Prof G Markets - America’s Economy Is Entering a New Era — ft. Noah Smith

Episode Date: July 24, 2026

Ed Elson and Scott Galloway are joined by Noah Smith for a livestream to discuss what AI issues he thinks deserve more attention and whether or not he is concerned about AI dumping. They also explore why he believes a sovereign wealth fund could help address wealth inequality and what America's fertility decline could mean for the economy in the decades ahead. For access to future livestreams, you can subscribe to our substack here.  Subscribe to the Prof G Markets Youtube Channel  Check out our latest Prof G Markets newsletter Follow Prof G Markets on Instagram Follow Ed on Instagram, X and Substack Follow Scott on Instagram Send us your questions or comments by emailing [e-mail] Learn more about your ad choices. Visit podcastchoices.com/adchoices

Support for the show comes from VCX, the public ticker for private tech. The U.S. stock market started history's greatest wave of wealth creation, from factory workers in Detroit to farmers in Omaha. Anyone can own a piece of the great American companies. But today, our most innovative companies are staying private longer, which means everyday Americans are missing out. Until now. Introducing VCX, a public ticker for private tech,

now available wherever you buy stocks. Visit getvcx.com for more info. That's getvcx.com. carefully consider the investment material before investing, including objectives, risks, charges, and expenses, this and other information can be found in the fund's prospectus at getvcx.com. This is a paid sponsorship. Hey, it's Ryan Reynolds here for MintMobil. Now, I was looking for fun ways to tell you that Mint's offer of unlimited premium wireless for $15 a month is back. So I thought it would be fun if we made $15 bills. But it turns out that's very illegal. So there goes my big idea. the commercial. Give it a try at mintmobile.com slash switch. Up-round payment of $45 for three months, $90 for six months, or $180 for 12-month plan required, $15 per month equivalent, taxes and fees extra.

Initial plan term only, greater than 50 gigabytes may slow when network is busy. See terms. Drop everything. The McDonald's, creamy parmesan and bacon quarter pounder is back for a limited time. So juicy, so cheesy. This burger demands your full attention. Made with 100% Canadian beef, topped with the creamy parmesan and black pepper sauce, Hickory smoked bacon and crispy onions. Oh, yeah. Try the creamy Parmesan and bacon quarter pounder today or treat yourself to your dream burger lineup

with our original quarter pounder with cheese or quarter pounder BLT. At participate in McDonald's restaurants in Canada, product availability varies by restaurant. Today is number five. That's how many cups of coffee per day are safe for most adults and may even have health benefits according to the American Heart Association.

Ed, what do you call sad coffee? What's that? Dispresso. A lovely james. joke to start our Friday episode. How are you doing? I'm doing great. I'm in, I'm in L.A., which I absolutely love. I would turn the camera around, but I'm worried the tech team would freak out. But I'm overlooking the Hollywood Hills, and I absolutely love it here. I love California. I love Los Angeles. It is 80 degrees, but it's dryout, so it just feels, I mean, when you and I were in

Miami and it was 89. It felt like we were in an air friar. Sufficating. Yeah. This at 80 feels pleasant. It feels... That's very nice. It's, yeah, I love it here. So I'm happy. I'm a little bit jet lagged. But, and unfortunately, what the podcast that start, you think I'd figure this out at 2 p.m. in London, start at 6 a.m. in L.A. So I was up at O Dark 100 hours this morning. How are you at? Enough of my I have enough about the Dispresso part of this. What's going on with you? I'm doing very well. I'm very excited because tonight I'm going to go see The Odyssey.

So I've been waiting for over a year. Have you seen it? No, I'm the same time. I want to see it in IMAX. I want to see it with my boys. So all of these things get in the way of actually... Going and seeing it. Exactly.

I just sort of cut my losses. We're not going to go see IMAX. We're going to go see it regular. But that's okay. I just want to get out there and see it. So I will give you my review. But, I mean, the reviews have been spectacular, 98% of rotten tomatoes, critics calling it a flawless masterpiece.

So I'm very excited about it. Yeah, well, I'm refusing to see it because a black woman is Helen and a transgender man as a warrior. That's just, that's just. That's the line. I'm being accurate to how you cast a myth, which to me is an oxymoron. You get to do whatever the fuck you want. It's a myth, folks. Anyways, I can't get over that bullshit.

I would like to know the percentage of people who are upset about the Odyssey casting and who believe that the Odyssey is a factual historical story. That would be a very interesting statistic that we should get some survey data on. But yeah, that wasn't enough to keep me away. I will be watching it, and I'll be very excited to see how it is. Who knows? Maybe I won't even like it. I think you're going to love it. Yeah, I've heard nothing. I go in with such a bias because two of my favorite films are from Christopher Nolan, Oppenheimer,

and I think the perfect movie, at least for me, whose favorite star of any movie is Hitler, is Dunkirk. I think you're going to say Interstellar, that's my favorite. Really? I think I need to do edibles and watch Interstellar again. I didn't watch it again. I kind of didn't get it. I re-watched it recently, and then I re-watched it again because it was so good, and I had forgotten how good it is. that is

profound and moving on multiple dimensions so that's my favorite Dunkirk's great too the Dark Night is up there for me great movie great movie

yeah Memento yeah Memento is good one of his early ones that guy has a bright future in Hollywood although I still am very upset the thing I didn't like about interstellar and it stuck in my craw what about the dude on the spaceship

who was up there for 22 years or 12 years just roaming around waiting for Matthew McConaughey and Anne Hathaway to get. I mean, think about it. They're on that wherever they were, and they come back after 28 minutes, and he's been there 12 years and they're like, hey, dude, what's going up? How you doing? What the fuck do you think is up? Nothing. I've been here for 12 years. I found that very upsetting. Very upsetting. That's kind of the point. But, yeah, I agree.

Well, we have a very good and exciting interview to get to get into today. We recorded this conversation live on Substack. earlier this week. If you'd like to catch the next live stream, we will be doing plenty more live streams. Then go to profitemedia.com where you can subscribe. We do a lot of these live streams. They're a lot of fun. I thought it was fun to do this live with an audience as well. So without further ado, I'm going to transition us into our conversation with Noah Smith. Now that we are past the halfway point of the year, we wanted to take a step back and look at some of the biggest forces shaping the economy. And to do that, we have invited someone we are a big fan of. He is a prolific

writer whose work spans everything from AI to inequality and demographics, and he has a way of explaining very complicated economic trends in a way that is both insightful and accessible, and that's why we really like his work. So we wanted to ask him some of the biggest questions that we are wrestling with right now, for example, are we in an AI bubble? And if so, what happens if it bursts is rising in equality, becoming a threat to economic growth, and also has America officially reshuffled in the global order? So, without further ado, here is our conversation with Noah Smith, writer of the no opinion blog. Noah, thank you so much for joining us on Prof. G. Markets and thank you so much for joining us on a live stream.

Very exciting stuff. Thanks for having me on. So I want to get started here with some recent AI news, which is the release of the Chinese model Kimi K3, which has caused a lot of concern in the AI world. I mean, what we know about this new model is that it's open source. It's very cheap, and it also outperforms Anthropic and Open AI's models on a variety of benchmarks, which has been.

A lot of people worried and thinking about this idea of, is China about to supersede the U.S. when it comes to AI? Are they about to steal America's lunch? Let's start with that question. What do you make of Kimmy K-3 and what do you make of the China versus U.S. AI race at this point? So with Chinese models so far, the pattern has been that they test very well on benchmarks and then that their capabilities are very brittle when you go beyond just sort of teaching to the test, if you will. That's one issue.

This has been a very consistent thing. So when we've seen, you know, more sort of more comprehensive tests of Chinese models' abilities, they've been increasing more slowly than the American models. There's also a question of, you know, to what degree this Chinese model is simply, you know, an American model that was copied. You know, so in terms of whether export controls are working, the overall story has been

again and again that export controls do work. In addition, China has far less compute available for inference. So, you know, most of what differentiates these models right now is, on their performance, is how much compute you can use for inference. You know, not how much you used to train the thing. It's how much you, you know, you can let them think, really. And so China has a lot less compute there, and AI and export controls have been very important in limiting China's compute on inference on the capabilities. So even if you exactly steal the weights for Claude, which maybe they did, who knows, I'm not a technical expert, so I can't say whether they, how likely that is.

Or maybe you use a distillation thing, so it's very similar to Claude. So even if you do that, you can't run at Claude's capability most of the time if you don't have the inference compute. So export controls have been extremely effective in terms of, and that's not even counting the export controls on chipmaking equipment, which have been very effective in keeping the Chinese chipmaking industry well behind the, you know, the Taiwanese, Korean, et cetera, and U.S. industry. And so I think export controls are still quite effective. And this may turn out to be another deep seek moment where we find out that the, you know, reports of extremely advanced, you know, like world beating capabilities turn out to be very exaggerated.

When you look at the AI landscape today, which are the issues or the topics that you think are not getting enough attention that matter more than perhaps the media or the conversation online give credit for? Well, I mean, obviously, AI biosecurity. Like, humans are vulnerable to viruses. And so there are, you know, viruses that kill all, you know, most of humanity or all of humanity don't naturally tend to evolve because those viruses will burn themselves out. But maybe you could make one if you intentionally design it. And we, our controls on biolabs are very weak. You know, we need to implement stronger controls on the places where these things could actually be made and disseminated in the physical world. It'll be a while before we get these fully automated robot labs. But humans can make stuff, and humans who insert some genetic mutation into a virus and then email it to you, or not email it to you, just mail it to you through the post office. They don't necessarily know that the mutations they inserted into common COVID-19 are going to turn it into a virus that just kills all humanity. They don't know that because they don't, they're not AI. They can't understand. So we need to deploy a defense in depth against doomsday viruses. And that's just not being done yet, although people are starting to talk about it more.

But if you're not scared about this, you should be more scared about this than you are. So one of our big themes or predictions is that China is engaging in modern-day dumping, but instead of steel, this time it's with AI. And I think there is evidence that token usage from Chinese models has exploded. Any thoughts on the notion of AI dumping? China could absolutely subsidize, you know, subsidize inference so that its models are temporarily cheaper and then this throws a wrench into the profit, you know, profit and loss of anthropic and open AI. And there's, you know, they cause them a crash, open AI and anthropic

go out of business or don't have enough money to like train the next generation of models. And then, you know, then China's eliminated some American AI advantage there. It's absolutely possible that that could happen. But at the same time, we have seen Anthropic profitability is expected to be positive on an operating basis this year. That's amazing for a company with that amount of growth. Usually if you're investing that much, you're not profitable. Like how long did it take for Amazon to be profitable? Many, many, many, many, many years.

Anthropics doing it much more quickly. Usage is exploding. I think Open AI will probably get there. And then so far the profitability of these companies doesn't necessarily seem to be under threat. If we see Anthropic and Open AI struggling to find the capital to invest in the next generation models, then we should be worried. Until then, there's nothing that AI dumping really does, except for attack, Open AI, and Anthropic. Now, maybe you can make an argument that it's hurting some of our players further down the line like Google could be getting hurt, or, you know, maybe

meta has been driven out of the AI race by the fact that they can't, you know, afford to keep up with Chinese inference because of dumping. So that's something that deserves to be looked into, but so far, if opening in Anthropic are still, if their profitability is still improving rapidly, I'm not so worried because that's the mechanism by which dumping would hurt us. I mean, it appears another theme we've seen is that we've gone from worrying about a supply crisis where there's not enough inference, compute, and energy to potentially a demand crisis because it feels like a lot of the initial demand creators who are supposed to create demand, whether it's Lama or XAI,

they have not created the demand they'd initially anticipated on the front end, so they've taken all of their infrastructure buildout and now are renting it to what appears to be a narrowing group of the demand side, specifically OpenAI and Anthropic. Do you think there's any fear that we're moving, or do you see any evidence that we're moving

from a supply-side crisis to a demand-side crisis? I haven't seen evidence of that yet, but it could happen. have not seen evidence of a demand-side crisis yet. Demand for all these things continues to explode. That demand is, you know, going to American companies. A lot of that. You know, we've seen demand for, you know, sort of cheap substitute from China also rise.

That's inevitable. You know, like I said, you might say that, like, oh, Gemini isn't as good. Google's falling behind because there's not enough. demand for their products because they're being, you know, as a slightly lower down the hierarchy, slightly behind the frontier model, they're being out competed by Chinese models with subsidized inference. So you can see that. I think the bigger danger on demand is, you know, sort of a, basically a collapse due to temporary overuse. So if you have every company that says, we need to token max, guys, we need to just use AI as much as possible to figure out what we can use this for,

And then they just token max, token max, token max, and then this drives demand up. And then, you know, OpenAI and Anthropic and Google and whoever make reckless capital budgeting decisions based on the idea that that growth curve will continue. And then it doesn't continue and it collapses because people are like, wait a second, we're not actually building shippable software. We're not shipping profitable software with all these tokens. What are these tokens going to produce? Then we can see a crash, right? And China's companies might be better able to weather that crash than us because of state subsidies, you know, keeping them alive. And keep in mind, by the way, that most of these state subsidies are delivered through below-market rate bank loans, not through the government mails you a check.

So that's how most of these subsidies are actually getting done. The United States has the capability to match that if we were to use, modify the Dodd-Frank legislation to get our big banks. lending to AI and also to manufacturers, by the way. If we want to do industrial policy, if we want to fund it, if we want to subsidize, it's not by mailing, I mean, mailing government checks like we did with the Chips Act can be effective, but I think the most effective and durable way of subsidizing things is actually through bank loans. And we already have the legislation in place. That's Dodd-Frank, to be able to do that. All we have to do is say, hey, guys, hey, JP Morgan Chase, if you lend to these strategic industries, whatever, you can get,

you can loosen your tier one capital requirements. And then they do. The loans start to flow. And then so we can do that. We've already done the repression part of financial repression. We just have to do the lending channeling part. If we want to do the kind of subsidies China does, we can actually match that pretty easily. There is an evidence that there is no demand.

I mean, we've seen the demand. We've seen how much money is flowing into this ecosystem. If I could kind of tweak it, it would be there is a concern that the demand that we are seeing, or at least the majority of the demand that we are seeing, is artificial. And you kind of hinted at it with the token maxing, where these companies are like telling their employees, like, doesn't matter about the ROI, doesn't matter how you do things. Like, you need to use as many tokens as possible. We had the issues with the circular financing stuff where you had these kind of circular deals where investors turned out to be various AI

companies, customers at the same time. You have the fact that, you know, a lot of the revenue that's being plowed, a lot of the business model is predicated on the spending of Open AI and Anthropic, who are currently money losing businesses. And so the money that they're spending, it's not their cash flows that they're spending, they're spending their investors' money. And then to your point, there's also the subsidies side of it. If you have companies in China who are receiving subsidies through whatever mechanism, and if we're going to start thinking about doing the same thing over here in America, that, translates to me as, yes, tens hundreds of billions of dollars in demand, but demand that doesn't

feel quite organic and that feels in a lot of ways forced or artificial, which makes me think, well, if we were to take our foot off the pedal here, then something might collapse, something might break in a big way. What do you make of that argument and that view? Do you think there's merit to that? Explain to me, artificial demand? Circular deals where you have Nvidia and OpenAI investing in each other and then being each other's customer. How does that create demand for the end product?

I don't understand. Well, it creates demand for Nvidia's chips. That's investment. So that's like, yes, it's demand for the upstream stuff. So right now it appears like a lot of the usage companies are because of these metrics around, such as we're tracking how many tokens you're using as opposed to tracking or putting in place real ROI.

And the thought of some of that demand might fall off. I think the more puncturing issue is, can they maintain this level of demand once they move to a pricing system that sustains? I mean, I use Claude, I think it's called Claude Pro, or the one that's $200 a month. Supposedly it's costing them $1,000 to $2,000 to service me. And the question is at what point, does this demand, can it sustain at anything resembling the growth that the street wants to see to support these private valuations once it's priced to its actual cost?

In other words, does this business model work? Do you think this business model works? I don't know. Fair enough. I don't have a way of tracking that. I don't know that other people do either. I think, you know, I wrote about that. I wrote a post about that, called how much software do we really need?

Like you, I hear anecdotes about token maxing. Token maxing being the term for the benchmarks, you know, the usage-based benchmarks that you're talking about instead of like profit or, you know, based benchmarks. Just how much input can we use? Like, that doesn't sound like a great benchmark for profitability. I hear anecdotes. I don't, you know, I have, I don't have hard statistics.

on how much this is being done, who's doing it. This is all internal. None of this data is public. And anecdotes don't equal data. And I think we should resist drawing excessive generalizations from anecdotes. I don't know how much value is being produced. Because value, first of all, it could be that lots of values being produced, but it won't be realized for a while.

So if a lot of this is learning, if the value of using this stuff is learning how to use it, we won't see that for a while. It's not going to show up like, I use this many tokens and I ship this much profit worth of software. I get this much profit from the software I ship with the tokens that I used. It won't necessarily be that clean. You know, if you look at how productivity is measured, like Chad Severson does a good job of measuring this productivity, but it's really hard, you know, like you can look at very established industries like ready mix concrete or something. And you can look at how you can really, you know the production process, you know the use cases,

you know exactly how this stuff is turned into value, you know, inputs are turned into value. But for AI, you don't. And so there's a lot of learning here, right? So the classic example here that everyone cites is electricity. When electricity was first introduced to manufacturing, by this time, everyone was using it to like light their homes or whatever. But when electricity is first introduced to manufacturing in like the 1900s, early 1900s,

they just, they had these vertical factories that were turned with these big crank shafts or the entire conveyor belt moved at the same speed because it was driven by a steam boiler. They ripped out the steam boiler, put an electric dynamo and drove the same thing, found out, wait, it's less energy efficient. Electricity sucks. They electricity maxed, right? they used an improper benchmark as there, you know, and then they tore out the dynamos, put back in the boilers, and ignored it. And then about 20 years later, people figured out, oh, if we completely

reorganized how the factory worked and ran workstations in parallel instead of in series, we could increase productivity a lot. And then you got a very rapid increase in productivity by a factor of like three to five or something like that, just this massive, massive burst of productivity growth with the reorganization of the factory. And so we got modern looking factories, not these like ancient-looking, you know, multi-floor factories that now are part of like old Mickey Mouse cartoons. And so was the initial use of the, of the, you know, electricity to try to replace steam directly, was that wasted money? I don't know. I can't tell you that that was wasted money because there was learning here.

Well, you learned what didn't work. And that was probably an important input into the, you know, people who redesigned the fact. later on and got this huge increase in productivity, they had to see what didn't work and think about what you could and couldn't use electricity for. And so I'm not sure that token maxing is wasted because all these people are learning what you can and can't use AI for, what it does, what it doesn't do. And it's forcing a lot of people to think about AI first production processes. So now we're, everyone's already talking about building AI first businesses and AI first business models. Maybe a couple people have, you know, my friend Edmund Zagrant has this company

where he's trying to do this. We're getting on this much faster than the people did with electricity because information travels, you know, travels much faster, right? So we know we need to do this faster than they knew that with electricity, okay? We're already on this. And so whether or not it's going to take us 20 years this time or maybe only five or something to build AI first business models that realize enormous productivity gains, I can't say. And whether that's even possible, I can't say. Like, there may be inherent limitations to AI technology once you scale things up from ask one AI some questions to have all these AI agents interact and do business together. There may, that may not scale up as well as I think as people think it does. I would guess it does

scale up, but maybe it doesn't. We don't really know yet. And I think if and when, suppose we do end up finding massive ways to massively increase productivity with AI, the blind alleys and dead ends that we spent all those tokens on before may have been an integral and important part of that. And so that productivity may be realized in a bit. Now, whether in the interim there's a big crash and what that implications are for producers like OpenAAAAAAAAAAAAAAAA, I can't say. And, you know, like, that's beyond my ability to predict, I think, but it could happen. We'll be right back after the break. And if you're enjoying the show so far, send it to a friend.

And please follow us on YouTube, Spotify, or wherever you get your podcasts. Support for the show comes from BCX, the public ticker for private tech. For generations, American companies have moved the world forward through their ingenuity and determination. And for generations, everyday Americans could be a part of that journey through perhaps the greatest innovation of all, the U.S. stock market. It didn't matter whether you were a factory worker in Detroit or a farmer in Omaha. Anyone can own a piece of the great American companies. But now, that's changed.

Today, our most innovative companies are staying private rather than going public. The result is that everyday Americans are excluded from investing and getting left further behind while they select few reap all the benefits. Until now. Introducing VCX, the public ticker for private tech, now available wherever you buy stocks. VCX, by Fundrise, gives everyone the opportunity to invest in the next generation of innovation, including the companies leading the AI revolution, space exploration, defense tech, and more.

Visit getvcx.com for more info. That's getvcx.com. Carefully consider the investment material before investing, including objectives, risks, charges, and expenses. This and other information can be found in the fund's prospectus at getvcx.com. This is a paid sponsorship. Support for the show comes from Upwork. One of the biggest breakthroughs for an entrepreneur is realizing that you don't have to do everything yourself. Upwork helps growing teams quickly bring in specialized freelancers so you can move faster

and take your business to the next level. Upwork is a one-stop platform to find hire and pay expert freelancers across more than 125 categories, giving you fast access to specialized talent without committing to full-time headcount. With Business Plus, you can access the top 1% of talent on Upwork and with AI-powered shortlisting, you'll get matched to the right freelancer in under six hours. No endless searching required. Upwork also cuts down operational hassle by handling things like contracts and payments in one place so you can spend more time running the business. Thousands of growing businesses already trust Upwork to hire flexible, high-quality freelance talent for everything from one-off projects to ongoing support. It's free to sign up and posting a job is easy.

Visit upwork.com slash markets right now and post your job for free. That is upwork.com.com to connect with top talent, ready to help your business grow. That's upw-O-R-K.com slash markets, upwork.com slash markets. Prime presents perfect pairings. What pairs with Reacher on Prime? Reaching for a sparkling punch delivered with Prime. It's an unbelievable feeling when you're enjoying a punch. At the same time, Reacher is enjoying

punching people. Wow, he's still punching people. Entertainment meets fast delivery. It's on prime. We're back with Profi Markets. Something we've been covering a lot related to this is just the public perception of AI. The more dollars we plow into this technology, the more we roll ahead, it seems the more pushback we're starting to see

in the general public and at the political level. We have obviously seen Kathy Hochel, the governor of New York, who has issued a pause on data center production. And it seems like that

is growing as a form of AI regulation. It's like just hit the pause button on data centers. But it also seems to be related to some of the wealth inequality issues that we keep on seeing a kind of getting worse every passing year in America, the top 1% of Americans control roughly a third of all U.S. household wealth, which is an all-time

high today. To what extent do you think that these two issues are linked? And how much of a problem do you think wealth inequality really is in America? What? What does it mean for the U.S. economy? When wealth inequality goes from, you know, when the top one percent has like this many trillions of dollars versus that many trillions of dollars, it's not a number humans understand. You know, regular people, you know, who are just trying to make their credit card payments, don't understand that. You know, you need data by, you know, and then you have, you can have like economists who

gather data on this and then argue about how much is really happening because it's hard to measure this top wealth. and then you have people arguing, I don't think the sheer number is like, oh my God, look at how much these people have. I don't think that's the way that wealth inequality becomes a problem in normal people's minds. I think instead people think about inequality of power

and people think about the companies that these people control and are these companies being run in our best interests. Everyone knows these people are getting rich. Small changes in exactly how much they're getting rich or even large changes in how much they're getting rich don't affect the basic idea that, you know, people are being squeezed,

at their jobs and they're being squeezed on their finances and they're being squeezed in their consumption and they don't necessarily see like and then they they they hear oh all these stories about how fabulously rich the people who own these companies are getting that's the way in which I believe wealth inequality comes home to real people that's what we need to address the idea of can the rich guys continue to do things that make them get richer while ever where other people struggle. And I think we need to attack that from two sides. Number one, we need to make sure that normal people don't struggle. Okay, if people can make their, like, if people have less credit card debt, and if people have, you know, more wealth in their houses, which is where normal people have

their wealth, by the way. Normal people have their wealth in the real estate market, not in stocks. But if people, you know, but also we can give people like, you know, wealth in terms of like a sovereign and wealth fund, what Alaska does with the permanent fund. But the other thing we need to do is to make sure people know that these rich guys are ultimately under the control of the people through the government. That the government, although the government doesn't necessarily tell you what to do with all your businesses, central planning, at some sense, these big corporations are serving the interests of the nation. And I think that during World War II in the early Cold War, people had a very strong sense that big companies like Ford, U.S. Steel, Alcoa, GM, and all these Boeing.

and all these companies were serving America. These were American companies working for America and working for the people of America through the nation state. We need to reestablish that. So Sam Altman floated the idea of a government stake in Open AI. He's right.

And that's how this has to happen. And so sovereign wealth fund is really the best policy we can do to do all of this. So sovereign wealth fund says, so Alaska has the Alaska permanent fund. You know what that is, right? The oil, right? So it takes Alaska,

oil revenues and natural resource revenues and redistributes the money to the people of Alaska. We need to do that with, with, you know, tech and AI and these strategically important industries where we don't, you know, have the government micromanaged and central plan and control business decisions,

but we redistribute some of the capital income. And if necessary, we could. If our companies were going rogue and doing stuff that was against the interest of the nation, we could have, you know, the government as a shareholder can sense in one of the shareholders' meetings instead of just regulation and law. In addition to that, in addition, we can have someone go to the shareholders' meetings say, hey, guys, you need to act in the interests of America and the people of America. And that's what we need. That we need people being reassured that they have some

sort of stake in this, not just financially, but also in terms of power. I think that solves the problem. I would push back on a no. I think sovereign wealth funds are used when, a nation that's blessed with an inordinate amount of natural resources has excess capital beyond their spending commitments. A sovereign wealth fund in the U.S. would be nothing more than issuing debt that Ed's going to have to pay back to fund the sovereign wealth fund such that a current administration could pick winners and losers, which governments usually aren't very good at. I understand the need to invest structurally in certain industries that have defense implications, but generally speaking, when the administration decides we should buy 5% of open AI,

with debt-fueled capital that will create regulatory capture and disparage or diminish the other competitors. I just can't find evidence of where that makes any sense and ultimately doesn't just end up with a warehouse full of unsold DeLoreans or Air France. A sovereign wealth fund, I just think is indefensible at this point in picking winners and losers by an administration, many of whom have no experience in business or their experience in businesses, bankruptcies, and failed subcontractor. So again, sovereign wealth fund, I think it's a terrible idea. Where am I wrong?

If you don't want a sovereign wealth fund to pick winners and losers, you can just have it by index funds. I invest in stocks, but other than, you know, I invest in friends startups, but I don't really sit there thinking, I think Ford's going to do well this week.

You know, I don't do that. I don't pick winners and losers. I buy the index, you know, and then I buy the, you know, Vanguard index fund. and you can have the government do that too. The only winner you're picking then is America itself. You're making a bet on America.

You're not betting on American stocks over foreign stocks. That's fine. But then you're not picking winners in terms of industries or companies or things like that. So if you don't want your sovereign wealth fund to do that, then you don't have to have it do that. It's really easy to not have it do that. In fact, the sovereign wealth fund proposals that I've heard involve explicitly buying the index.

don't, you know, so I think that ad hoc stakes in companies like Open AI are dangerous for this reason, but then you can easily solve that just by buying the index. So if you look at the labor share of income, it's going down, it's going down by any reasonable measure. And so if you buy a very broad index of stocks, all you're doing is buying the capital, some of the capital, share of income and then redistributing it. So you're doing that. So that fear, I think, of picking winners and losers. I also think that picking winners actually has a much better track record than people think. I think that the idea that this always, picking winners always fails, it doesn't. It has some pretty predictable failure modes that we see a lot, sure, but it's not

like the simple easy nostrum of, oh, picking winners always fails, guys, if, if, if, if, If I can label something picking winners, then it's something we shouldn't do. That's been proven wrong again and again. But that said, we don't have to pick winners with a sovereign wealth fund. We can just buy the index. So that's not a worry, I think, you should worry about, or it should be a reason to encourage a broad sovereign wealth fund that buys index funds rather than the kind of policy Donald Trump is doing with ad hoc stakes individual companies, which.

I do have very strong misgivings about. I think that makes sense. You're saying, as opposed to picking winners and losers, support the U.S. economy and also give consumers or American citizens similar to that. I think it's the Super Return Fund or whatever it is in Australia, a chance to participate in what has been largely compounding. I guess the issue is, or the caution I would throw out there, I don't want to get your response, is that we don't have surplus.

So any investment, any incremental investment in an index fund or a company is going to come from debt, which will ultimately be taxes on either current or future generations. Why not just try and maintain lower deficits and let individuals decide what companies? In other words, ultimately that's saying that the government is a better stock picker. And indexes have effectively become stock picking. When you buy the S&P index, you're buying 40% of it's going to seven or ten companies. Isn't this taking, isn't this me saying, infantilizing the public saying, I'm going to run up your kid's credit card, and then I'm going to pick, I'm going to invest in index funds rather than just getting government out of the way

and letting you buy your own index funds? The main difference there being that most people don't have the money to buy index funds that don't have the liquidity to borrow money to buy index funds. And so if the government does it, how does the person who doesn't have money? It's a redistribution. Okay, so it's a redistribution because then you would have some sort of mechanism for returning the gains to citizens. Yes, that's the, yes, that's the whole point, redistribute capital income to regular people.

So you got some person who works at a gas station, that person could theoretically, you know, like scrimp and save and buy a tiny bit of index fund, but the transaction cost are high, the information costs are high, they don't understand how to do it. And practically, they can't borrow. Like, what does the government borrow at? A couple percent? Like, even in, now that interest rates have risen a ton, what's the 20-year treasury rate? I don't even know.

It's like a few percent. Do you think the gas station worker can borrow it a few percent to buy an index fund? You're crazy. That's absolutely not. There's two pieces here. Number one is it's redistributionary, because if you, if you have the government by services, you know, like pay for old people's health care or something, or maybe build highways or whatever, or provide for the national defense, if you have the government

do those things that benefit the gas station worker just as much as they benefit a rich guy, okay, that's redistributionary. And it's, and so that's the first benefit. The second benefit is insurance. So if you, debt is unconditional. You know, it's like you pay the same interest rate back whether or not AI takes over our economy. But if you buy the AI company, part of the AI company, and then you're able to redistribute that income in the future, it functions as insurance. Because in the state of the world, it's a state contingent asset. Because in the state of the world where AI takes over the whole economy, normal people get a piece of that. But the amount of debt that taxpayers have to pay back doesn't change its state uncontingent. It doesn't change based

on whether or not AI takes over the economy. Sounds like you'd be a fan of Trump bonds or baby bonds. Yeah, baby bonds. I mean, that's been done in other countries and it's been, it's been successful. Yeah, absolutely. 100%. And that's exactly what you do. That's, that's a thing you can do. So, so you can choose to pay down America's debt if you think inflation's a problem and you need to, you know, pay down some of the debt to reduce inflation. You can use it. You can do baby bonds if you think it's better to just give people cash, do, you know, UBI, fund UBI. You can also just do public goods. You can use this for research and development spending, you know, to improve productivity. And that Paul Romer sense, you can use this to do national defense if we need to build up for big war. You can use this for building highways. If we need better infrastructure, you can use this to build trains. You can use this to build, you know, critical infrastructure. And you can use this for in-kind redistribution, like buying old people health care, which is, you know, not as broad-based as baby bonds, but it's pretty broad-based because everyone gets old and everyone gets sick. And so, you know, you can do a lot of these redistributionary things. So there's the insurance benefit,

there's the redistribution benefit. And then there's the fact that the government's borrowing costs, you're talking about running up these kids credit card, but you're not talking about their credit card, because a credit card interest rates are through the roof, where the government interest rates are a few percent. Even after we raised interest rates to fight inflation, the government is the cheapest borrower.

Right. And unless you believe that there's a good chance that the government is going to get below treasury returns on its stocks over the maturity of, government debt that it takes out to buy these things, which is a possibility. I mean, that's, we haven't seen negative returns, or we haven't seen below treasury returns, I think, on 20-year, you know, S&P, whatever, ever before, but we could see it for the first time. It's not out of the realm of possibility. We could take a bath. We could actually lose this, right? If you use state

uncontingent debt to buy state contingent assets, there are states of the world in which you take a bath. You lose money. Okay, the government can lose money. And this is the argument that people used to stop Bush from investing social security funds in the stock market back in the day. But had he done that, we would have either less federal debt or much larger social security benefits, a much more about social security trust fund today than we do. It was actually the right idea. And people use, you know, sort of the specter of risk to to dunk on it, to stop it back in 2005. And yes, there is risk, but it's aggregate risk. Because if corporate America is doing so bad that the stock market does really bad over a 20-year period or whatever, even a 10-year period,

but especially a 20-year period, then America as a whole has done badly. It's hard to see a a case in which the stock market does crap, underperforms treasuries for 20 years in which the regular Americans are thriving. That's a weird state of the world. It's a state of the world I'm willing to bet a lot against. It seems like you and I, I mean, you would agree that the Trump accounts have been a good idea.

I person, I've talked about this on our show. It's one of the only policies that I've seen out of this administration that I really support and that I think is moving in the right direction for a lot of the reasons that you're describing. One, I just want to confirm that you think that Trump accounts are kind of the right idea. I wish we could just change the name because I don't see why we need to politicize this thing, but the idea of gifting a child at birth, an investment account, automatically buys them into the S&P into the aggregate capital markets of America.

That, to me, seems like a very good thing and a step in the right direction. So one, I'd like to just get your reaction to that, And then two, we're tackling like the biggest, most difficult, most unanswered question in economics, which is like at what point is U.S. debt unsustainable? How bad is it really? I'll pose that question to you. I know it's kind of unanswerable, but I'd like to get your take. U.S. debt is on a very bad trajectory right now. That said, if you take out debt to buy U.S. stocks, to have the government by U.S. stocks, and you have the option to

use the proceeds from the stocks to repay the debt. If you structure the sovereign wealth fund that way, then it is not, it would be more sustainable than it is. And, in fact, this is what Japan has done. So, Japan, on paper, has a higher debt to GDP ratio than we do.

And, but they've taken, you know, they've had the government invest like a hedge fund and made very good returns. And because of that, their debt is less of a burden than ours is now, even though it's higher as a, you know, headline share of GDP. We would be the same. We'd be betting on us. I mean, Japan often invests foreign assets. But sometimes in the Japanese market, too. So, you know, the BOJ buys stocks and Japanese companies and things like that. And the, you know, the topics

has done well. So even if we took out more debt than we have, in order to buy shares in corporate America, that would, in every state of the world except the long-declined stagnation of America state, in every state of the world except that, it would make our debt burden less of a problem, right? Just like if you have a bunch of debt, okay, and then you take out some more debt to buy stocks, and you're able to hold that for 20 years, right? then your personal financial position actually just improved. You can't do this because of liquidity constraints. The government doesn't have those liquidity constraints.

Curious to get your take on the controversy around data centers. Some of the local harms of data centers that people tout are real, for example, if you're very close to it, it's noisy. It does draw power from the grid. And that can compete for power with local communities. Some of the things people worry about are not real, like water. Water usage, that's not a real thing. Data centers have a reservoir of water that they run to cool stuff,

and they lose extremely little water, and they don't pollute any water. So the water thing is made up. The other harms that people tend to emphasize a little bit less are actually real. If you don't want data center near your house, that's your right. You can, you know, I don't. Because I think data centers can go elsewhere. Like, you know, if New York won't build the data centers, Texas will, Alabama will. Places will.

They bring benefits to the local economy that can then be seen in tax revenue. Local tax revenue, you can tax them. So I'm not so worried about this data center nimbiasm, first of all. That's my first take. My second take is that I believe without solid evidence, so, you know, I don't have solid evidence for this one. But I believe that the local nimbism over data centers is to some degree, to some substantial degree, not all, but to some substantial degree, an expression of broader fears over AI itself. That people don't like the idea that AI will take their jobs, you know, basically take their jobs. I don't think most people think that AI will kill them yet.

I think people are mostly scared of jobs. And that fear manifests as an instinctive opposition to AI that filters through to opposition to local data centers. I think it's ineffectual opposition because banning local data centers in New York or whatever community won't actually reduce the amount of compute available to the AI industry. But I think that that's what's motivating it to some degree. We'll be right back. And for even more markets content, sign up for our newsletter at profjimarkets.com. Gregus believes in the art of living in choosing the pleasurable over the practical. It's why we craft our premium vodka in France.

using pure spring water from Jenzac. And why you hear the celebratory club of our quirk reminding us to make every martini cocktail count. So go on and make every moment count because there's no better time than now. Gregoose. Make time wait. Sip responsibly.

There's a civil war happening in the Democratic Party and if there's one place that that's playing out most clearly, it's in Michigan. A crucial Senate primary. battle in Michigan that could determine control of Congress in November. Congresswoman Haley Stevens and Abdul Al-Sayed. A progressive Democrat and a moderate Democrat. In the end, it all comes down to the dreaded e-word, electability.

But in Michigan, one candidate is trying to turn the electability concept on its head. If we think that voters walk around asking, where do I sit on some theoretical left-right spectrum, then in theory, the bulk of the voters are somewhere in the middle. The problem, though, is that that model hasn't really accurately predicted our politics for a very long time. If Michiganers wanted moderate, why would they have elected Donald Trump twice? Dr. Abdul-El-Sayette is making the progressive case for America first, and he's trying to settle the Democrats' ideological battle in the process. This is about the many versus the money.

I'm Instead, Herndon, and this is America Actually. Catch us every Saturday on YouTube or wherever you get your podcast. One, two, day. I'm stand-up comedian John Marco Seresi. And I'm actor-penis model, Russell Daniels. The downside is our podcast where we bring on guests to talk about how miserable their lives are. Because let's face it, things are not getting better. Every episode we talk about what's wrong with our lives, our guest lives, the world.

But in a fun way. Bottom line is, you're going to walk away feeling better about your life. We've had so many cool guests, Caleb here on. Busy Phillips. Stavros Halkias. Byrne Cox. Hassan Piker. Alana Glazer.

I promise you're going to have a good time. Now on the Vox Media Podcast Network, this is the downside. We're back with Profty Markets. A trend that we have been watching and kind of following your coverage of is the fertility crisis in America. The fact that the U.S. fertility rate fell to another record low in 2025. How big of a problem do you see this, this fertility? crisis and what kind of impacts do you think this will have on the U.S. economy, perhaps some impacts that maybe people aren't seeing or recognizing right now? It's going to be bad,

but the bad is going to be, is going to slowly trickle into our economy in ways that it's hard to notice. It's not like there's some cliff we fall off and like, ah, you know, fertility's low. It's, you know, old people will clog the top ranks of companies and will be more less nimble, less able to invest in the future. You know, less, our companies will get just less creative and dynamic. Slowly, you know, the debt burdens and tax burdens will go up and social security benefits will be cut. And, you know, there'll be the slow trickle of making it harder for, you know, just naturally harder for people to live because you have fewer workers supporting greater number of retirees. People have to do more uncompensated elder care. People will

have to pay more taxes to support. Elder care. You'll see immigration become actually more contentious because on one side, you'll see people say we need to bring in more immigrants for labor shortages. On the other side, you're going to see, like, immigrants will go up faster as a percent of population because we're not reproducing local people as fast. And so that's going to freak people out more. So you're going to see immigration become more contentious, even than it already is. So that's going to be difficult. There's just all these problems you get from a low TFR that trickle in slowly and subtly, and you don't see them like, like, oh my God, disaster. It's corrosive. Do you think that this is the kind of thing that will force us to sort of rewrite how we think about

Social Security, how we think about taxation, like, is this one of those era-defining events that may change the way we organize our entire economy? Yes, but I think it's less of an event than a trend. It's going to happen slowly. We already should have realized there's something very wrong with the pay-as-you-go method of social security, because as, as Paul Samuelson famously proved, in his paper about dynamic inefficiency back in, I think, the 50s or early 60s, social security, a pay-as-you-go system like Social Security is efficient, you know, if and when you have a growing population. When you have a very slowly growing population or a shrinking population, in the long-term, Social Security is dynamically inefficient. You just, it has big economic problems. And so that's,

that's hitting already. That's hitting already. And so, so pay, we needed to shift to a fully funded system. We can't now because it's expensive one-time shift. And instead, we're going to have to have constant bruising political battles over benefit cuts and retirement age raises and, you know, lifting the cap on taxes and making the system more redistributionary and all these things. We're going to have all these fights over. There's no way to avoid these fights. And it's baked into our system. And so, yes, it's happening, but it won't be a thunderbolt that happens all at once.

Is the best solution, in your view, some form of redistribution? Is that what we need to all get on the same side on? when you describe these kind of political debates that are inevitable, is that the main thing? It's absolutely part of what we need, because when you have benefit cuts, that hurts. I mean, if you're rich and you get your Social Security benefit cut, benefits cut, do you care? No.

But if you're poor and you get your Social Security benefits cut, you starve. Like, that's all you have. And so, obviously, the more we cut benefits, which we'll need to do to maintain the solvency of the system, the more redistribution will need to do to prevent human suffering. So the insolvency of Social Security creates a greater need for redistribution across income gaps. And so I think that's absolutely one thing we need to do. We obviously need to be looking for ways to boost the fertility rate and to continue bringing immigrants without, you know, freaking out, freaking people out.

We need to do both those things, especially high-skilled immigrants, you know, add lots and lots and lots to our revenue, to our coffers. They're a huge fiscal plus. So we need to look at how to do more of that. which we're really not doing right now. It's the opposite of what we're doing right now. And we need, so we're going to need more redistribution. We're going to need to tighten our belts and do more redistribution at the same time in order to keep the system solvent and keep it supporting old people and have old people

not be synonymous with poor people as they used to be. And at the same time, we need to be looking to address the root causes of the problem here for the long term. What would you say to someone who believes that redistribution is very, very, very important, on socialism. People who say, I mean, obviously, we've seen the wealth tax proposals, we've seen the billionaire tax proposals, which are kind of way over here on the extreme end of things. And a lot of people say, that's not the right direction. And I think those arguments might have some merit. But there are also arguments that, you know, increasing taxes,

redistributing in and of itself is not the right path. It's not the right solution. We're kind of waiting on government, waiting on some centralized system to solve all of our problems, to the people who take that view, what is your response to them? My response is, what's your alternative? What's your better idea? They have nothing. Nobody who just like, well, that would be socialism has any, says any alternative other than just like, you know, more money for my bank account. I think we're using the words incorrectly, though. Socialism is supposed to mean, at least.

government controls the means of production, which just doesn't have a good track record. And then when you ask, where does it work? They'll point to Northern Europe. I would argue Northern Europe in many instances is more capitalist than we are. I think what most socialist today really mean is a capitalist economy that has a stronger social safety net, more progressive taxation and more investment in social services. It feels as if a lot of this argument is an inefficient argument around semantics, that those of us who, think of socialism as east versus west Germany, North Korea versus South Korea, Soviet Union versus, you know, or USSR. I mean, the 50s and 60s here had much higher tax rates, much more public infrastructure investment, great society.

But we didn't call it socialism back then. We just called it healthy capitalism. Isn't a lot of this words? It's all words, yeah. You're exactly right. Like the word socialism isn't magic fairy dust that you can sprinkle on something. to like make it go away. Like that that that might have worked for like a few years in the 1980s, but but that dog will not hunt now. And so you're right. It is it is totally semantic. And that's,

isn't that just socialism is a, you know, like there's a shrinking and rapidly aging cohort of people for whom that sort of argument carries any weight at all. Yeah, you're talking to one of them. I love that. But I'm glad that you, you, you and I are on the exact same page about, you know, the idea that ownership and control of the means of production and is a very different thing from, you know, redistribution like social democracy, kind of the stuff like Sweden does. If you look at the Scandinavian countries, those countries are less regulated than we are. That's exactly right. They're more capitalistic than we are on the production side, and then on the redistribution side,

they do a lot more. And so is that more or less socialist than us? It is a totally semantic argument, and I wish people would just stop, like saying, well, isn't that socialism? And they soon will because that, that, you know, like, that, that's going away. We need to come up with a new term. I thought what George Bush did was pretty, or W, compassionate conservatism. Isn't there, is there a room for compassionate capitalism? Woke. You're talking about sloganeering and I'm not necessarily good slogan ear. I want a bumper sticker, Noah. I'm not good at bumper stickers, except for maybe silly ones.

But then, but I think that what I would, I would just use the word redistribution to, use a word to mean what it means. Instead of tying it to a whole package of ideologies and stuff by calling it socialism, just call it redistribution, which is what it is. Call things what they are. Use, you know, terminology to mean specific, useful things. My final question, if you know it. Is there any economic trend that you're seeing today that makes you optimistic about America

that makes you bullish on America or the world or humanity? at large. What are so? I mean like obviously AI, you know, is the big reason for optimism. You know, productivity growth, we're already starting to see some productivity growth that's probably from AI. We're not sure. Without apparent job loss. You know, like all these people thought AI was a job destroyer and a human remover. And so far, it's turning out not to be that. Like, you know, there's a few occupations that are under stress and under pressure from, from AI. But that's true of any technology. That was true of machine tools. That was true of anything that improved our prosperity in the past that now we're not worried about. It's true of databases and computers.

It was true of the internet. It's true of everything. And AI, so far, like our job markets, you know, there's less hiring and less firing. How corporations hire and fire people is going to change. I think we'll go more toward a Japanese system where we have like salariman doing generalist jobs for companies and like, you know, instead of highly specialized things and a whole lot more independent entrepreneurs. We're seeing a lot more independent people start small businesses. There's a positive trend. We're just seeing a massive surge in dynamism, a surge in business formation. Individuals can start businesses more easily than ever before. My substack is a tiny business. You know, it is an S-corporation. And so we're seeing a ton

of that, and we're seeing no big negative effects so far of AI on aggregate employment outcomes, and yet we're seeing apparent productivity acceleration. And that is great. that's a much better outcome than we could have, than people would have hoped for even two years ago, even one year ago. Even now it's better than what people think is happening. You know, people think AI is out there destroying jobs, and so far it's not.

And so that's something we need to be more vocally acknowledging of. We need to say, like, look, guys, AI is not destroying jobs, and yet productivity is accelerated. And there's all this dynamism, all these people are starting these small businesses. That's a big reason for optimism. We will need to have plenty more conversations in the future, but for now I will have to wrap us up. Noah Smith writes the No Opinion blog covering economics, technology, geopolitics, and culture.

Previously, Noah was an economics PhD student at the University of Michigan, an assistant finance professor at Stony Brook University and an economics columnist for Bloomberg opinion. He left Bloomberg in late 2021 to blog full time. For more of his work, head to Noopinion, that's Noah, his name, pinion, dot blog, to subscribe. to his newsletter, Noah. Thank you so much for joining us today, and thank you to our live audience for tuning in. We will see you all next time. Yeah, it was great to come on, and please have me back any time. Congratulations on your success, Noah. This episode was produced by Claire Miller and Alison Weiss and engineered by Benjamin Spencer. Our video editor is Jorge Carty. Our research team is Dan Shalon, Kristen O'Donohue, and Mia Silverio. Jake McPherson is our social producer. Drew

Burroughs is our technical director, and Catherine Dillon is our executive producer. Thank you for listening to ProfG Markets from ProfG Media. If you liked what you heard, give us a follow and join us for a fresh take on markets on Monday.