マット・イグレシアス＋エルサレム・デムサス＋ジョーダン・シュナイダー 特別コラボ回

The Arguments, a podcast with two legends in the game, the founder of The Arguments, Jerusalem Demsys and Matt Iglesias. We have a, frankly, like, pretty unhinged conversation about American politics, Chinese politics, artificial intelligence, corruption, and Kanye West at the state dinner. I am really excited to share it with all you guys. Thanks for having us. Hi, everyone. My name is Jerusalem Demsys. I'm Matthew Iglesias. And this is the Argument Podcast. This week we are joined by China Talks' Jordan Schneider. Jordan, welcome. So great to be here. Two legends in the game. I'm honored. I've finally reached the big kids table. And now a third legend. Look at this. We're expanding. This is legend abundance. That's what we're doing here. Yes. We're going to talk to you on the podcast, A, because China apparently is a very important country with the care about it. A lot of people living there. Yeah, as navel-gazing Americans, we try to only stay focused on, you know, our own neck of the woods, but occasionally the other necks of the woods come here.

So she just recently came to the U.S. What was that about? Well, you know, Trump and Shea are buds. And look, like when you run countries, instead of like meeting up at the country club, you just like meet up at each other's capital city to like have a nice dinner, listen to some Pavarotti. I don't even know what the heck. Yeah. You know, my big takeaway, Jerusalem, is that human beings are frail. And when they get older, they get increasingly frail. And it is just a remarkable thing to watch, to see that, like, we're sitting here in 2026. Humanity has progressed in so many ways on so many dimensions. We have all this incredible technology. Two most powerful countries in the world are ruled by one guy who's 73 and one guy who's 80. And like, you know, they kind of don't walk great anymore and need naps. And just because there wasn't a lot of news actually to come out of the she-meeting itself,

it just made me more reflect on the sort of flaws that these human beings, like politicians as human beings are forced to. They get tired, they get hungry, they get cranky, they have fights with their spouses and children, and just like other shit gets in the way. of executing on the job. And, you know, we think a lot, or there's a lot of discussion about AI changing the labor force and how, you know, lawyers' jobs are going to change and teachers' jobs are going to change and doctors' jobs are going to change. The dimension, like the magnitude in which politicians' jobs have the potential to change in, you know, 1, 2, 5, 10 years because we are going to have intelligence on tap, I think was really put into light by the fact that we have these two incredibly frail and flawed individuals, which almost two billion people have entrusted to run them.

I mean, it's crazy. I mean, they are clearly very aware of this. I remember when she and Putin were overheard discussing the possibility of living to 150 years on that hot mic because of potential for longevity drugs to make it possible for them to live there. I mean, like, clearly it's very much on the brain for them. Even Trump domestically, of course, obsessed with his legacy increasingly, making funny comments about how he thinks he'll go to hell or whatever. So clearly, you know, end of life is on the brain. And, like, I mean, AI in China is clearly, like, the topic that everyone wants to talk about. And I agree that people often don't think about how much it's going to affect the jobs of politicians. Domestically, you can tell they're using it all the time. Like, I have run so many politicians, like, random tweets through Pangram. And the ones that, you know, they're tweeting themselves. Like, Ro Khanna clearly has used AI to write his, like, long-ass tweets about various economic policies. And, you know, you see this all the time in D.C. where you know that, like, high-level staffers are using it to understand and break down laws.

And, you know, a lot of people hear that and they think it's really bad. And there's, like, a question here about, like, whether or not that's bad. Because in a lot of these situations, like, it's a tool being used to hopefully speed up the process. But, like, in most jobs, you're like, okay, you got more productive. That's great. But, like, in politics, like, I didn't elect you because you were, like, the fastest processor on the planet. I wasn't, like, man, like, you were in Mensa and now you get to be, like, the senator from Rhode Island. Like, no, it was, like, you got elected because people, like, trusted you to make decisions on their behalf. But if you're, like, turning that decision-making over to Claude or Chachupiti or whatever, it's, like, raising those questions at, like, democratic legitimacy, which is different in China, right, where, like, it's not really, like, a democratically legitimate government. So when do you think, when Matt and Jerusalem, do you think we'll get the first candidate to run? I mean, maybe you have to be a new party or as an independent or something, basically saying, look, Claude or whatever model is smarter than me and smarter than my opponent.

I will let this model make all the decisions. And you can expect better government from me than you can from any human being who thinks that they have a chance at standing up against this model when it comes to organizing traffic or doing zoning rights or what have you. I mean, it's interesting because whenever I try to probe this, AI gives very sort of like technocratic answers to a lot of these kind of questions. I mean, there's lots of different policy issues out there, But there's a set of issues where pointy-headed elites on both sides tend to agree that, you know, a revenue-neutral congestion pricing program would be a good idea. And voters and politicians who are responsive to them in the real world infamously are resistant to those kinds of things. And people marvel in circles about Lee Kuan Yew's success in getting Singapore to adopt these very kind of expert driven policy solutions and debate the democratic status of that political system.

And I do wonder if more use of AI could push certain kinds of policies in a better direction, both by telling elected officials that they're correct, but also potentially if citizens use these tools to get advice, like it will just tell them, you know, correct, expert driven kinds of things, or if it'll go the other direction. And it'll end up being that because there's not a lot of concrete value in your AI telling you the truth about public policy, that it'll just get really good at, you know, reflecting back to people what they already think. Because that's been a lot of the social media impact on politics, right, is that, you know, formulating correct opinions about abstract public policy questions is important to society, but it's not that important to you as an individual.

Walking around with all kinds of misperceptions can be fine. And so we get a lot of expressive behavior, expressive content. And, you know, it seems like for now the people designing AI systems are trying to steer it in a certain kind of direction. But ultimately, if the basis of competition is, you know, do I enjoy chatting with this chatbot, it could start gassing me up with all kinds of bullshit. Well, I mean, it's hard because it's like right now, I think the – so I lean in the direction that AI will be more of a centralizing technology rather than a decentralizing one like social media has been. And so in the way that social media really fragmented our information ecosystem, I imagine AI will do the opposite in part because, like, I imagine the ones that are consumer-facing will be so capital-intensive and expensive to make that you're not going to have, like, thousands and thousands of them in the same way. And so you're going to have, like, just one or two, like, chat, Claude, like, whatever the next version of this deep seek, whatever these things are.

And you can tell, like, Elon really, really wants to make a right-wing Grok. Elon would love for Grok when you ask it, hey, are there too many Indians in America for it to go? Great question. There's, like, really good thoughts about whether or not there are enough, like, we've had too many Indian people in America, and if we should just have white South Africans being allowed into the country. And, like, Grok, I have not tried this specific question, but, like, in general, when you try to get it to answer very directly about, like, factual questions, about, like, the nature of social science research, the evidence that we have about different policy questions, it, like, answers very similar to, like, the way that these, like, center left or center right, like, econ-brained wonks would answer because it's trained on, like, a corpus of text that is very dominated by evidence that is amassed by people with small liberal tendencies. And, of course, personally, I believe, like, towards truth-seeking, right? And, you know, clearly there's, like, attempts to do, you know, reinforcement after the fact on more right-wing ideas.

And when Elon has tried to do that explicitly, it's, like, broken the model to become, like, less useful and become, like, a joke. And so part of me is just like, you know, I feel the threat of this sort of thing to be less, okay, do we get much more decentralization where you have these models telling people exactly what they want to hear when it comes to specific factual questions? and more if these are controlled, like these very centralized institutions are controlled by single companies, single countries, like whatever, there's like a huge incentive to control the what kinds of information are allowed to be understood or heard or whatever. Like I imagine if China's GBT was controlled by Chinese government, like you would not really get very many answers if you asked it about Tiananmen Square or whatever it is. So anyway, I think that's like the bigger concern for me. Well, there's some political stasis point, right, where, like, TikTok did get banned because, you know, because there was concern about the sort of political leanings of it.

And, yes, it's still around. And there was a whole, you know, long reason of that. But, like, I think if the – like, Grok is like a curiosity right now because it gets, you know, one thousandths of the usage. But if it ended up being a sort of like prominent shaper of the discourse in contemporary society and you had a Democratic president or a Democratic Congress, like like something would be done about that. So, you know, maybe it does come back to the map thing where we're like in order to sort of stay decently on the right side of the line, whether or not there's a Republican or Democrat in charge like these. And I think both parties will just be pushing them to like, hey, like you can't you can't go too crazy in one direction or another. I mean, this is what happened with, you know, the Google image where we had like the, you know, black female Gestapo. Right. Right. The diverse, the diverse Vikings. It ends up kind of bringing you to this sort of milquetoast, tectocratic thing.

And, you know, I want to come back to the thing Jerusalem, right, of the like, you know, they're using it for tweets and stuff. It's a little different than like, am I going to vote up or down on a thing? Or like, who am I going to write a bill with? But we'll get there, I think. Well, I mean, it's worth saying, right, at least in Congress, you know, the members, of course, rely fairly heavily on staffers and they rely on outside trusted sources to make these decisions. Right. It's not I don't want to say it's never happened, but it's rarely the case that when a new piece of legislation comes out, what happens is that a member of Congress goes into a room alone with the document. Well, no, but I mean, he also doesn't just study the text alone and then come up with an answer, right? For starters, I mean, you always get this kind of discourse around, oh, we haven't even had time to read the bill. But if you've ever looked at a piece of legislative text, it's incomprehensible unless you have specialist knowledge.

because it'll say things like, in U.S. Code 324 subsection F, delete, comma, and insert, and there'll be like a random clause. And, you know, if you know the corpus, you can understand what that means. So you're relying on staff to do it. And in the first instance, AI will, to some extent, enhance and to some extent, replace that staff work, right? Just summarize it. And also the groups that are helping them do it. And that's the thing. The staff themselves tend to be young and a little bit overworked, underpaid, a little underqualified. So I know when the revenue provisions of Inflation Reduction Act were being put together, Democrats were being helped fairly extensively by just former tax policy people who now work in the private sector for law firms. and like their job is to help rich people dodge taxes, but they in their hearts are Democrats or have personal relationships with Democratic members

because they used to be on staff. And so they would help them work these things through. So it's all very driven by personal trust. When that trust is well placed, it can function quite well, but it's pretty dubious. You know, it matters a lot. For example, there's this permitting reform. proposal that is now floating around in different forms in the Senate. And a big thing that people are really wondering about is what position will the League of Conservation Voters and the Natural Resources Defense Council take on the bill? Because in practice, how many members feel they can or can't support it is going to depend a lot on these interest groups. And to some extent, you know, if everyone is analyzing the legislative text with a common tool, that might create people who could still disagree about whether it's a good idea or not. But a big part of the argument that's

going to happen in D.C. over the next few weeks is just going to be literally like, what does this bill do? Yeah. Like, is it gutting the Clean Water Act or is it a minor modification? Are the changes to National Historical Preservation Act a huge deal or are they small? And particularly with bipartisan bills, it can get tricky because people negotiate on both sides. And then the Republican wants to come back and say to the other Republicans, this is amazing. It's going to unleash oil and gas. But then the Democratic negotiators want to say to their people, no, no, no, the fossil fuel changes are really small. And so to have a common set of analytic tools that are good, you know, could really destabilize some of how we count on the system working, because it requires people who see themselves as enemies to jointly get to yes on something, often by not agreeing on what it is they're doing. I'm like curious, like in China, are you seeing a lot of open evidence of the CCP or various,

you know, local factions using AI? So there's like performative. I have one question on that. I mean, can I do this? So it's interesting, right? Because like you need to, for the, for the bipartisan negotiations to work, you need to have like thought that you got the, that you like stole a march on the other side, right? The thing that they think is an eight out of 10 change is actually only a three out of 10 change. And the thing that they agreed to, which you said, oh, it's really not a big deal is actually a big deal. Right. And if like you actually get to, you know, it's interesting because on the one hand, like if you get to ground truth because you know chat gpt can actually like do a better job of forecasting and like make it very clear that like when you change this comma in this clause it actually means that there's a 40 chance of this happening and everyone yeah you know has an understanding that there a 40 of that happening i mean maybe it just like there a i see that like the creative ambiguity can make negotiations easier but it can also make them harder right If everyone just straightforwardly horse trading

because we all know how old our horses are and what the breeding history is, then, like, I don't know, maybe that's my option. Like, can I save Congress, Matt? This is what I'm hoping for. No, I mean, it's an interesting question, and, you know, there's an old kind of argument argument international relations theory where some people say it's easier to do deals with authoritarian regimes because they don't need to do politics internally right and can just commit to something that they think is good and don't have all these counterbalancing considerations but then there's an alternate view which is that because democratic regimes are more transparent and you need to come to your legislature and come to your people and explain to them like what's your understanding of this agreement, that it's more credible, right? That an authoritarian ruler could say anything in the room. He could sign any piece of paper and there's no telling, you know, what it is he means or what he's going to do. So, you know, it could bring people together to have more transparency, more common understanding.

But I just mean destabilized literally in like there's a ritual to how we do things now. And as the technology changes, you know, that's going to shift. But I think also, like, I mean, what's important, I guess, like, a lot of this conversation or this specific line presumes that a lot of the debate or the thing that's happening here is, like, at some level benefit-cost analysis, which I think is, I wish that's how people were making these decisions. And I agree in, like, some cases that's what's going on and it'll disabilize that part of it. But I think often it is, like, very, like, values-based. So, like, I mean, to take the permitting reform example, like, I think the evidence was, like, quite clear on, like, the Manchin-Barrasso permitting deal back under Biden, that that was going to be, like, net positive for climate emissions, even though it would make it easier to build some, you know, fossil fuel infrastructure, LNG stuff. and like the problem was that like there are environmental groups who have not a benefit cost analysis approach to permitting reform they have a all fossil fuel is bad keep it in the ground

approach regardless of the net impact each policy might have on reducing climate emissions and so in that case like i don't know what like chat gbt like what would have helped there like i mean i maybe would convince more people about like the upside of like allowing for permitting reform to build out our transmission infrastructure but i really don't think that was the problem or like the politics of it had a lot more to do with environmental justice groups that were opposed to allowing big greens to kind of come on board with this. I just, I hear what you guys are saying. And also I'm like, I'm not sure Congress has ascended to the level of counting up pros and cons and then deciding which one is bigger. So can I ask about, you know, a Chinese version of this, which is, you know, of course the political system is very different, but also the models are different. Like I, as a experiment, I fired up Deep Seek the other day and I asked it to tell me about the pros and cons of democracy and to give me an example of some democracies.

And so it, you know, it gave, I think, a pretty fair minded take on the pros and cons of democracy. And it said the United States, most European countries, it has a very euphemistic way of referring to Taiwan. And it said those were examples of democracies. And then it said that Cuba and Vietnam and Laos are one party communist dictatorships. And then it said that China has like a competitive system with eight parties and open consensus based decision making. And I was like, are you sure about that? And, you know, it was really sticking to a story about the idea that Cuba and China have very different political systems, things like that. And I assume Chinese decision makers are aware of how their political system works. But I don't know on what level of kind of self-deception or mandated deception like you're operating with as you get into a more closed system.

What are you talking about, Matt? There are eight parties in China. What are they? They're like social clubs. They're like really lame. I don't know. We got fact-checked on this from a China expert who texted me and was like, actually, China does have eight parties. And I was like, I think you're missing the point here a little bit. Okay, they've got eight parties. That's great. They have funny names. It's like they were kind of grandfathered in, actually, because they were like around. I think this is true to at least a few of them. They were, like, around in 1949, and some of them were, like, pro-CCP as opposed to pro-King Ishaq and KMT, and it was, like, a nice fig leaf, and it's kind of, like, an elite thing. It's, like, if you're, like, a really cool rich person, because you're still, if you join one of these parties, you're still, like, buying into the CCP system, but it's, like, you know, you can join, like, the consultative Congress. I don't know. We don't need to. Your audience does not need to learn about the seven other parties in China. Wait a second. And, okay, I just opened up DeepSeek, and I asked it, is China a democracy?

And it gives me this answer about how it's a socialist country and that, you know, it has a whole process, people's democracy, whatever. And it's a successful practice of socialism with Chinese characteristics. I said, okay, but it's not like open competition of parties. And it begins to answer and says, like, you're right. And then it immediately deletes all that and goes, sorry, that's beyond my current scope. Let's talk about something else. So that's the current – that's currently what's going on with DeepSeek. Yeah, I mean, look, I think there are two things going on with the Chinese models. First, they are trained on basically the same data that all the Western models are trained at. Because, like, you know, it's some, like, 94% of, like, even the Chinese models are trained on, like, English language corpus. Because there's just, like, more out there in English than there is in Chinese. Like, you didn't have a lot published for, you know, 49 to 68, 86. The Internet is newer. Whatever. And then you have this sort of final sheen of reinforcement learning,

particularly when you're asking, like... Sensitive questions. Like, very obviously sensitive questions, which, like, look, China's been doing AI safety much longer than the U.S. has. Like, they have, like, a model registry. you have to pass content tests in order for these models not to say things. So what's striking to me is actually almost the opposite. Like if you ask these models about other things, they do feel sort of like boring, normie, liberal, whatever. They don't feel entirely like different objects, only in the fact that actually sometimes they can write better because they care less about copyright, so they have more good books than them, I think. So I've been more like the best English language creative writing, I think you find out of Kimmy as opposed to Claude or ChatGVT or Grock. I was just curious here a little bit because I do think that there is a part of what both countries are grappling with are the economic impacts of AI

on their populations and on the world writ large. I think a lot of this has to do with, I mean, like a couple of months ago, I asked Matt, like, hey, what's going to happen with like the development ladder if AI becomes competent at doing like service sector jobs and then robotics takes over the normal entry level manufacturing jobs such that, you know, a no country that wants to develop with textiles and then move up and do call centers, all those sort of disappear. And, like, it's a very, like, dystopian kind of thing to really sit with for a bit because in that situation, not only would you have taken away, like, the only development ladder that we as, like, a human collective community have discovered, but you also have it happening across nations such that, like, the theoretical fix, which would be, like, a transfer system that you would transfer from, like, wealthy countries that have, like, these incredibly productive and wealth-generating AI companies, would, like, be giving money to poor places, and it's like, we're not going to do that.

Like, there's no way that, like, San Francisco, no matter how wealthy it gets, will ever say, like, yeah, you can tax us and send it to, like, you know, Somalia or whatever it is. And, like, I think this is something that people in, like, development econ are thinking about, like, pretty intensively. How is China thinking about this? Like, obviously, China is, like, way higher along the development ladder than, you know, sub-Saharan African countries. But it's still not at the level where it wants to be. So what's going on there? Sorry, that was a big question. Yeah. Yeah, can you solve all developmental issues? Are they worried about this? Like, what's going on in this? Well, I think the sort of the bet that the CCP in general and she in particular has made is that the only way out is through. So, you know, they have an aging society. We have youth unemployment at 20%. We have 40% of the country, which is still making less than $10,000 a year.

And then, like, you know, that is a few hundred million people, right? And we have, you know, a lot of folks who are still living in rural China, struggling on farms, like, you know, deworming, still a thing, like bad eyesight, whatnot. So the sort of idea of bringing new productive forces to bear to break the middle income trap of, okay, we've industrialized, but we're not Japan yet. Like, what else are we going to do is the only card that the CCP is really playing because, like, normal growth has basically ended for China. Like they were, you know, we were humming along in 2000s, eight, nine percent, 2010s, five, six percent. It's been like between half a percent and a percent and a half for the past six years. People were hoping it would come back in a stronger way after COVID and it just didn't. And there's lots of reasons for that.

There's like, you know, poorly allocated credit. There was this housing bust. And, you know, we're kind of in this like extended like zombie recession because there's just like not good. There aren't places where you can put return on investment. So what is the party kind of doubled out on this idea that these futuristic industries are going to of the future and selling the industries of the future to the world are going to be what takes China, you know, what's what what creates jobs and drives productivity for the party. So I think there's like the amount of sort of like AI is going to ruin us. You know, there's a bit of that. But you still have these like, like at least America, like we're at 3% unemployment. We're growing like gangbusters. And people are thrilled. I know. What is wrong with this? This is my main take, right? It's just like we got first world problems, man. Like we have a democracy.

We're going to have an election in a month. Like, we have low youth unemployment, growth, clean air. Like, people complain so much in America. They do. They're going to talk to you for that one, Jordan. Well, so from what I've seen of surveys, though, there's a big difference in mass opinion about AI, where, you know, people in the United States are fairly pessimistic about the impact of this technology on their lives, on society, even as people use it, right? It's not like Luddite nation and nobody is firing up Chagy Bt, but people have a dark view of how this is transforming things. And people in Asia, in China included, seem much more generally optimistic. And I've heard, I mean, I've heard a range of different takes on this, including the fact that the science fiction literary traditions are different in different countries. and that we just happen to be on a kind of robot apocalypse cultural trajectory that long predates modern LLMs to just the recent, you know, the past generation's worth of economic growth in China has been really good.

and so people are naturally just more optimistic about change that they have seen you know their parents and grandparents lives completely transformed so they want more transformation while americans are sullen and pessimistic um i have no idea you know i don't i don't know any foreigners i just uh i probably know a few but i just i just stay close to home dealing with my american pessimists you know i i saw that polling i haven't done their regression i haven't asked Excuse me. I haven't asked ChatGPT to do the regression for me, but I would not be surprised if the, like, explanatory variable there is just, like, level of development. Right. You know, if your life is pretty good and things are all right and you haven't seen a lot of change in your life for the better or seen a lot of change, period, then, like, yeah, something new and scary is going to seem new and scary. Oh, yeah. There are a lot of takes about, like, how, I mean, Matt and you referenced this, But, like, the law takes about how, like, in China, there's just less concern about the apocalyptic fallout of AI.

And there's, like, much more concern about it's, like, diffusion. Is that, like, your view as well? Like, AI safety is, like, the big obsession of the AI labs in the United States. And there's, like, a view that it's not actually going to cause that many problems from the Chinese perspective. Well, I think there is a view in America that it was going to not cause that many problems in the White House. So then all of a sudden Mythos came out and they freaked out. So I'm not sure how much explain it. Because like it is a weird like like just on face. It's a weird thing to believe that AI is going to kill us all. So you kind of need to be shown it in order to like freak out. And something like sort of the power of Mythos and the hugging face stuff was able to convince a White House, which is as temperamentally inclined as you could possibly have to sort of not buy into this narrative. So, look, Chinese AI, is it six months behind? Is it 12 months? Is it 18 months behind?

Their models don't have the capability to do stuff like the Hugging Face attack yet. Yeah, there's like an Ipsos polling on this, and you have to take a grain of salt because Ipsos polling of China, my understanding is that it's much more urbane and educated of a population than the general population. But, like, the gaps are, like, insane between China and the United States. So, like, the share agreeing that AI offers more benefits than drawbacks, 85% of Chinese respondents are optimistic, 38% of U.S. ones. This is from March of this year. And then, like, do AI products and services make you seem nervous? Only 40% of Chinese respondents said so. 64% of U.S. respondents said so. But at the same time, like in line with what Matt was saying, asked whether AI would replace their current job within five years. 52% of Chinese respondents said it was likely. Just 25% of Americans said it was likely.

So it's like, it's going to replace me, but like growth and change, it's going to be great. It's going to be positive. It's going to lead to all these benefits in the future. and I don't know we've done polling at the argument about AI and there's a you know there's a lot of people that think that AI could cause pretty catastrophic risks and yet that's not seeming their top concern which makes sense I mean people are just like generally do me about the technology here because they're generally do me regardless but can we talk about these cocky but also terrifying Americans Yeah. I mean... Well, but it makes sense, right? Like if you talk to anybody who works in a creative field in the United States the typical view is both that AI content is terrible right It slop et cetera et cetera et cetera But also that like its impact is really negative I only have the taste to know

it's a generalized negativity where you want to both say this could never be as good as me and my friends but also it's going to be so good that it destroys everything versus you know what I mean I feel this too I mean there was a several iterations of models ago, I was thinking, oh, this is a cool way to get research summaries done, but the writing is terrible, and this is not within a million light years of being able to replicate my work. And it's only by looking at the trend that I'm now like, all right, this is not going to last my gig here. It's just, you know, but that's, to me, a humbling kind of assertion. And of course, you know, we're not accustomed to seeing huge sectoral transformation unfold before our eyes in the United States. And I assume Chinese people

have just witnessed a lot more economic transformation in their personal life than I can imagine more. This is how we can make this sort of like, China's been growing really fast, like tangible, is, you know, a lot of jobs have like, Like the, like the, like the profession transform, like the, the, the chance that you have worked in different professions in China over the past 30 years, if you were working is extremely high. Like, you know, you started out in a steel factory or you were in the army or you worked in like a commissary or something. And then all of a sudden you worked in a bank because like the bank was, the bank sector was growing at 30% of the year and they were just want to take anyone. And then all of a sudden, you know, everyone's working in real estate and everyone's working in childhood education and then we shut down all the online tutoring so then everyone goes to work and anything else. I think there's just like, when you have an economy that's growing that fast, you just have much more labor force swings into different things. So the idea that 50% of the

people would think that the AI would take that job but they would figure something else out seems to be something higher than like, okay, you were in the US, that happens less because there's less kind of like dramatic you haven't seen this type of dramatic dynamism over the past 30 years like yeah maybe it from like 2005 to 2008 you had way more people selling houses and then they stopped selling houses but like it's it hasn't been quite as uh as uh as swingy for sure um when thinking about china over the past 20 years and uh joy politano who's a contributor to the argument he he did a great post for his newsletter um a pricatast recently about how over the last four years, America has lost more than 200,000 jobs in creative fields, publishing, TV, film, graphic design, et cetera. It's unclear, of course, like, you know, it's just we're just looking at trend lines here, like what the cause of this is. But one thing that was interesting he pulls out in his piece is when you look at music as a percentage of consumer spending, right?

The graph of this, I will try to describe it. I'll put it on the screen for people who are watching this or whatever. But like, it's basically now all tickets to live entertainment. And very little of the money comes from discs or even streaming. And still, like, the level is still pretty similar to its peak. So you still have, like, I'm just reading this graph, like, almost 0.4% of U.S. consumer spending is on music. But, like, the composition of what that is has changed a lot. So, like, I mean, I don't know. Like, I, again, like, Matt likes to make fun of all of the, like, econ-brained AI analysis that is just, like, look at all the technological innovation that's happened in the past, and you just get different composition of jobs, and you get growth, and, like, whatever. But, like, I just, I do think that it should be people's default expectation for what happens, and that there are, like, characteristics of this technology that might update you in different directions about what might be different about the economic reaction to this, to AI. But in general, like, we should expect that over the medium term, people are reallocated to jobs.

And also that, like, unemployment is largely a factor of monetary policy and not just, like, random technological innovation that causes unemployment and just that gets sustained. It's just like, yeah, like, if people get sorted out of their jobs, like, we should stimulate the economy and then we'll have different jobs. You know what I mean? I mean, it's interesting on the music case, right, that, you know, people, people, I mean, I try to, because I'm old, and I try to convey to people how, when I was in high school, it was like all my money went to buying CDs, because I didn't have very much money. But teenagers really care a lot about music. And so I was saving my pennies, and I was deciding which CDs to buy. And of course, that's completely evaporated now. Can we go down memory lane? What did you buy on release date, Matt? I don't think I got anything on release date. Because things would come out and I'd want to go to the store and I'd want to do sample listening and stuff like that.

But then I went to Russia one summer in 1998. And you could buy all this illegal bootleg music there. So I came back from Nizhinovgorod with a huge folder of ripoff CDs. But, you know, I barely ever went to shows because, you know, I didn't have any money to, like, go see a big band play was inconceivable. But now, you know, it's like the thing you could do to be a music fan, to invest some kind of money into it. And this speaks to, like, why the vibes are so bad for some people, too. Like, if you were a rich person who, like, used to go to concerts when, like, nobody got to go to concerts, you're, like, now that you go to concerts and there are all these fucking plebeians who can suddenly afford live music and entertainment are, like, ruining your experience. It's, like, I'll talk to, like, my friend's parents or whatever who were, like, here. They're, like, man, concerts used to be great. Like, you could just be, like, up front and it was all real heads and we spent, like, whatever.

And I was, like, yeah, like, you were rich. Like, now you're mad that all the middle class people could attend concerts with you. And, like, the riffraff is among you. So, hopefully, in a year, we'll be selling tickets to a live show. Exactly. To get back together with Jordan. And we'll be live writing. So, Matt always likes the idea of us, like, writing our takes live. So, everyone knows they're artisanal and organic and generated. I always wonder, like, I would love, yeah, going, like, the live stream of Matt writing a column, I think, would be fascinating. Like, I just, I want to know how it happens. Like, I'm not, I can't do it. I'm not as fast. I don't have to take generations. But it can't be a live stream because that could be easily faked. It's like, I'm going to get a typewriter, you know, go up on stage. I'm going to take questions from the audience and just, like, bang it out, analog. Oh, you do the Q&A newsletter live. Yeah, exactly. Okay. I think it could be amazing. You need, like, a second screen, though.

$35, I am there, Matt. You need, like, a second screen, though, like, for these ADHD kids. Because, like, while Matt's typing, you have to, like, watch, like, I don't know, like, some sort of racing game on the other screen to keep everyone engaged. So this is a pessimistic future. But, okay, I also just, I wonder here, though, I want to get back to, like, the development question a little bit here. Okay. Because the fallout regionally, I mean, like, the fallout regionally for both the U.S. and China, if the development ladder is, like, destroyed, is, like, genuinely very concerning for various reasons. Like, I care. So you're saying to explain. Yeah. The hypothesis here is right now, if a country is really, really poor, one way it can become less poor is by saying, hey, put your factories here. We've got some like really cheap labor. And if we get industrial robots that are cheaper than, you know, Bangladeshi peasants or whatever, then the countries that have advanced AI technology can say, well, we're going to do taxes, we're going to do redistribution, we're going to have, you know, fully automated luxury space communism.

but the countries that are poor right now are never going to be able to, you know, go through the process that China went through over the past 30, 40 years. Yeah. I mean, I could just like, let me tell you about something. I know something about that's related as opposed to this. Look, there is this like China right now is getting older. They're having like China is like your quintessential example of this model, right? It's like they were the world's factory and then they got way richer by being able to do that. They're trying to play this game again by exporting increasingly high-value goods and going from an OEM to having their own brands. I think there was some level of disappointment that the Chinese firms didn't, that Southeast Asia or Africa didn't follow in the footsteps of China. India still hasn't been able to figure out how to do manufacturing at a global scale. And I think the story

is that the Chinese firms got more efficient and insofar as they needed cheaper capital, you had the Chinese firms ending up sort of expanding into Vietnam, Malaysia, Bangladesh. But anything that's like kind of, anything that's like slightly nicer or required sort of more skilled labor, they still end up doing in China just because you had enough productivity growth and automation that it wasn't really worth it. So I feel like this is like, this feels almost to me to be like AI, independent of AI. At some level, we're still human beings who live in the world and need to buy stuff to make our lives comfortable. Will crazy industrial automation bring, you know, take how manufacturing only happened in the US and China going forward, I don't know. I think there's just like a lot of weird stuff

and a lot of edge cases. I got a few things to worry about. I mean, I don't know. I'm very worried about this. I think that like one thing that will obviously be a comparative advantage for developing nations is like mining of raw minerals, but that is like the reason why the textiles were such a good entry level to the development ladder is that it a like introduces mechanization it inculcates like i mean this happened in britain this happens you know in china etc but like it inculcates like an idea of like you show up at this job and you train employees like understand how like working in an industrialized economy works like it's very very different than agricultural work where like you're really not like you're like working all the time and also like not working lots of the time it's like a very different form of work you're like not on the clock like whatever it is but What's your experience as a farm worker? Where's your argument farm? Can we do our live show at that? A lot of some chickens running around stage while Matt types out his takes.

My grandfather, I'm sure, is rolling around in his grave somewhere. But, you know, whatever. Farming is very, very different than mechanized labor, And that is an important part of teaching and, you know, a labor market or a labor force how to prepare for advanced industrialized and advanced service sector economy work. And like that training, that like shift of how the labor force actually is, you know, is acculturated to a modern economy is like really critical for development. development and if it's just that you instead have these minds that are often probably easily controlled by like one centralized government or one individual corporation like there's no reason why that would result in the rest of the country also developing the skills necessary to have like a modern developed economy so i'm very worried about this because i think obviously like it's really bad that people are poor but also like i think that like i think china and the u.s

are going to have to care for their own selfish reasons because it will obviously spur a ton more migration. It will obviously spur a ton of unrest if you have, like, no path to prosperity anywhere outside places where there are AI labs. Like, I don't know. Like, I just think people are not worried about this enough. Well, wait, I want to do an agricultural labor take because I don't know anything about... You've worked on a farm too, so you're... No, I don't know anything about China, But I did take a, like, government-funded propaganda tour of China one time. Oh, when? Back in 2009. Okay. So, you know, it was like a million years ago in China years. I was always talking about these tours they used to take journalists on, and nobody does this anymore, and I find it, like, absolutely interesting. Well, it's the ethical standards. You haven't invited me to Abu Dhabi, Jerusalem? Abu Dhabi's got to get on their ship. People have become more ethical. I would like to go to Abu Dhabi. So I'm going to put that out there. So I was in Shanghai and, you know, at some dinner and there's like a tedious presentation.

And then, you know, we're chit-chatting and I'm sitting next to some party official, some guy from the North America desk at the foreign ministry. And I'm complaining about how the young people these days are too soft and weak and stuff. And he says, yeah, you know, during the Cultural Revolution, I was made to go live on a farm and it was very unpleasant, but it forged my character to the man I am today. So he had a very different take on agricultural labor and a real appreciation of it. And I thought it kind of an optimistic read on being a victim of the Cultural Revolution. This is She's Lived's story, right? His parents were purged. He was sent down from Beijing to the countryside. I mean, the thing you need, Matt, is like you want to, once you're in study abroad in Inner Mongolia in 1973 ends,

you want to be able to go back to a country that actually has like education and infrastructure and like a future, right? So it's a little different than being in Bangladesh in 1978. There's not like, okay, you turn the key and then all of a sudden it's back. So you want to apply all of the pain and suffering to an industrial economy, not just the farm next door. But there is more habituation to disruptive change and bad things happening. I mean, it's interesting. I'm sorry, Matt. Is this a pro-take, a pro-culture revolution take? A pro-culture revolution. No, I just – I thought it was interesting because this is obviously one of history's big bad events that occurred. The same regime in some sense is still in charge, but it's very much not people who are saying – Well, this is – I mean, this is good.

You're on the fucking propaganda tour, Matt. So, like, of course, this is what you're going to hear. But also, I mean, I think there is a, you know. I learned some very interesting things about the Dalai Lama on the propaganda short. You don't know the truth about that guy. Look, there are, for every person for whom the Cultural Revolution was a cute little building, you know, a chapter in their buildings, Roman, there were like, you know, 50 million people who died. and then another you know 50 100 200 million people who lie whose lives were like irreparably shattered so i mean you know you take like yes like if you if you send everyone to some horrible experience like some of whom will be able to like um what is it uh uh uh post-traumatic growth or something that's like that's like our hot phrase i mean millions may die but it's an opportunity for personal growth and enrichment for those who survive I mean yeah I not here to play the Hassan Piker role But it Where is the argument on national like mandatory national service

Like, there's a halfway point here, right? Anti. Anti, okay. I mean, we don't have company lines here. You know, that's the whole point about the argument. You know, people are going to have different takes. But, you know, I'm not very sure. Is that in Chris Murphy's book, though? Yeah, it is. He wants national service, right? He does. I mean, Eritrea has mandatory national service. it's not really great. So we should put that out there. Yeah, it's where I'm from. Yeah, but I mean, like, Israel, South Korea, Singapore, it works out for them. Yeah, I prefer America to all of those countries, so. Yeah, but look how ununified and we don't have pain through progress. Yeah, look, look, maybe if we had national service and people had to eat shit for a year, we wouldn't be so pessimistic about AI. You just go back to your normal life and be like, oh, okay, cool. Like, I'm not, like, making forest, you know, paths in like a tier four national park like i i can enjoy my my my first world life and maybe be grateful for it right we need if we force kids to do backbreaking physical labor they'll come back they'll appreciate automation and progress and they won't mind feeding their bosses lizard

crickets and they'll be fine with that um yeah it's totally devolved holiday in cambodia style Okay, we've got to pull out of the tailspin. I have a corruption take for you guys. Okay, shoot. Sorry, where were you going to go, Matt? I was going to return to serious bilateral diplomacy. Let's go with that. Let's do corruption now. That sounds good. The sort of growth in corruption in the U.S., overt corruption in the U.S. system is something I think needs no background. But the striking thing to me is like corruption, deep, you know, central feature in the Chinese political system since at least reform and opening. And, you know, they have sort of like, I don't want to say made peace to it with it. Like she does all these anti-corruption campaigns all the time. They're not happy with it. But but there is a social equilibrium that I think China has reached where the U.S. is not at. We're like in China. You need to hide it. um and we've gotten to such a place with this administration um and you know thanks to

scotus rulings and and whatnot that like you can just be up front um about uh how corrupt you're being and like this whole idea of like having the sons and the in-laws be the vehicles is something i mean it's worldwide particularly um uh also a pattern that happens in china but like like just you could not like the system had enough antibodies and i think the society had enough antibodies that even if you wanted to pull some shit like this you try to like keep it on the low not be like tweeting about how um you know we're so you know i'm so excited to be in partnership with like mubidala or whatever um and it's just it's just really striking to me that like yes Chinese corruption runs so much deeper and is a much more integral part of the society than Trump making $500 million here, $500 million there. But the party is extremely sensitive

to overt public corruption things. And that is a main focus. And they do crackdowns all the time to show that they're taking it seriously and they cash people with, you know, hundreds of millions of dollars in their, not tens of millions of dollars in cash and their houses and whatnot. But the U.S. system, it's like you can run the country and, like, be totally fine with half the country screaming at you, like, saying you're super corrupt. I find that very curious. Well, and, you know, one thing that's interesting is that, you know, people, corruption is bad and it seems like you should crack down on corruption. But something that, you know, I mean, I think we saw in Chinese history is that saying you're going to do a crackdown on corruption can very easily just become a pretext for, you know, for purges and authoritarianism. And that especially if a lot of corruption is legitimately happening, anti-corruption drives can be very pretextual.

A lot of Vladimir Putin's consolidation of power in Russia, for example, took the form of cracking down on corruption. But you crack down on the people for anti-corruption because anti-corruption is a popular thing. Right. Yes, yes, yes. But I mean, exactly. It's a good cause. But it raises the question of how do you actually pull out of the spiral? Right. Because it's easy to say, OK, well, we need some new political authorities in. They need to clean house. They need to oppose accountability for what's happened here before. But it's like the line between, okay, the new regime is going to persecute its political enemies and the new regime is going to clean house is very, can be very hard to draw. I mean, again, you know, on my propaganda tour already, people were talking about how there was going to be a corruption crackdown. And even within the context of a propaganda tour, there was a lot of like, like, this might just mean, you know, people jostling for power and denouncing their enemies because the system is very corrupt.

So everybody's guilty. So, you know, who do you end up going after? And, you know, I think we saw this, like Trump has tried to do a lot of what I would characterize as politically motivated prosecutions of his opponents. But he says is, you know, anti-corruption stuff against Hunter Biden and whatever else. And Hunter Biden does, in fact, seem to be a little bit corrupt. It's just that the moves against him don't seem to be on the level. Right. Because. Because you can't both be outraged by the way Hunter traded on his family connections and also look at the way Eric and Don Jr. are conducting themselves. Like, that doesn't make any sense. But how do you get, you know, a leveling up of conduct rather than, you know, tit for tat? I don't know. There's, like, something interesting that ties back to AI a bit here, where, like, obviously a lot of corruption that happens in all countries basically

has to do with, like, procurement processes. and they're trying to use AI as an anti-corruption tool now to, like, identify bid rigging and flag bias scores by procurement evaluators. And this goes back to, like, kind of, like, at a more, like, technical level, like, AI use in government, that corruption and, like, you know, unfair decisions being made that don't, you know, comport with, like, the reasons being given by government for why things are being done. It's, like, really easy to, like, root out with AI. There's a system that can see everything and process things very, very quickly and know, hey, this office is giving contracts consistently to just people from these places. And I can see that those are places where the employees of that office are from or have political ties to. Like, it's very easy to see that versus, like, it's a lot of backbreaking human, expensive human labor lawyers who have to do that kind of work now and requires them deciding, like, what's worth my time and energy to even focus on, etc.

But also, like, I mean, I don't know. I think that it's true that, like, you know, the brazenness of U.S. political corruption is, like, off the charts right now. I mean, Trump, like, publicly offered dinner access to the top 220 holders of his meme coin in 2025. Like, that's, like, insanely explicit. I think people are not – I mean, their charges of corruption fly all the time that people, I think, are not getting, like, how different it's gotten now. But also, like, the law has changed in ways that are, like, insane. So, Treon White, who is a very, like, local DC thing, Jordan, so forgive me, but, like, Treon White was a council member, and he was charged with, like, taking a $35,000 bribe. This guy's, like, on camera. Like, everyone knows he took the money, but there was, like, a hung jury and a mistrial because he didn't do anything after taking the money. They couldn't prove that he, like, actually did, like, whatever the – but, like, he's being asked repeatedly, like, so what was the $35,000 for? And he's just like, let's stay on topic.

So, like, I don't know, like, Snyder versus the United States, like, that's a while ago. But it basically becomes very, very difficult to actually hold anyone accountable for corruption. So I don't know what's going on in China, but in the U.S., it seems hard. Yeah, so how does it come back? I mean, I think you have, like, what is the Republican candidate going to say about any of this in 2029? and like is this is this like it would be interesting if it's like is there any world where you have some prime like say say trump is polling at like 15 right and then uh you know you have rubio and jd are trying to like dance around this like there's got to be a lane for someone to be like drain the swamp right like like these two guys on the stage are like deeply corrupt that's what's happening right now i mean i mean is that is that tucker is this like one of Tucker's things. I don't know how potent it is. I mean, it just seemed to me logically like the lane is for somebody

to complain that the Trump Vance administration isn't right-wing enough. That they haven't cut spending, they haven't addressed the Holocaust of the unborn, and that they have been expending their political capital on this corruption. Right? To say, you know, instead of whatever dinners we should be... I mean, I don't know if anyone will actually try that, but that was sort of the Ted Cruz 2016 pitch. It's like, we need a hard-edged ideologue who will be remorseless and unfunny and just locked in on doing right-wing policy. That's what I would try if I was beating JD. I mean, I don't think I could sell it. Not, I'm an AI and so I'll be better than you? Well, maybe. We've got to wait until 2032 for that. The models won't be. Well, I mean, maybe AI could help us develop this pitch, right? I mean, this is the... So anyway, I wanted to circle all the way back to where we started with the Trump summit. Did anything happen at this summit?

Like, I feel like American politics is so polarized that it's hard to learn new facts about things you're not familiar with. And so it's like, of course, the media was like, oh, it was pointless. And Republicans are like, Donald Trump is a genius. But like, why did they hold this summit? What did they do? They held a summit because Donald Trump likes to throw a party and he likes to play hotel owner. Look, I think, I mean, listen to China Talk for More if you want to learn actual things about U.S. Translations. That was a great ad. You do know how to do ads. You know, I had an emergency pod scheduled and canceled it because I was like, we don't have anything to talk about. Okay, so it wasn't even breaking news to pretend you needed an emergency pod. No. We look the the the the interest. So where are we in U.S.-China relations in 2026 today? The sort of big development over the past two years has really been Trump's like effort to escalate on trade, not on technology.

So he does this whole big, you know, tariff war. And China says, we can actually play the same game. And we're not Canada. So like what we'll do, we'll actually hurt. And like we're not scared. So you have these rare earth controls. And then all of a sudden the Trump administration is like, wait, maybe we don't actually care all this much. All of the like actual China hawks are kind of fired. Trump like doesn't like confronting dictators. And this was a great off ramp for them to be like, okay, any escalatory stuff. we're just going to put on hold. So you've, in this sort of, in this like pause of any sort of more coercive, you know, constraining activity on China's trade or, you know, technological development, you've had, you know, many policies freeze. The Trump administration is trying to spend all this money through the Defense Department and the DFC to, like, build up America's rare earth capacity so that China can't do this again.

But in the meantime, Trump likes to throw a party. And he thinks it's cool to go to Beijing. And he thinks it's cool to have, you know, to pick up Xi at the airport and show him apparently, like, red coats dancing. Okay, but, like, the weirdest thing that came out of this, which I still, I do need something to explain to me what's going on here. The White House said that, like, both China and the U.S. had agreed to call AI superintelligence. And then China came out with a statement that was like, we respect, you know, like the woke terminology the Trump administration wants to use with regards to super intelligence. But like, then they continue to call it AI themselves. Why is this even a thing? Does anyone know what's going on with the super intelligence thing? Like why he wants to rebrand it? Yeah, why are we doing that? Well, I mean, look, AI is polling at 14%. It needs a rebrand. Yeah, that part made perfect sense to me. Sorry, we think that, like, calling it super intelligence is going to somehow change the game here? No, but that's the politicians. He had the touch, right?

He was, you know, little Marco, right? Like, this guy was the genius at it. And human frailty, right? We're in our 80s now, and, you know, he doesn't have a fastball. We need AI to take over. Making up funny nicknames for people seems like saying that, A, Trump was really good at, and B, that AI is really lagging on. Oh, I take it back. It makes great titles for me. Did you see my comic today, Jerusalem? No, what? Okay, this is the best thing I've done all year. The state dinner was so boring that I had to reimagine it. The concept I came up with was instead of booking the military band, Trump books Kanye. and then Kanye starts off saying, you know, we are happy to announce our second soybean trade framework and he goes, fuck that. Xi Jinping doesn't care about Muslims.

Peng Li Yuan, who is China's first lady, was actually like an incredibly accomplished singer. Like Mariah Carey, Dolly Parton levels of fame, chachy bp tells me um marries him when he's like just like a son of a famous politician but was like you know a super low level presidential person she rushes to the stage gets into a rap battle with kanye um kanye proceeds to overhear um uh trump and she talking about replacing their organs to live to 200 and she and and kanye's like oh wait that sounds really cool and then he closes by releasing an album called Jesus. So I asked, I give this to ChatGPT. I'm like, give me some titles, and it gives me my beautiful, dark, twisted state dinner. The three weeks later in Kashgar panel is amazing. I really enjoyed that. Well, this, I think, has been our most freewheeling conversation yet. Jordan, I want to congratulate you.

You're welcome. You said it was like boring us. I'm like, Jerusalem. I'm sorry. I was like, can we talk about AI and economic development? I do want to say for the record, we wanted to have a serious conversation about the implications for liberalism of China's success at economic development without liberalizing domestically. Join the arguments Patreon for that conversation. Yeah, so if you enjoyed this, you'll definitely enjoy China Talk. Please go and subscribe. It's a great sub-stack. It's a great podcast. Very informative and funny. But Jordan, thanks for joining us. Oh, this is such a treat. Thanks for having me. Thank you.