinging to that illusion that the zero-to-one step can’t happen in China?
Maybe half-half. I think that there is still plenty of clinginess among the quote-unquote serious people. But I think that there is also a readiness.
I mean, it starts sounding a little bit absurd. “Oh, well, they can make all the toys and T-shirts, and they can make the large-capacity batteries for the EVs, but they can’t.” You know, these sort of things, statements, don’t really make sense to me.
I think that, you know, 2025 was an important year in terms of the DeepSeek moment. Maybe that scale of success was overstated, you know, there was this big market reaction to deep-seek, and that was a little bit hard to predict based on the release of a paper.
But, you know, EVs are definitely taking over, and, you know, we don’t have to over-index our expectations of the economy based on EVs alone. Transportation equipment is not a giant sector of the American economy, even though the auto sector is politically important.
But I think that people now are ready to hear this, especially given the moment we’re in with Donald Trump eroding a lot of alliances, making the world question American leadership. I think that it is good for there to be a little bit more worry about what is the American position and try to be a little bit more clear-eyed about what are American weaknesses.
This question of whether, in a less free political atmosphere, scientists can’t make big innovative leaps. I still hear that.
Yes. Do you still hear that?
Yes, and I think this is one of the sins of the lawyers, that they think that, “oh, we need so much free speech in order to produce scientific breakthroughs.”
In my book, I spent a bit of time thinking about the example of the Soviet Union, and there’s a particular historian of science from Harvard named Lauren Graham, who studied the Soviet science system. She pointed out that there were two big cataclysmic events for Soviet scientists:
- The first was the terror of the Stalin purges throughout the 30s.
- The second was the dissolution of the Soviet Union throughout the 1990s.
Which event was worse for scientists? Actually, it was the dissolution of the Soviet Union, because that was when the money stopped. A lot of scientists had to close their labs and leave.
Where, yes, very strangely, throughout the 1930s, under Stalin, there are many, many examples of scientists who barely staggered out of gulags and then produced their Nobel Prize winning scientific innovations.
And so, you know, part of this is because Stalin funded science. And when you throw in a lot of money at science, you’re probably going to get a lot of science. At least this was true about 100 years ago. Maybe not so true today.
And there is also this view that, you know, why were the Soviets so good at pure mathematics, as well as chess, as well as a lot of science? Well, potentially, they treat doing science as an intellectual escape when they do not, cannot enjoy broader political freedoms.
So, I don’t think that, you know, authoritarianism is good for science. I definitely will not go that far.
What I will say is that money is good for science, for the most part. And right now, the Trump administration is cutting somewhat some scientific funding. University budgets are on the back foot. But the Chinese are funding a lot of science, and I think they’re going to get the science.
Yeah, I agree with you.
So, I finished my PhD in 1991, which coincided with a huge influx of former Soviet scientists to the West.
Right. The best of the best from the Soviet Union, former Soviet Union, came to the United States and competed with young Americans like me for the same jobs. So, it was a very, very vivid experience with me, and I spent many, many hours in so-called Russian seminars.
Our seminars are usually scheduled, oh, it’s an hour, and if a few people want to stay longer and talk to the speaker, they hang out, but everybody else leaves. The Soviet system, they talk for hours, and they fight until they get to the bottom of the matter. So, even the style of Soviet physics and mathematics I was very familiar with. When you talk to those individuals, you get to know those people after they become professors here and collaborate with them. They had very, very fond memories of the Soviet system.
It was a system. There were even cities, if they were adjacent to the weapons program, they would live in a secret city where they didn’t have to worry about food, they didn’t have to worry about salaries, everything was provided, and they could just focus on the science. And they had a chance to have an orchestra and play chess in their free time or write poetry. So, really did not interfere.
They knew there were certain things they were not supposed to do. They should not suddenly write a tract about why the wheat harvest was so terrible in Ukraine last year, right? Okay, avoid that topic, but if you want to learn about electrons in super lattices, no problem. So, they always felt very nostalgic for that system, and then what happened is when that system collapsed, just as you said, they literally couldn’t eat.
They had to grow potatoes in their front yard in order not to starve and just plead with Western colleagues to send them a few dollars. So, they could actually survive, you know, so they could actually survive, you know, in the Soviet Union. Eventually, they all left the Soviet Union.
So, this idea of connecting the political system to the productivity of the scientific effort is a very tenuous connection. I even wonder why that system took place. Like, how did we forget the Soviet Union produced a lot of amazing math and science? How do we forget about Nazi Germany?
I think most people don’t have direct experience with how scientific breakthroughs or technical breakthroughs are produced. So, it’s always a great remove. Like, you read about it in a textbook or you watch a movie about it. But for the people who really see how it’s done, the Aspie genius who’s going to improve your large language model, transformer model, that guy’s not potentially thinking about any political things. He just wants to actually sit in his office and do his own thing. And he could do that just as well in Beijing as in Mountain View.
Yes. But I do want to say that I think that Aspie genius would do even better in a system where he was allowed to criticize the Ukrainian harvest. We can state a minimal case in which authoritarianism does not have to defeat science. But I think we can still embrace a stronger case that freer societies can get better science because the American science ultimately was better in probably most ways, with some exceptions.
Now, there is a more extreme version of this hypothesis, which, for example, people from the French system and people who really experienced the Soviet system will advocate for it. I don’t necessarily hold this to you, but they will advocate for it.
For example, the French really outperformed the Americans in winning Fields Medals. It’s because they don’t have a system that glorifies money and becoming a billionaire. They have values. They have cultural values. Mathematics is one of their things that in their society is highly esteemed.
And a kid here who is really talented at math is thinking to themselves,
“Well, how can I make a lot of money in cryptocurrency when he’s 17 years old?”
That doesn’t happen so much in France and it didn’t happen so much in the Soviet Union. And so they claim that on a pound-for-pound basis, their math and physics was better because they were not corrupted. I’m not talking about intellectual freedom or political freedom. They were not corrupted by capitalism, by materialism. And that is an argument that some people, even French scientists today, will still make.
Well, France is interesting because they’ve had several leaders with mathematics degrees. Yeah. Like even modern senators. And that is interesting. Yeah, the prestige in France from having a mastery of mathematics is quite high still.
Here in the United States, it’s not so high. Although maybe it’s coming back with AI and crypto and stuff like that. I don’t know.
I doubt it, Steve. I don’t think mathematicians will be cool. I can only dream. Let me throw one more quote at you. I think we still have a little bit more time. This is from your letter.
I sometimes hear that the U.S. will save manufacturers through automation. The truth is that Chinese factories tend to be ahead on automation. That’s a big part of the reason that Chinese Tesla workers are more productive than California Tesla workers. China regularly installs as many robots as the rest of the world put together. They are also able to provide greater amounts of training data for AI.
We have to be careful not to let automation, like superintelligence, become an excuse for magical thinking rather than doing the hard work of capacity. There’s all these wonderful claims now that AI, AGI, superintelligence, whatever you want to call it, is going to solve all of our problems.
- “It is going to solve our scientific development.”
- “This is going to cure cancer.”
- “This is going to fix our industrial base.”
And, you know, I fear that before it cures cancer and rebuilds the industrial base, AI is simply going to derange us all. People are getting fooled quite easily on the Sora 2 with these video generation tools. How can you tell what is an AI video today? Like, are there even actually any watermarks here to say? Hard to say.
There’s teens with their AI companions. I feel like social media is already trolling me to insanity without this super intelligence hamming that up. So, I’m worried about AI. And I think that, you know, there’s all sorts of opportunity, but it seems like there’s also a lot of threats here that I don’t really know how to assess.
What’s your view on AI?
Steve, when you think, when you are in this, you’re doing your tour of Silicon Valley, you’re living in Berkeley this month. When engineers come up to you with these really apocalyptic claims about AI or pure utopia claims, how do you personally react? How do you assess it? How do you maintain epistemic hygiene and sanity? And I want to know your mental toolkit for dealing with these claims.
Yeah, I think this is the question of our era. And I think this is ground zero.
You know, on the one hand, on Saturday, I’ll be at a dinner, I’ll give a talk, and there’ll be a hackathon at AGI House, which is a huge mansion that’s been rented by these AI researchers. It has a view, it’s a twin to the Twin Peaks neighborhood of San Francisco, and has this unbelievable view of the bay looking to the east.
On the other hand, at Lighthaven in Berkeley, where there’s an assembled crew of rationalists who are all doomers and worried about AI killing us, they have a beautifully manicured campus where you can sit outside and discuss things in a philosophical tone.
So here we have both poles of the most optimistic, abundance-focused view of what AI is going to give us and the most doomerish existential risk view. And those two communities are living maybe 10 miles apart from each other. So this is the place to examine those views.
You know, for me, I would say we’re close to spending 2% of GDP now on CapEx investments for data centers and model training and NVIDIA chips and things like this. And back to your quote, the assumption that that will suddenly make us competitive in terms of industrial robots or automation in our ports or assembly lines where cars are made, things that China is way ahead of us on, I don’t see that the AI that we’re currently developing is necessarily going to close that gap automatically in some magical way.
I agree with you. It’s sort of magical thinking on the abundance side to think that these large language models are suddenly going to solve all those problems for us. So I think that’s pretty unrealistic. I think that’s a problem for U.S. strategy. If you’re sort of a hard power person thinking about U.S.-China competition, I don’t think these AI investments are going to, at least for a while, close these gaps. They’re very concerning.
As far as the long-run existential risk from AI, I think the people who are afraid are not wrong. I think in the long run, if we do create things which are more intelligent than us, more powerful thinkers than us, but also eventually have their own desires and wants and maybe even a kind of consciousness, that it will be hard eventually for them to displace us in some way.
I think there is a real long-term existential risk. I don’t think it’s as proximate as some of these people in Berkeley think. I think it’s going to take longer. I think there could actually be a period of human flourishing where we do create these great intelligences, and for a while, they are harnessed to our needs. They are well-designed. They are aligned in some sense.
That alignment, I think, can’t be guaranteed in any rigorous mathematical way, and so in the long run, they could diverge, and we could create godlike things that eventually don’t care about us and accidentally smush us. And I think that’s the argument that the Berkeley crowd would make.
What is the long run? Is that a matter of two decades or two years? Very hard to predict because the problem is that, and this is the key thing everyone’s on the lookout for, is when do the AIs become so good at AI research and software writing that they start improving themselves and at a pace where human engineers can’t quite follow the way in which they’re improving themselves?
I don’t think we’re close to that right now. I think I would guess we’re probably five or ten years away from that inflection point. Other people have much more aggressive timelines than I do. Yeah. It’s hard to know. It’s one of these things. It’s like, oh, what is it? You know, if China does succeed in becoming wealthy and fully developed, what will their government system evolve to? I literally don’t know the answer to that.
And similarly, I do think AI is definitely going to continue to advance rapidly, but I don’t know the timescale over which it will become threatening.
Do you believe that there will be a point in which it starts self-recursively improving?
Yes, I do.
Okay.
I mean, this is one of these hard things where, you know, it is really easy for, you know, once you look at all of these exponential curves and we see them and, oh, it’s like a log scale is a perfectly linear, you know, it becomes really, really difficult to think of anything but, you know, the extent of these curves.
And, you know, I don’t know if the saying to do, saying thing to do is to just rule out the idea of, you know, the idea that this will self-recursively improve because once we’re in that scenario, then all bets are off, right? And how can we predict anything?
It’s like the first order impact of AI is hard enough to understand. And then we have to like think about the 17th order impact. That’s just way too hard, right?
And so this is where I’m wondering whether, you know, this sort of totalizing aspect, it is occluding our view of what is really important. And maybe it is rational because there’s plenty of smart folks who are telling all of us that we’ll get to this point of recursiveness. And I don’t know, it feels like it is just going to be, it’s hard to come up with the epistemic tools really to deal with that argument.
So, I was in a war game held at the Tate Modern, this museum in London, and the war game was sponsored by people who are concerned about AI safety. The British government funded some of it, and some of the big AI companies funded it.
My team was the China team. I was a leader of China in this. It was literally a war game where every turn is like a year and you’re making decisions about:
- how your society is going to allocate things
- what you’re going to do to other countries during that turn
The head of the U.S. team was a guy called Stanley McChrystal, who was a prominent former general. Very serious simulation.
The outcome was though very optimistic because what happened is that when we, in our respective countries, saw this recursive capability start to appear, we negotiated a slowdown and mutual inspections of our AI labs. So the American side could send AI scientists to see what was going on in the Chinese labs and vice versa. The game ended with us slowing down the development at a place where we might lose control over the AI.
Now, just before I came over here, I was listening to, as I mentioned earlier, an interview of Demis Hassabis and Dario Amodei at Davos, which I think is happening right now. They both said, they just kind of mentioned during their conversation, which is mostly not about this stuff, but a little bit about this cataclysmic type development.
They both mentioned that they’re in communication because they’re both on the lookout for this, and they are competing against each other and they’re competing against Chinese labs, but they are watching for this development. They do have in mind that they are going to start talking to each other when they start to see it. And hopefully I think it would just be wise that if we start to see that we slow down.
Now for the doomers, this doesn’t solve the problem because the doomers will say:
- “Oh, but there’ll be some lab in China that doesn’t stop.”
- “Or there’ll be some dudes in UAE who have a big data center and they won’t stop.”
So it doesn’t completely solve the problem, but there is a plausible future where we delay the real catastrophe for a long time by being smart about this process.
Well, if we’re sure that the catastrophe is coming, why don’t we slow down now?
Those guys were actually interesting. So very, very interesting remark because both Dario and Demis said in this conversation, they would both be happy if they could slow down. They both said that.
- Dario has a one to five year timeline
- Demis has more of a five to 10 year timeline to get to this point. And they both said, Dario said, “I actually prefer your timeline. I wish it would — I hope it’s going to take five to 10 years and not one to five years.” They both said, “Yeah, we’re both more comfortable if we can slow this down.”
So, but it’s a collective action problem. It’s a little bit like global warming. Like, can people actually put aside their narrow self-interest to cooperate to avert this catastrophe? I don’t know.
Well, global warming is challenging in part because there are so many countries that have to coordinate together who have to trade off against economic benefits, presumably with coal.
But why don’t we send General McChrystal to start inspecting those labs right now? Because we can do a lot of that right now.
When they designed this war game that I was part of, the people who designed it are actual specialists who designed war games for the Pentagon and also for the British defense ministry. They were honest.
So they designed the dynamics of the game to be as realistic as possible. However, the makers of the documentary were hoping that the ending would be an optimistic one where the competing sides could cooperate.
And the view is like, this provides a template because a movie version of it is very real for the ape brain to see humans enacting the story. It’s a narrative, right?
And the hope is that this will plant the seed in the Chinese leadership, the U.S. leadership, and the corporate leadership that yes, we do need to be ready to pause when things start to get dangerous.
So I think it is in everybody’s mind, whether it’ll, it could still be a tragedy. It could still be a runaway arms race, but I’m optimistic.
Actually, narratives matter. We are all ruled by ape brains. And so this is why the lawyers are in charge, and the scientists and engineers need to get better at narrative.
Very good. Well, that’s a good place for us to stop.
So thank you very much, Dan. It’s been a wonderful experience. Hope my audience enjoys this.
- Read his book
- Read his annual letter
We’ll put some useful links in the show notes to some of these things like industrial maximalism and other topics that came up.
Thanks very much for your time.
“Thanks, Steve. This was wonderful.”