Is AI Really Taking Our Jobs? — With Noah Smith
56:43 Share

Star economics writer Noah Smith joins us to talk about AI productivity, immigration, why our politics are so silly, the state of the US-China relationship, and much more. That's coming up right after this.

Welcome to Big Technology Podcast, a show for cool-headed and nuanced conversation of the tech world and beyond. Today, we're joined by one of my favorite writers, Noah Smith, aka Noahpinion, is here with us to talk about whether AI is actually leading to productivity, but also a lot of bigger and broader questions about the state of the U.S. and the global economy, including the U.S.-China relationship, immigration, and tariffs. Great to see you, Noah. Thanks so much for joining the show. Thank you.

Great to be here, man. Thanks for having me on. It's great to have you. I definitely want to talk tariffs with you. Um, we're going to do that in a little bit, but for our audience, really the bread and butter here is artificial intelligence. And a big question is, you know, amid all the hype, all the bluster is AI having any impact on productivity. And for someone, you know, in your position, you've examined this, you're an economics writer. Um, I thought we would just start there because we'll, we'll, uh,

cover the stuff it's definitely be of interest to everybody and then we will go broader as we go right so in terms of impact on the productivity statistics it's hard to tell because ai as we know it has only been around for a couple years and um gen ai llms yeah so like other ai you know machine learning just like computer vision or stuff like that um

old ML, old AI kind of stuff that probably had some effect, but probably, you know, there's a lot of other stuff going on. We've had good productivity growth since the pandemic, like not spectacular, but solid. And you can always think about whether that's in the face of headwinds from anti-development restrictions or curtailment of trade, you know, like not just tariffs, but like just general deglobalization, China buying less of our stuff, things like that. And so we've

There's a lot of other stuff going on. And so we've had decent productivity growth and it's all been in services. None of it's in manufacturing actually.

Manufacturing productivity has slowed down. Oh, yeah, go ahead. Manufacturing productivity? Service productivity is the kind of thing we'd expect LLMs to add to, but the service productivity boom has been around a lot longer than LLMs, so it's hard to tell. And we haven't really seen mass business adoption. Obviously, ChatGPT is a popular website that has many millions of subscribers.

But so, you know, in that sense, it's like, but it's hard to tell whether that increases productivity at all. It's hard to tell whether that translates in terms of like companies using AI to do stuff. It's too early to tell with that. Like we haven't seen that happen yet.

We just had a McKinsey report come across the big technology discord that I found interesting, where the consulting firm finds that nearly eight in 10 companies have deployed generative AI in some form, but roughly the same percentage report no material impact on earnings. They call it the generative AI paradox. I mean, it's not that companies aren't trying. It's just they aren't seeing success. What do you think about that?

So I think that here's a story from the, this is one of the economists' favorite stories about technology. And it's the story of how electricity was incorporated into factories.

So originally you had steam powered factories, which would essentially had a large boiler. The boiler would turn a belt or a crankshaft or something like that. And then everything would turn at the same rate. And then people would have to like slap stuff together at the same rate. And you were limited by your rate limiting step, your slowest step, right?

And so that happened. And then people, that's how factories worked. And if you see old like Mickey Mouse cartoons, he's like on this conveyor belt that's like everything like stamping something at a uniform interval. That's because that's steam powered. You couldn't slow down. And so when they, when electricity got invented, factory owners were like, okay, how are we going to do this? Let's rip out our steam boiler and put in an electric dynamo. You know, it's electricity instead of, you know, you just run a electrician

cord and power your crankshaft or your belt that way. It turned out that electricity released much less power than steam did. And so it was just a lot weaker. So you had to pay a lot more for the same power. It's like trying to heat your home with like electro-resistive heating versus gas. And so people were, so they were like, this is useless. This electricity stuff sucks, man. So they ripped it out, put it in a steam boiler again and went back to the old way. And this happened, that lasted for about 20 years until somebody

Brilliant entrepreneur said, wait a second, we could just do all of manufacturing just differently because electricity, unlike steam, can easily be routed in parallel instead of in series. So you took wires and you routed the electric power to different workstations on a flat factory. And then once you did that, you could carry pieces from one side to the other. And what this allowed you to do is have each step work at its own rate instead of all having to work at the same rate. And this allowed you to apply a lot more resources toward the rate limiting steps. And this

This plus, you know, some other advantages you got from electricity that required you to reorganize the way your factory was shaped ended up increasing productivity by about a factor of five. So you got a 5x increase in productivity, this massive, massive, huge increase in productivity from factory electrification. But it took decades for people to figure out we have to change not just what technology we're inputting, but the whole organization of our production process we have to reorganize.

And with AI, you're seeing the initial impulse toward AI as, well, I'm just going to rip out a human employee and put in an AI chatbot instead of the human employee. We've seen that in a couple situations that actually work. So like call centers...

In some call centers, depending on how much help you need, call centers have been able to just replace humans with AI. We've seen a few cases where this happens. Everybody mentally thinks this is how AI's got to work. It's a human replacer. The chatbot sounds like a human, so you have to use it like a human. Well, no.

In the end, you know, and of course, my friend Dario Amadei of Anthropic believes that AI will get so good that eventually you will be able to just use the dumbass use case of just replacing every human with AI. But I say that we're not there yet. We keep finding reasons why AI is not quite like a human. And so replacing humans with AI doesn't quite work. However...

machine tool isn't quite like a human or electricity isn't quite like steam either. And so eventually we may like people will come along and say, you know what? I could just organize my company entirely different to take more advantage of AI. And I don't just have to replace my humans with AI. I just have the humans do different things and the AI do different things. And I reorganize the production process in ways that

I know a Smith sitting here in 2025, I'm unable to imagine. And so then we will see amazing productivity growth at that point. But it would, you know, with electricity, it took 20 years for people to sort of think of it.

so then what do you make of the fact that people are saying well because of the introduction of llms into companies entry-level jobs are harder to find these days i mean this is a narrative that's taking off we've discussed it on the show that it seems like it's harder to get an entry-level job today than it ever has been

maybe outside of recessions, like in a good economy, it seems like it's difficult. So do you ascribe any of that to the AI boom or is that just something else that's happening and we're sort of conflating the correlation and causation? I think we don't know. So the answer is that we don't know what the hell's going on. You know, I've seen...

Data on like entry-level jobs, you know, there has been a slowdown entry-level hiring, but there's a number of reasons this could be. So one reason is the macro economy. One reason is, um, two, there was over hiring of entry-level jobs during the like 2021 during the bubble. And now people like glutted on entry-level jobs and are now cutting back for a while. Um,

There's, you know, there's other stuff, too. There's the glut of people who just all stampeded into the same few majors and career paths. And so, yeah, so there's all these things going on. And it's not it's very much unclear as to what's going on. So there was a good Economist article by someone who really doesn't believe this thesis.

You know, claiming to debunk it all. And I don't think they completely debunked it, but they raised lots of other possible explanations that were pretty sufficient, I think, to explain it. We don't know which is which. It could still be AI. This could still be what's happening. But I think that we don't have enough evidence to conclude one way or another what's going on there.

Okay. And you mentioned that you're friends with Dario. His prediction, I think, is that we could see 50% of entry-level white-collar jobs eliminated in the coming years. But it doesn't seem like you agree with that. So where do you think the employment levels will go as this technology gets better? And how do you think it will just change our economy and our society?

Right. Before I answer that, do you want me to quote you a couple of the statistics from this Economist article that I just looked up? Definitely. Okay. So it said that young graduates' relative unemployment started to rise in 2009, long before generative AI came along. So this trend long, long, long predates AI.

It's been going on the relative unemployment of of, you know, college graduates relative to other people has been rising for quite a while now. You know, that's that's 15 years. And it said and their actual unemployment rate at around 6 percent remains low. Six percent.

That's not super low, but it's not high. It's not catastrophic here. It says, returning to a measure reintroduced into 2023, we examine American data on employment by occupation, singling out workers that are believed to be vulnerable to AI. These are white collar employees, including people in back office support, financial operations and sales. There's a pattern here. We find no evidence of an AI hit.

those occupations as a percentage of total employment are still rising steadily and have been rising on basically the same trend since about 2000. So that trend does not appear to be disrupted. We're still hiring a lot more back office support, financial operations, sales. All those people are still getting hired more and more. Those positions are still increasing. And those are the positions that people ahead of time thought would be very vulnerable to AI. And, you know, wage growth is strong, blah, blah, blah. And then, yeah,

An official measure suggests that less than 10% of American companies employ AI to produce goods and services. So AI is probably not big enough to have like a major impact yet there. But so I think that the evidence probably leans against a big impact from AI like driving these trends, but it could be in there. You know, this doesn't rule it out. Like I said, there's always a lot of stuff going on at any given time.

Right. And it's hard to tease it all out. But there's no there's no like white collar job apocalypse from AI that we can see yet. I mean, there won't be one. So let's talk about the future. So Dario thinks Dario is an engineer and a researcher. I mean, he's more research than engineer, obviously. But then but he I've known him since since, you know, well before the anthropic. So he is.

Researchers target human capabilities for simulation with AI. They say, what can a human do? Let's look at the specific thing humans can do and engineer a system that does that. This is not just unique to AI. A lot of technology does that. So if you're looking at a machine tool, you think, okay, humans fasten a thing like this. Look at how my hands work. Maybe I could make a little gear, robot, assembly, whatever that fastens things on.

So you're trying to, you know what humans do, and then you try to make a machine that does a similar thing, accomplishes that purpose, right? So you're trying, in some sense, a lot of technology is intentionally trying to replace human labor. And we've been doing this since, you know, the dawn of technology. Saying a human carries water. What if I made a pipe to carry water? A human turns this wheel. What if I made a boiler to turn this wheel? You know, and so you intentionally...

engineer and research based on observing human capabilities to know what's possible. Man looks at the birds and wants to fly. This is old like 50s shit. Oh, yes. That's how we do it. And then, so when companies like McKinsey come along and they say, which jobs are vulnerable to AI? How many jobs are vulnerable to AI? What they do is they sort of

Chop jobs up into tasks by asking people, what tasks do you do? And then people give them a list of tasks you do. And then you go to engineers and say, can you engineer a system to do this? And if the engineer says, yes, yes, I can. Then you say it's vulnerable to being replaced. This is

complete who we and they all come up with really huge percentage like 40% 60% of your jobs are going to be replaced by AI so that you know every there's a million groups doing this and they'll come up with different things and it's who we but for a number of reasons number one the people that you ask which tasks do you do are not giving you a list of the complete tasks there's a lot more stuff you do than the stuff they say they're

The second thing is that the jobs are deemed to be vulnerable to AI when half the tasks or something, some percentage of the tasks are, you know, like AI doable according to what these engineers say. Right. And so,

If a machine, instead of taking half of our jobs, takes half of each person's job, that just makes us more productive, right? It frees us up to do more things. Number three, the engineers don't necessarily know what they'll be able to automate. They say they can, but they can't always do it.

And so, for example, with AI stuff, it's all fun and games until catastrophic hallucinations cause you to lose a big client at your AI automated sales shit, right? So things you say you can automate, often you can't really automate. And then that's three big reasons why these studies are bunk. And then one more big reason, the most important reason, is that these studies say nothing about additional tasks you could add. So suppose that my job is...

consists of answering emails and then like, you know, sort of like planning strategies.

And now an AI comes along that can answer all my emails. It's just taken up half, that took up half my day. Now I can spend the other half of my day. You know, now I can spend the other half of my day doing other things. So when a very stupid naive way to think about it is, oh, well, your wage will be cut in half and you'll be hired for half as many hours and you'll only do the one part, the strategizing part of your job because the email part is gone. Goodbye. Wage cut in half, hours cut in half. That's bullshit. That's not how things work. I will have, I will find something else to do.

Maybe I'll do even more business strategy than I did before or maybe I'll find yet more tasks that I didn't know that I could do because it freed me up. And if you look at the history of technology, if you look at people who are freed from carrying water, plowing the field, blah, blah, blah, blah, blah, all this stuff, they found other, guess what? They found other things to do. And factory workers, as we did more and more machine tools, we found more things for factory workers to do. You know, so...

We also find more things for people to do at not even at the job level, but at the industry level. So like your company could hire you for a different kind of role. So we saw this thing where IBM fired all these people because it thought they could replace their jobs with AI and then had to hire them all back, had to hire like a ton of people back, maybe not the exact same people, not doing the exact same things, but for, you know, other similar related tasks, right?

that AI couldn't do yet. They had to hire all these humans back who were doing slightly different things than they did before. And I think this is

None of these like predictions, these dire predictions of what a of jobs as destroy take this into account. And I think when Dario looks at the stuff, he's looking at it from the typical perspective of this of the engineers that the McKinsey guys ask, can you automate this away? He's just a bigger version of that. He's thinking, yes, my company can automate this away. But there's all these other questions, all these other things about like human capabilities that I think that.

Dario is an expert on AI, but he's not necessarily an expert on what humans do in every job in America, right? No one is. - Definitely. I mean, I think if you're a salesperson, just to take one example, this really hits home. I've been in sales before. I can tell you no salesperson is on top of their pipeline.

they might be concentrating on like the two or three clients that are most likely to close and have a hundred accounts that they have to reach out to or nurture or nudge or, you know, meet with, and they just don't have the time. So if you make a salesperson, let's say a hundred percent more efficient now, instead of concentrating on those two, three, four, five accounts, maybe they can spend time with 10 and then all of a sudden they might be doubling their productivity if they're motivated. Yeah. Yes. And so,

In the past, we've always found stuff for people to do. And you can say, well, this time is different. AI is a fundamentally different technology and there will be nothing for us to do. First of all, that's actually wrong. There's a thing called comparative advantage, which I'll talk about in a bit. But also, setting that aside for now, we don't know what human capabilities are. We don't know.

We don't know what AI capabilities are, but we don't know what human capabilities are. That sounds like a bizarre statement because we've had lots of humans working in our economy, doing lots of different things for a very long time. However, at any point in time, I can point you to massive numbers of occupations and tasks that did not exist 10 years ago, 20 years ago, 30 years ago. OK, we didn't know humans had this capability to do these kind of things, but they did. And we just thought, oh, you know, it was a matter of course for us.

If you look at the history of technology, they've been saying that AI is going to cause mass, not AI, I'm sorry. They've been saying technology is going to cause mass unemployment since the birth of the Industrial Revolution. People said that industrial technology will mean laborers will have mass unemployment. They won't have anything to do and they'll revolt and they'll riot and society will fall, you know, because you have all these unemployed guys. You know, the threshing machine is going to replace all the farmers.

And the, you know, so self-driving or drivable tractor, like, you know, can replace all the farmers. The prophecy was that was it. What are we going to do? And so then we found something for people to do. And now people say, well, AI is different. AI is different. That's going to replace all the people. And maybe they're right. I can't say they're wrong. Right? I can say they've been wrong every last, every other time they've predicted this, but then sometimes something that never happened before happens. And so I can't tell them they're wrong. And yet, conceptually, the...

idea of like well we made this thing to do all these tasks that humans do is not a is not by itself a reason to think human labor is going to be mass replaced okay so now let me ask you this question in terms of economics that i think would be really interesting to get into since you've brought up the fact that we've had technology and then it has automated things but yet people have found new things to do and you've also talked a little bit and written about the fact that

that our hollowed out middle class is not as hollowed out as people think it is. I think there's a narrative that technology may have found other people, may have found people other things to do. But if you look at the economy more broadly today, it's much more unequal than it has been in the past. And technology is a big reason why, because if you concentrate all the...

This is just a narrative. I'm actually kind of curious what you think, right? If you look at the gains that technology has led to, it's concentrated those gains in these big tech companies and in the hands of people like Elon Musk, whereas everybody else sort of, I don't know, they get what's left. So from your standpoint, looking at the economics of this, talk a little bit about this narrative and what's actually true. Right. So

This narrative really took hold in the early to mid 2010s when we were just coming off the Great Recession and people really paid a lot of attention to the arguments of Thomas Piketty and some other French economists who claimed that inequality relentlessly rises until social upheaval emerges to push it back down. That was the Thomas Piketty thesis. And some people have interpreted the unrest of the 2010s as society's attempt to push inequality back down.

inequality increased by less than those people say. It did increase. Inequality has increased since like the 70s, but it has increased by less than those people say. And it is, it had cycles before, you know, where inequality increased and then was pushed back down. Inequality, so right around the time that this thesis started becoming very well accepted and popular,

wage inequality started to go down and wages started rising more at the bottom of the distribution than at the middle or top. And so we've seen that for now over 10 years. We've seen wage inequality go down. What is that? Like, how does that fit in the model? A lot of people when they hear this fact, they pivot to talking about all the white collar job, middle class jobs that AI is going to destroy and all this stuff. But no, don't pivot to that. Tell me why from 2013 to now,

Working class wages, wages for like simple laborers, waiters, clerks, just, you know, the working class. Why did these rise faster than wages for rich people, for educated professionals, all this stuff? Why? And so a lot of the same exact trends that people are now using to say, oh, AI is replacing the entry level workers, blah, blah, blah. Those are the exact same trends that broke the previous trend of rising inequality, right?

You know, it's like, oh, look at college grads, relative unemployment relative to like non-college people is rising and rising. Well, that that's because non-college people all get a job and their wages have been going up. So it's, you know, their wages haven't been going up because of scarcity. We imported a bazillion immigrants, as every like rightist will tell you. They're really mad about this. But then we imported a billion immigrants. We outsourced to China. We did all this stuff. Working class wages have been going up and up.

And if you look at when wages have stagnated, actually, wages stagnated in like the 70s and like 80s. That's when wages stagnated. The Rust Belt was real long before AI and before the internet, before really computerization affected anything.

So we can discuss why that happened in the past. But if you look at a graph of median wages or wages for like the working class, you know, the 25th percentile or something like that, it's like up in the postwar, then this flat period through like the late 80s and then up again until now.

You know, somewhere in like 1994 or something like that, it just starts going up again. And then that that, you know, the Great Recession broke that trend for like a few years, like three or four years. Right. And a lot of these negative narratives came right after that. So it looked like you had this this temporary dip, but then they started rising smoothly again. And I don't think Donald Trump did anything, man. I think our economy just righted itself. Fascinating. So why do you think the wages have gone up then? Because I believe productivity is gone because new technologies are

new technologies and new markets added to our productivity and allowed us to do more stuff. And then humans were compensated for that by getting paid more.

Now, have we seen a lot of inequality? Have people at the top risen more than people at the bottom since like 1980 or late 70s? Absolutely. Although that has, the wage inequality has gone down a little bit since like 2013. It's been on a slightly declining trend, but that declining trend has not been big enough yet to reverse the big run up in like the 80s, 90s and 2000s. And there's a large, I mean, there's an asset gap also, right? Of course there's an asset gap, there's an asset gap, yeah.

And assets have gone up way more than wages. Absolutely. Absolutely. But that's not productivity related. Well, it is, but it's more related to like financial market factors. So the big question then is if we are having decreasing inequality and wages are going up, then why is there so much unrest in the world? So I want to capture that and we will discuss that right after the break.

and we're back here on big technology podcast with noah smith aka no opinion you can find his work at noopinion.blog he also hosts a great podcast with eric torenberg it's called econ 102 you can find that on your podcast app of choice noah great having you here so before the break we talked about the fact that hey

It ain't that bad, right? There's been some, you know, decrease in the ability of entry-level workers to get jobs, but it's not catastrophic. There is inequality in our society, but there has been actually decreasing wage inequality, or the wages at the bottom have gone up. And yet it does seem like we are living in a society with a lot of unrest. It's certainly reflected in the politics. How do you square the fact that things are getting better in your view, but it's

politics is there's politics of grievance and things getting worse that's something I don't actually know I think it's primarily social media

I think that social media, so I think that Americans, America was this very big diverse country. You had ideological diversity. We had geographic diversity, not just like racial diversity, which is what we typically think about. We had all these forms of diversity where you could really spread out and find your people. Like in 2010, you could go, if you wanted to live a conservative life, you could go to Texas. If you want to live a liberal life, you can go to California. Yeah.

But then social media came and especially Twitter. And suddenly everyone was thrown in one little room with everyone else and not just people from America, but English speaking people from all over the world. And we have the global language, right? So we've got people weighing in from Argentina, from Bulgaria, from Pakistan, from, you know, UK and places like that.

And so immediately became this giant scrum where everyone was constantly exposed to people who hate them and despise them and disagree with them. We couldn't spread out, right? I think that this caused conflict in our society at every level of society because people who had been able to segregate were no longer able to segregate. And so social media is what caused our thing because every day social media will find whoever hates you and concentrate them directly in your notifications tab.

Right. That's what it does. But it's also it's interesting because the politics is all about, again, like if things if things are getting better in society and in the economy, it's right. I think you're right that social media has brought us all into contact with each other. But the question is, like, why has this messaging of of you're being wronged and your economic interests are not being looked after and they want to make you broke?

Why has that been this? I mean, it's been I know you could argue it's a culture war message, but there's also been an economic message that I think has really started to work in the US. Is it I mean, is it also the fact that like it's kind of this like FOMO culture where we are going to we're you know, we see all all the activities of the super rich all the time because it's on, you know, TikTok and Instagram or is there something is there something else that I'm missing?

Yeah, I think that is basically what's going on. Here's the thing. A lot of the conflicts that people immediately had were racial conflicts or things like that. Certainly during the first decade and a half of Twitter, first decade of Twitter, let's say, racial conflicts, I mean, there were some gender conflicts too, but I think racial conflicts dominated.

So obviously Black Lives Matter is about that. But you saw conflicts over representation and wokeness and white supremacy and all this stuff, right? You saw there was like, you know, years of racially centered conflict in America. People probably get tired of it and people probably want to fight about something else. Economic conflicts never really got addressed, never really got resolved. Right.

Well, let's talk about the sort of the racial conflict in the U.S. and sort of some of the puzzling things that are happening right now in the aftermath in our politics today. We're talking on Wednesday, June 25th. This episode will go live the following week. So stay tuned.

There will be many new cycles between now and then, but we have the results in from the New York City primary where Zoran Mamdami has won the Democratic primary. And you posted this on Twitter, and it's quite interesting that I think this is the racial breakup of people that voted for Zoran or Anwar.

Andrew Cuomo. So Hispanics, Cuomo won by 20%. White or Caucasian, Zoran won by 13%. Black or African American, Cuomo won by 44%. So can you explain what's happening here that you would imagine that Zoran is the spiritual inheritor of the Black Lives Matter protests and the woke left, and

And yet it seems like the minority groups within New York City are going towards his opponent while white and Caucasian groups are going towards him. Asian, by the way, Asian voters, 58% towards Zoran and 16% to Cuomo. So what's your read there?

I think that you're seeing partly this is education polarization. I think that white people are more likely to have like college degrees and stuff, you know, so there's been like a lot of sort of lefty ideas have really caught on among the educated populace that have not caught on among the working class. Barbara Ehrenreich called this PMC socialism. She called it professional managerial class socialism.

Right. And a lot of people, like infinite people, have remarked on this tendency. I think it's the most important political tendency in America today. But then certainly, yeah, I think that you definitely see just lefty ideas appealing a lot to the educated professional class. So I think you have an educational divide there. Also,

This is a Democratic primary. So you have people who are registered as Democrats. And I think in general, there's two reasons you'd register as a Democrat. One is that you feel that like Republicans are against your group of people and the other is that you're a liberal, right? And so I think that with white people, you don't think like, you don't think Republicans are against white people. So you'd be a Democrat if you're a liberal. But if you're black, if you're Hispanic, you might think, well, Republicans, you know, you might be a kind of conservative person who doesn't like like trans stuff or whatever. And then you might, you know,

You might think, but you might also think Republicans are against me, you know, and so I better be a Democrat, even though I'm kind of a conservative Democrat. So I think conservative Democrats are less likely to be white. So I think that's a, that's another thing that you see that's going on here, but I'm not an expert. So this is just me, you know, spitball. Of course. But it is, it's so interesting because the politics is tied to the economics in this, in this case. I mean, it was an economic campaign in many ways that,

Zoran ran and is running. And it's been interesting to see this professional. I think it's interesting, the professional managerial class coming out in support of him. You shared a tweet. It said Zoran's left wing base, so 400,000 here, hedge fund employees have concluded that capitalists, the bourgeois bodega owners must be instrumentally liquidated as a class.

via a fall in the rate of profit in order for socialism to triumph. Talk a little bit about that. Yeah, that was a good joke. Yeah, I know. So the point is that if you look at some of Zoran's proposals, one of them is to have government run grocery stores. Now, the reason Zoran wants government run grocery stores is because he has a deep seated belief that that government can run industries better than the private sector in many cases. That's why he's a socialist.

But also he thinks that poor people lack cheap food. He thinks that's a big problem for poor people in America, that they lack cheap food. Their food is too expensive. It makes sense because poor people spend a lot of money on food. You know, so of course the price of food is really important for poor people. We all know that.

America gets cheaper food than other countries do. And actually, we do a pretty good job getting food cheaply to poor people compared to like most countries. But that's not on Zoran's radar right there. So he's like, food is too expensive. And of course, when we saw inflation, if you see inflation rise across the board, if you see prices rise across the board, well, food is going to hit the poor people hardest, right? Even if prices rise for like, you know, like,

cars or whatever like it's not going to hit poor people as much because they won't spend as much money on a car or maybe it'll it'll hit some poor people like a lot of poor people drive obviously but um so what but like things that rich people buy like the price of like vacations it's not going to hit poor people as hard but food is going to hit them hard so it's like how do we get so so zoran wants things to be cheaper for poor people and he thinks how do we do that

Let's do grocery stores. And so he's kind of bought into this Elizabeth Warren rhetoric that grocery stores are taking all the money, man. They're making all this profit off your back and selling you this expensive food. Well, it's wrong. It's just factually wrong. Grocery stores make almost no profit. And like your corner store bodega makes like no profit basically. And your little indie grocery store in New York, which is actually what poor people mostly buy from. They're not buying from like a big chain. Poor people are buying from like a little bodega or like a little like independent supermarket in their neighborhood.

Those places, you know, sometimes their costs are a little bit higher because they're inefficient. They just don't make any profit, but they don't have these big supply chains and logistics that like Kroger has or something, you know? Walmart has. But even Walmart and Kroger, they don't make, like Kroger doesn't make any profit. Like Albertsons doesn't make any freaking profit. Safeway doesn't make any profit. These companies make like almost no profit. So like they're skating on the edge of like unprofitability here. And so like there's really not much more water to be wrung from the sponge.

So if you make government-run grocery stores, it'll do two things. Number one, it will out-compete all the little indie grocery stores and bodegas, thus like a meager source of capital and wealth accumulation for like poor people in poor neighborhoods. That's currently like how poor people do entrepreneurship. That's where they make a little bit of money so they can like advance to the middle class through starting like a little grocery store or bodega or something like that. Most grocery stores, most bodegas are run by like Latinos and Yemenis actually. Yeah.

In New York City. And so so you're really pushing out the you know, you're going to push out those people. And then also because American government is has this thing that the authors of the book Abundance talk about where American government hobbles itself. It self regulates to the point where government stuff costs more, costs too much.

because it does that the government-run bodegas are not even you know they're going to have to be subsidized in order to compete so it's just going to be like a fiscal drain if you want to have them actually produce stuff cheap enough to like sell to people it's just a point of uh it's just like a sort of a boondoggle you know can i make the so let me see if i can because we haven't really given like the argument for why someone like zoron might have uh risen yet or i haven't let me just make the argument and throw it out to you i mean

If you look at what's going on, we talked a little bit about asset inflation. It really is difficult to make it, I would argue, in New York and in many places in the United States,

largely because real estate prices are really high. And so it might just be as simple as, you know, if you look at the culture war stuff, you throw that out, throw the socialism stuff, the label out the window. Maybe what we're seeing in American politics right now, and I know this is not a full microcosm of American politics, but this flavor is starting to arise, is just that

people have found that they are doing like the quote unquote right things and still unable to make a go of it. The American dream is inaccessible to them. And when you look at the, you know, sort of joke about the professional managerial class, but a lot of people that have gone to college right now, they're settled with hundreds of thousands of dollars or a hundred thousand plus in debt. They can't afford a home and they're pissed.

And so maybe it's just that simple explanation as to why a candidate of this nature or candidates of grievance are winning. Yeah, but so Zoran, I understand that sort of anger over student debt, anger, you know, sort of middle class anger over student debt and, you know, expensive homes. I understand why that's a thing that people are upset about. But also, it's true that Zoran isn't really going to address either of those things.

Like he hasn't talked about it. It's not in his main policies and he doesn't have the power. I mean, the rent freeze is, I think, and just addressing that in some way. Well, is it? Because the rent freeze is for rent stabilized units only. Right. Right. How many people who just graduate from college do you think live in a rent stabilized unit? How do you get in that rent stabilized unit if you had it? Rent control doesn't transfer when you change tenants, right?

Correct. Okay, so you get your rent reset when you change the tenant. So yeah, maybe in the future, they live in a rent stabilized unit that then gets rent frozen in the future. But then they don't, they're not the people sitting on these like old apartments that have been rent stabilized forever. Those are all older people. And so also, if you're talking about the difficulty of buying a house,

You're then and you're talking about rent freeze as a way of appeasing the people who are pissed about the cost of buying a house. Well, then you're looking at long term renting of a rent stabilized unit as an alternative to home buying. Right. You're saying, well, I can't buy a home. At least I'm going to be able to have my rent stabilized and live here forever. I've seen the people who live in rent stabilized units for like 30 years in lieu of buying a home and they're kind of poor.

They're not, you know, they're not, they're not poor, poor, but they're, it's, it's a bit of a, you know, a shabby bohemian existence to stay in that, in that rent controlled space for 30 years. And some people are going to like that. And maybe some people, maybe some young people see that and they're like, wow, that's, that's, that's the best I can do. I'm not going to buy a home. But I think that people still want to buy homes. You look at surveys, millennials still want to buy homes. Zoomers still want to buy homes, even though they're too young to do it yet.

And so I think that, yes, I do think people care about these things. No, I don't think that was Zoran's main appeal. I don't think, I think that, you know, his main, like, yes, people, like young people would like a bit of rent stabilization. But I think that people like Zoran's attitude. He's handsome.

It does oftentimes explain a lot in politics. Yeah, he's a good looking guy. He, you know, he does a good video saying like, I care about you. People just want to vote for someone. And like Cuomo's a schmuck, you know, and a giant NIMBY and just basically like just a disaster of a politician.

And, you know, nothing really going for him. And so people look at Zora and they look at this young, handsome guy who says, I care about you. I care about these prices you're paying, blah, blah. And I don't think that people necessarily look at that deep economic impact and analyze like the economic impact of like, oh, how would government run grocery stores actually do? Or like, you know, would we really want free transit fare, you know, free bus fares and stuff like that? There's a go. This guy wants to reduce costs for us. He's like young, handsome guy. Looks cool. Does a good video. I'll vote for that guy.

You know, and if you look at surveys of people who are voting for Zora and a lot of them think like that, especially when your opponent is this complete schmuck. And so, yeah, let's talk about what the Republican answer to these to these conversations, at least coming from the White House. And there have been two policies that you've written a bunch about, which has been tariffs and restrictions on immigration. Let's start with immigration first, because I know that you're passionate about this one.

What do you think the net impact of the Trump immigration crackdown is going to be on the U.S. long term? I think the net impact of the Trump immigration crackdown will be fairly economically will be negative, but not huge.

Okay. The reason is because fundamentally immigrants are a source of both supply and demand for labor, for everything, for products. And the reason why a flood of immigration does not actually reduce wages is that immigrants are also buying things and the demand from buying things creates labor demand for native born people. And so immigrants really don't do much to wages. And what that means is when you kick out a whole bunch of immigrants, you're not going to do much to wages either.

what you're going to do is cause disruption. You know, and by the way, this also applies to prices. So like immigrants will demand more stuff, which puts prices up, pushes prices up, but they also supply more stuff, which pushes prices down. You know, you get a bunch of, like it's 1995 and you get a bunch of Mexican guys moving into LA. Yeah, they're going to bid for houses and your rent will go up. Okay, but also at the same time, they're going to provide like cheap lawn care and cheap like daycare and stuff like that. And those prices will go down. So immigration pushes up demand and it pushes up supply as well.

Okay. And so then what about – can I make the counterargument? I don't want to say it's going to have no economic impact because it's going to cause disruptions.

If you're in an industry, it's going to cause bottlenecks. If you're in an industry that relies entirely on immigrant labor, illegal immigrant labor, like some agricultural industries do, I'm not saying that's good that they do, and I think we shouldn't force wage laws and blah, blah, blah. But I'm saying that if you do have an industry that has become reliant on that labor, and then suddenly you cut them off completely from their source of labor, that industry will be disrupted and that will create bottlenecks in our system, which will create some disruption, could create some temporary inflation.

Let me just make this argument about the AI industry, because a lot of talent that's working on AI today is people coming from overseas to work in the United States to work at these labs. That's right.

and try to deport people. But I think that there's potential long-term consequence here where the best and the brightest who wanted to come to the United States aren't thinking about doing that as much. So I know it's hard to measure because it's a hypothetical, but I wanted to run that by you and get your thoughts on it. I think that that's really important. So cutting us off from skilled immigration would reduce our innovation, which reduces our long-term productivity growth.

And I think it reduces clustering effects, which affects productivity growth a lot. So like, where do companies want to invest? They want to invest where the smart people are. And if we don't have the smart people here, we're in trouble because we'll get less investment. So I think that when you look at the economic impact of low skilled immigration, it's not a lot.

Either way, but when you look at the economic impact of high skilled immigration, strongly positive. So we're really playing with fire if we start kicking out the smart people. Okay. And the other side is tariffs. We have a tariff deadline coming up. It's going to be on July 4th. No, July 4th is the tax and spending bill. And July 9th is the expiration of the global tariff pause. This is according to Bloomberg. Noah, from an economics perspective,

been the impact of tariffs so far? We have like some conflicting reporting. Some say everything is held up in like the L.A. and Long Beach ports. And others are like, well, you know, everything is back to normal. And Trump will, you know, everybody's anticipating that Trump will chicken out again. And you have the S&P 500, which, you know, the day we're recording is up nearly 4% on the year. It was down 15% on the year April 8th. So an unbelievable rebound from the S&P. So what's going on?

So I do think that the reason why the financial markets are doing well is because people have decided that Trump's going to chicken out and that these pauses will be permanent and then they'll announce, quote unquote, deals that allow us to cut all these tariffs and that basically Trump's a bag of wind. I think traders are betting on that. It's pretty obvious. And the reason why financial markets abruptly crashed when Trump announced the tariffs was because they thought this is for real. When they decided it wasn't for real, they bought and financial market prices, asset prices went back up.

However, do I think that that's really true? I think that there's some issues. There are some negative effects that these tariffs are going to have that even if Trump does chicken out, there's uncertainty about what if Trump doesn't chicken out on this or that. Also, it's not like tariffs haven't gone up. They have gone up just like much more modestly than Trump was threatening.

That will gum up a bunch of things. How big an impact it's going to have is hard to say because currency movements adjust and other things adjust to like balance it out. But like it's going to be there's going to be some pain from that, especially in certain sectors. And the threat of tariff driven inflation will lead the Fed to keep interest rates higher a little longer, as we just saw with Powell. And, you know, conflicting with J.D. Vance, all the administration is like, no, push rates down because they want like an economic boost.

So that's going to be a factor too. And then there's uncertainty. There's people like companies, how much do you want to invest if you don't know what the overall economic policy regime of America is going to be in 10 years, five years? What if the next person who comes after Trump, what if J.D. Vance gets elected? What if he does tariffs for real because he doesn't chicken out? Do you want to do an investment that takes 10 years to pay off if Trump's only going to be there for three more?

You know, and so like there's those questions. Are tariffs now just the consensus policy of the land? Have we just all become fucking protectionists? Can I curse on your show? Sorry. Yeah, yeah. You're welcome to. I agree. Yeah, definitely. Yeah. So then so I think that that those will erode our economic strength in ways that don't look like an abrupt crash, but they look like data getting like a little bit worse and a little bit worse. That's my prediction. But again.

Hard to say. It's hard to predict, which is, I guess, part of the problem. So, all right, look, before we end, I heard you in another podcast sort of make the case for why the United States and China are

I would say maybe more adversarial than a lot of people can see on the surface. I would love to hear you just kind of share that with our audience about like why the United States and China have interests that are competing and are potentially leading them on a collision course and what might happen as a result. Well, I think that that's now become conventional wisdom. The idea that China's leaders do not mean America well and mean America harm has become conventional wisdom in America since I wrote that stuff. My position has now become the mainstream.

And if you ask most people, the attitudes toward China are very negative, although that's moderate a little bit as people realize that Trump, our leaders suck too. But then China is ruled by old guys, by boomers. And they're feeling their oats from a massive run up in development. It's very hard to point to countries that got rich without starting wars and being aggressive. If you can think of any, let me know which ones those are.

I mean, Switzerland, like little countries maybe, but like big countries that became like great powers. China's building 100 nuclear weapons a year. 100 a year? 100 a year. That's a lot. That's a lot, man. A lot of nukes. I feel like once you get to 50, you're good. But I suppose not. Call them and tell them that. I don't know. You can argue that with them. But they are building 100 nukes a year. They're building one of the most impressive military buildup in the world right now and in absolute terms.

the most impressive military buildup the world has ever seen, ever, is happening in China right now. What are they going to do with that, do you think? What do you think? What do I think they're going to do with that? Fight a war, Alex. That's what that's for. Okay.

They'll fight a war with the war tools they're building. Otherwise, it becomes a sea of rusting metal. You know, the Soviet Union built up all these tanks, massive amounts of tanks, and then they never used it until Ukraine. And now they've used them all. Right. Chekhov's gun. OK, what do you do with a military? Well, we did. You know, we built up our military in the Cold War. We built up this giant military and then we went to war in 1991. We went to war in 2003. You know, Iraq war, war and terror, all this stuff.

Maybe the fact that we built up this giant arsenal was one reason we used it. Okay, it's like we got this tool better use it You know, we got all this stuff and so like yeah, man They they build that shit for a reason and I think it's gonna be Taiwan That's the likeliest. I don't know if they're gonna go to war It might be that they don't that we deter them that they have internal dissent that there could there's a lot of reasons Why would that we might not go to war? I don't think it's certain that we'll go to war. I really hope we don't that would suck. I

Great power wars even worse than normal war. And so we could we might not go to war but I'm saying that It's a reason why we need to establish deterrence and if we don't we're just being dumb, right? We're like being oh look look at those guys building that massive army over there. Cool How's that working out for y'all then they're like, you know, like bye motherfucker out. See ya. And then we're like, oh no Oh, no

I mean, this is a dumb question, but how do you establish deterrence against a country that's building 100 nuclear weapons a year? Well, I mean, like you said, if you have a sufficient number of nukes, you don't need to build 100 a year. You have enough to deter them. But then also you build up a large army and you get a bunch of allies. We need allies. And Trump doesn't realize this enough. And Trump administration's really under-emphasized this. You get a gang together. There's a big guy on the block, big tough guy. You get a gang. You roll deep. You know what I'm saying? Yeah.

And so we need to ally with India. We need to ally with Japan. We need to ally. I mean, we're already allied with Japan. We need to ally, but we need to strengthen that alliance and get Japan to like re-militarize Korea, obviously. Vietnam, we should be focusing on. Indonesia maybe could possibly be helpful to us in some way. Australia, obviously small, but they're around. Europe, you know, Europe needs to handle Russia, which fortunately they may now do.

So there's all these countries that we need to ally more with. And Trump has been tearing up, not just tearing up our alliances, but also tariffing people and just being a general asshole to everybody we like. This is going in the wrong direction. It's not establishing deterrence. But then Democrats tend to cut military budgets, which is also not great. And so we need to build up our own army and get the armies of our allies because China is very big and powerful.

We'll see how much China is inclined to use that military. Like in the past, rising powers usually have a small war before the big war. So you saw Germany had the Austro-Prussian War before they did the Franco-Prussian War. And then that was before World War I where they tried to take on everybody at once. Japan had the Sino-Japanese War and the Russian-Japanese War before it did World War II and tried to take on everyone at once. So you see we had the Spanish-American War, OK?

So you see countries have these little wars before they start the big war. So we'll see what China does. And maybe if China starts a little war, it'll wake us up to their aggressiveness. And I think that's one reason they're holding off because they don't want people to realize they don't want people to think of them as being aggressive. They want the chance to start the big war as the first thing. But politically, that's difficult. You know, big war is a big, big jumping off a cliff when you don't even know how well your army knows how to fight.

So let me just wrap up the whole conversation because we started with AI. We're ending with China. Let's just cross the two and then we'll let you go. Does this make the U.S.-China competition on AI much more meaningful than a lot of people are taking, I mean, yeah, giving it credit for? I don't know. I think AI will be very important. We need things like computer vision, autonomy. Obviously, autonomous drones are going to rule the battlefield. That will rule the battlefield. So we need that.

As for whether we need better LLMs, I don't know how militarily useful LLMs are. Daria would say a lot, but I cannot vouch for that. So you got me. But whatever it is, we need to be good at it. We need good military technology. We also need large-scale production because technology alone does not win wars as Germany found out in World War II when they had great rocket bombs and jet fighters and that did absolutely nothing to help them win. So yeah, no, we need as good AI as possible everywhere.

you know, military competition is only one of many reasons we need good AI. The website is noopinion.blog. Highly recommend you go and subscribe. And you can also listen to the Econ 102 podcast on your app of choice. Noah Smith, great to see you. Thanks for coming on the show. Good to see you, Alex. Thanks for having me on. All right. Thanks, everybody, for listening. And we'll see you next time on Big Technology Podcast.