00:00:00.000Adam Becker. Welcome back to Dreaming Against the Machine. I'm your host, Adam Becker. This week's guest is Matt Schurer. He's a fellow at the Open Markets Institute and a friend of mine. And he is a man on a mission. He is working to make sure that when the AI bubble bursts, as it will, we do not have a bailout for the big tech billionaires who are behind it.
00:00:25.880And in fact, he has a report titled No Bailout for Big Tech Billionaires that there's a link to in the episode description.
00:00:34.140So I understand that this may sound like it's a kind of depressing subject for an entire episode of this podcast.
00:00:41.800But trust me, it's a great conversation.
00:01:12.840So why don't you go ahead and introduce yourself for our listeners?
00:01:16.260I am a fellow at Open Markets Institute. I started this fellowship at Open Markets Institute at the beginning of the year. It's focused on the AI bubble, or what I certainly think is an AI bubble, and trying to both explore the breadth of it and try and convince people, especially in the policy world, that the risks that it presents are serious and should be treated as such.
00:01:40.740okay so really sunny stuff uh and easy to talk about too i'm sure well there's certainly plenty
00:01:48.540oh well yes there's plenty to talk about but yeah no i think and wait hold on before we dive into it
00:01:54.620your background i'm a lawyer by training okay uh which uh if you look at my resume there is
00:02:02.040absolutely no reason i'll admit that anybody would think that i am qualified to write about
00:02:06.460macroeconomics and the threat posed by the ai industry okay but but hold on just to interrupt
00:02:13.360to make a joke but are macroeconomists qualified to talk about macroeconomics
00:02:17.640i do not think that they are the funny thing is i what i always describe this as is like
00:02:26.600there is such a thing as being too close to the subject you are studying i do not by any
00:02:32.500stretch of the imagination want to suggest that there are not trained economists out there who
00:02:37.220are not as worried about this as I am. But by and large, the economics profession, number one,
00:02:45.880when you are looking at the sweep of what you've been working on over the previous 5, 10, 15, 20
00:02:52.780years, there's a tendency to kind of forget the more distant history of how things can go wrong
00:03:01.360when, uh, you, you invest too much money too quickly, uh, in something. And there's also
00:03:08.740just a tendency to get sucked up into group thing. And I think that that is absolutely
00:03:14.920what has happened with the AI bubble. And I, the way that I described is how I kind of got into
00:03:22.900this. Well, number one, I graduated from law school in 2009, which means that I was looking
00:03:29.660for a job right around september 2008 yeah and i'm sure that went really well so i i ended up
00:03:38.080employed so that is more than a lot of my classmates yeah yes recently but but i but that
00:03:43.840was by the skin of my teeth i got i the offer from the law firm that i was planning on working for
00:03:48.620never came um and the truth is from that moment i became kind of like what the hell is going on
00:03:56.260with the economy right now? How did we get here? Why is this happening? And I've kind of been
00:04:01.840somewhat very closely following both what's going on currently in economics in any given time since
00:04:08.140then. And I've done a ton of reading on both what happened in the run-up to the last financial
00:04:13.180crisis and what happened before previous financial crises. As a lawyer, I started writing and
00:04:19.960speaking on AI around 2015, right when it was becoming kind of a bigger deal in the public
00:04:26.540consciousness. At the time I was an employment lawyer and eventually I kind of found a way to
00:04:31.480meld the two. And I became an employment lawyer who dealt with AI related stuff in the workplace
00:04:37.840and labor market. And eventually, you know, a long, long story short, I ended up doing that
00:04:44.000at the center for democracy and technology for about five years. I led their workers' rights
00:04:47.660project. But as I got late into my tenure there, trying to convince policymakers to enact regulations
00:04:56.640and guardrails around the use of AI in employment, in consumer settings, you name it, what I kept
00:05:04.240running into over and over again is no, AI is the future. We do not want to be viewed as stifling
00:05:09.900innovation. And so we are not going to impose any guardrails on this stuff. And I kept trying to
00:05:16.600explain like, look, you are assuming that the tech industry is, is giving you an accurate sense of
00:05:23.940this technology. And I don't think that that is true. And I kind of reached a breaking point where
00:05:31.280I was like, you know what? I think that the most impact that I could make would be by kind of
00:05:36.500trying to explain why all of the assumptions that people have in the policy world about
00:05:43.420this ai future that they see is misbegotten so that's how i ended up here well i certainly don't
00:05:51.700see how that's related to anything that i've ever worked on or why you and i are talking right now
00:05:57.240i mean that all makes a lot of sense to me i imagine there are a lot of people who don't
00:06:01.360want to hear that but um speaking of things that people don't want to hear why don't we talk about
00:06:07.040this bubble. And then I don't want to give our listeners a sense that this episode is once again
00:06:17.000going to be a complete bummer. We really are going to talk about what we can do about these
00:06:22.120things. But first, we got to be clear on the problem. So let's talk about the bubble.
00:06:27.060You've written a report for Open Markets Institute, where you work, called No Bailout
00:18:50.540Aviation was a new thing at that time.
00:18:52.860And people were pouring money into deeply unprofitable airlines and the, in the end,
00:18:58.180uh, airplane building companies and automotive companies.
00:19:02.100Um, back then those were the high tech plays and they were, they were, they were, they
00:19:09.040technologies that had the potential to, and that eventually did transform the economy.
00:19:14.260The problem was they were not ready to transform the economy yet, number one. And frankly,
00:19:22.240they didn't transform the economy nearly to the extent as people thought they would,
00:19:26.400because subsequent inventions ended up displacing a lot of the technologies that
00:19:31.920were being invested in at that point. So it wasn't though the stock market bubble alone
00:19:40.880in 1929, as big as it was, that led to the Great Depression. What's really eerie is that if you
00:19:50.660read the last chapter of a book, it's one of the best books of economic history ever written. It's
00:19:55.380The Great Crash 1929 by John Kenneth Galbraith. And he said that there were five things in 1929
00:20:02.840that led that stock market crash to kind of spiral into the Great Depression.
00:20:09.000Number one was high levels of income inequality, which basically meant that when all these rich
00:20:16.960people who'd gotten rich off the stock market weren't so rich anymore, that the economy kind
00:20:22.980of just inherently was going to slow down a lot because it was the rich people who were driving
00:20:28.000it in the first place. The second was you had a lot of interconnected banks and financial
00:20:33.360institutions that kind of relied on each other to pay their bills. That's absolutely true today.
00:20:39.320There, he talked about how there is a deep imbalance of trade. Back then it was actually
00:20:44.480a trade surplus, but today it's a trade deficit, but there's kind of a similar effect there.
00:20:49.980And another was that policymakers, I'm forgetting the fifth, but another was that policymakers were doing all the wrong things. And that is absolutely true today, too, because instead of pumping the brakes on all of this, you've got the Trump administration about to open up Americans 401ks to private credit.
00:21:11.320Uh, you've got the federal reserve loosening capital requirements, which basically allows
00:21:17.780banks to take on more risk instead of less.
00:21:21.140Um, and in general, you've just got kind of an eerie coalescence of the same ingredients
00:22:01.100the stock market is worth two and a half times almost us gdp which means that any stock market
00:22:08.460crash today would wipe out more wealth than anybody could have ever thought was imaginable
00:22:14.280in previous crashes so that's why i'm petrified about what's happening with this okay i know you
00:22:23.440said that was it wasn't going to be all doom and gloom but it's pretty gloomy yeah this is not a
00:22:29.140video podcast but again i'm making a face as is so often the case when when you're talking to someone
00:22:35.280and you're recording it uh the the best line came before we turned the recording on you at one point
00:22:42.520said well we were getting ready to start uh let's party like it's 1929 and now i'm understanding why
00:22:50.080you said that. Oh boy. Okay. Well, that all sounds horrifying. It is. And the other actual
00:23:00.740close parallel or somewhat close parallel that I can think of was 1873.
00:23:06.920Ooh, a deep cut. Yeah, very deep cut. But there's a new book by an author whose name I'm sure I
00:23:12.980will butcher. I think it's Liaquat Ahamed. But it's called 1873. And it is about that year's
00:23:20.720stock market crash. And really, it was mainly a bond market crash. But that was really the first
00:23:27.400kind of modern financial crisis that the U.S. experienced. And it was built around the railroad
00:23:33.860bubble that had arisen at that time. The reason that that, I think, is an important parallel
00:23:39.220is that it really, even if you do believe that AI is transformative and that we will have
00:23:47.680a completely revolutionized economy as a result of it in a few years, that's exactly what they
00:23:54.420thought about railroads in the 1860s. And they were right then. Railroads did revolutionize the0.83
00:24:01.480country. The problem was people poured way too much money into too many railroad lines,
00:24:09.980way more railroad lines than we needed. They wildly overestimated how quickly
00:24:16.360railroads would transform the economy. And basically what ended up happening was they
00:24:22.040took on a lot of debt. They weren't able to pay it back fast enough. And it all came crashing down
00:24:27.800when a few of the original Robber Baron schemes started to fall apart. So I think that that is an
00:24:33.060important lesson too, because what it shows is that even if you do believe that AI is transformative
00:24:38.840and revolutionary, that doesn't mean that you can throw as much money as you can get your hands on
00:24:46.580into it and expect everything to turn out okay. Especially given the fact that well-built railroad
00:24:53.220tracks last decades, the GPUs that are at the center of the spending on the AI bubble have a
00:24:59.960shelf life of at most about five to six years. So we are burning a lot of money on infrastructure
00:25:08.200that won't even last as long as the infrastructure that tanked the economy in 1873 today.
00:25:14.340You know, it's funny. This is not even the first time that the railroad bubble has come up on this
00:25:18.220podcast uh i think i was talking with um um shannon velour uh a few episodes back and
00:25:27.500she mentioned the railroad bubble and she also mentioned i think it was her shannon i'm sorry
00:25:33.960if i'm putting somebody else's words in your mouth and i apologize to whoever the other person was
00:25:37.360but i'm pretty sure this was shannon that at one point um she's you know giving a talk and
00:25:43.280somebody said, well, you know, like regulating AI now would be like, you know, regulating
00:25:48.480the railroads when they were transforming the economy. And that would have gotten in the way
00:25:53.280of a truly transformative technology. And we don't want to let that happen. And her response
00:25:57.280was, do you think that we didn't regulate the railroads? We totally did.0.97
00:26:04.100Here's the thing. One of the reasons that we regulated the railroads, we didn't regulate
00:26:08.320the railroads at first. At the very beginning of the railroad boom in the mid-1800s, they didn't.
00:26:15.340And people kept getting hit by railroads. And that's why we have grade crossings. That's why
00:26:20.820there are signal systems that are standardized across the entire system. And that is why railroads
00:26:29.240are also regulated as common carriers and subject to intensive supervision by authorities wherever
00:26:35.320they operate. But a lot of that didn't happen until number one, people started dying. And number
00:26:42.220two, uh, the railroad industry started to lose some of its clout because, uh, their bubble burst
00:26:49.160and people, they, they didn't no longer held all of this credibility of, Hey, if you go against
00:26:55.580the railroads, you are going against the future of the U S economy. So it's very parallel in a
00:27:01.460lot of ways to what's happening now well and so this brings us to you know some real grounds for
00:27:05.820hope right we want to we want to regulate the tech industry well when this bubble pops and it will
00:27:13.500that's going to give us a chance to do it i mean that that's certainly my hope but i obviously
00:27:20.140would prefer if we did it beforehand and did it in a way that could mitigate whatever damage from
00:27:27.040the bubble bursting uh is is is possible at this point and the truth is i don't know though like
00:27:34.140if there's one thing that i've learned from reading about the history of speculative bubbles
00:27:39.340and crashes it's that they have a life of their own they they have the way i describe it is they
00:27:47.260have an illogic of their own um and their rational arguments about hey look at the company's
00:27:56.900earnings relative to how much they're spending on this and look at how well that's always worked
00:28:03.340out in the past whenever that's happened. It just never seems to take hold because everybody always
00:28:08.480thinks this time will be different. This technology is different. This is the technology
00:28:13.400that will bring about change so rapidly and so many riches to so many people
00:28:18.800that the amount of money that we are spending on at this time will pay itself off.
00:28:24.620One of the books that talks about that is actually called This Time is Different. It's the most dangerous, messed up. How is it that we never learn as a society our lessons from these things that is in all of certainly industrial era history?
00:28:43.100Yeah, I mean, it's just that Arrested Development meme. No, it never works for them. They always think it will and it never will. But it just might work for us. This is awful, but I also kind of love this. I don't like that this is happening, but this is such a good articulation of this problem.
00:29:04.620And it's also coming from a different perspective from the place where I talked about it in More Everything Forever, right? Like, in my book, I talked about people saying, oh, you know, this is going to happen. You know, we're going to have endless growth forever. And I talk about all the, like, physical reasons why that's not going to happen.
00:29:24.220But this is also just a wonderful historical perspective on, yeah, this isn't the first time that somebody promised us this, and it never worked before.
00:29:36.620And, you know, what are you going to say?
00:29:37.780You're going to say, you know, well, the past failure to growth forever is no promise of the future lack of infinite returns.
00:29:44.960Like, that's not going to, that logic is not going to fly.
00:29:48.300But of course, these are people who think that there is no relevant history here, that this is a truly unique thing in human history. And again, we could get into like why they believe that and all that stuff. But instead, let's talk about what we can do about it.
00:30:04.980Like, do you see a way for us to break that cycle of it logic and do something to mitigate
00:30:11.720this bubble before it pops or I don't know, pop it sooner rather than later.
00:30:28.960It'll hopefully be out in about six weeks or so that has an in-depth section on this.
00:30:34.000But number one, so much of the revenues that are floating around in the AI ecosystem, you know, your audience can't see me doing this, but revenues should be in scare quotes, scare quotes, because it's circular.
00:30:54.360It is companies, it is NVIDIA and Google and Microsoft investing money in companies that
00:31:02.520are using it to buy services from those companies.
00:35:00.980i am not forecasting when this bubble will pop um although if you have a guess matt i would love to
00:35:09.000hear it but what does that look like what effect does that have on the world what does what effect
00:35:15.180does that have like around the world um listeners of this podcast as far as the timing of it first
00:35:21.780um no i have no idea i think that's the only real answer yeah i mean there's a a famous quote
00:35:28.760attributed to John Maynard. Yeah. I think I know what you're going to say. Yeah.
00:35:32.380Markets will stay irrational longer than you can stay. So that's the one. Um, yeah.
00:35:39.060And that, that is absolutely true here. I do think that there is kind of a
00:35:43.560sell by date that is that, that, that is happening, but that date is not until 2028 or 2029.
00:35:50.580Um, I think, and if this bubble keeps inflating for another two or three years,
00:35:57.320I don't even want to think about what it would look like at that point. Something else that
00:36:02.000a lot of people, um, and this is a transition to what the world would look like afterwards.
00:36:07.840A lot of people seem to have a, an impression that crashes and crises happen almost instantaneously.
00:36:18.480Um, they don't remember, they remember the collapse of Lehman brothers in 2008 as this
00:36:24.100instantaneous thing that sent everybody into a panic. And to a certain degree, that makes sense.
00:36:29.700That's when it really kind of became impossible to ignore. But credit markets froze up a year
00:36:37.400before that, and it took extraordinary interventions by the Federal Reserve to keep the economy from
00:36:43.160collapsing at that point. Bear Stearns, which was like Lehman Brothers and other big investment
00:36:48.420bank. It had collapsed about six months before Lehman Brothers did. So, you know, like it wasn't
00:36:55.120an instantaneous thing. There was actually a one year lead up to, uh, September, 2008 and Lehman
00:37:02.360Brothers collapsed. And even in 1929, um, the stock market crashed in October of 1929. Banks
00:37:10.240did not start failing on Moss until Roosevelt was elected president and was about to be sworn in
00:37:17.440almost four years later. Wow. That's how long it took for the stock market crash of 29 to spiral
00:37:23.640into the Great Depression. And in the dot-com bubble, it burst, but nobody actually knew when
00:37:29.360it burst. It wasn't as though there was this day where everybody's like, oh, the dot-com bubble's
00:37:34.420over now. The stock market reached a peak. It started to fall. It stabilized a bit, and then
00:37:41.080it fell a bit further. But basically what happened was it reached a peak, and nobody knew the day
00:37:46.340that that happened, that that, that the stock market was never going to go higher again for
00:37:49.620another six or eight years, whatever it was. So that's the thing about economic crises and
00:37:56.280bubbles is that the end of them can sneak up on you in some ways. Again, like there's, there's a
00:38:03.040dark side to that. And there's a bright side to that. The bright side is that it means that there
00:38:07.220are opportunities to kind of manage the fallout from a crash before it happens. And again, like
00:38:15.040one of the big things is making sure that you have steps in place where you can allow
00:38:21.300big companies and even banks to fail without bailing them out, which was a problem that we
00:38:27.060kind of started to address in the aftermath of the 2008 financial crisis, but I worry that we
00:38:32.820didn't do nearly enough. But it means that there's still an opportunity to pass legislation that
00:38:37.580makes it easier to wind up big failing companies so that they can collapse without taking down
00:38:46.840large chunks of the economy with it. But the downside of it is that people often hold on
00:38:53.140to the illusion that inflated the bubble or whatever the speculative frenzy was in the first
00:39:01.080place, they can maintain that illusion long after it should have been apparent that that
00:39:09.520narrative had been destroyed. That was certainly the case in 2007 to 2008. Here's a fun fact about
00:39:16.100the 2008 financial crisis. Even as Lehman Brothers was collapsing into the ground,
00:39:23.540they were going around telling the Federal Reserve and telling bankers,
00:39:26.540we're not insolvent. We're still good. People are freaking out because they think that these
00:39:32.580subprime mortgage bonds are never going to be worth anything. They're wrong. They're still
00:39:37.060worth something. You should lend us more money. So the ability of the people who are inside a
00:39:43.980bubble to engage in self-deception and self-delusion is virtually unlimited. And that
00:39:49.080creates a real danger that even after the bubble bursts, more people will put money into it.
00:39:55.860And usually the last suckers are the people who are least aware of what had been happening before.
00:40:01.620And that's why it's so scary that Americans 401ks might be about to get exposed to this stuff because they are the ones that are kind of haven't been maybe paying attention as much to the balance sheets of these companies and could easily get sucked into it even as, you know, the rug gets pulled out from under the industry.
00:40:23.940i i guess fun yeah you have a different definition of fun than i do matt interesting i should have
00:40:30.600said interesting uh yeah i know the 401k stuff is terrifying especially uh with the spacex ipo
00:40:37.760and all that but you and i oh god you know what we're not gonna do that particular one right now
00:40:43.080because you and i have a separate thing we're we're working on about that uh which uh i didn't
00:40:49.800say that and we might yeah long story short if you know everybody who's listening if you have a 401k
00:40:55.960make sure that it's invested in funds that are not exposed to spacex that's that let's let's just say
00:41:01.560that definitely true okay so i'm gonna be really annoying for a second here um or maybe more than
00:41:10.640a second depending on on your opinion of me um you didn't really answer my question you explain
00:41:17.920that we don't know when bubbles end you explain people's capacity for self-deception but you
00:41:22.060didn't actually tell me okay what does it look like if this bubble pops in the next six months
00:41:26.900or or if it pops a couple years from now so i'll give you best case scenario worst case scenario
00:41:34.140uh best case scenario i think that there's still a chance that if it pops tomorrow or today that
00:41:41.700it looks like the the dot-com bubble okay you know when the dot-com bubble burst there was a
00:41:47.200mild recession you know obviously people who lost their jobs during that recession it did not feel
00:41:52.700mild but in the grand sweep of history the bursting of the dot-com bubble was not like 2008 1929 or
00:41:59.8401873 um there was another similar situation actually in the mid-1960s you know where stock
00:42:07.380prices got kind of wildly out of proportion and things kind of settled down. A lot of investors
00:42:16.060lost their shirts, but the economy as a whole kept going. The things that would be necessary
00:42:20.760for that to happen are probably, again, you would need to, it would be nice if a couple of the
00:42:27.520smaller players started collapsing so that people could see the writing on the wall before the big
00:42:32.220fish started to collapse. And that's kind of what happened with the dot-com bubble. You had
00:42:36.980well, you know, global crossing and, you know, is maybe the most famous.com. You're a company
00:42:43.020that just went totally bust based solely on too much speculative debt building up,
00:42:48.320but gold global crossing was not a fortune 10 company. And the fortune 10 companies from the
00:42:54.940dot-com year, like Cisco and Microsoft and those other companies, they're still around.
00:42:59.300they survived their stock prices went down but they you know so the best case scenario would
00:43:05.580be something like that i think that that is highly unlikely because there's so much more
00:43:10.620debt that has built up today but there is a pathway where you know like again if regulators
00:43:18.960were kind of doing their job and you could find ways to ring fence the debt so that private equity
00:43:25.120funds that are exposed to it, they kind of get wiped out. But the commercial banks, I think that
00:43:31.400there's still a path that you can kind of save the commercial banks from being too deeply exposed to
00:43:37.240this. And maybe just barely the biggest AI and tech companies could hang on. So that's kind of
00:43:47.100the best case scenario. Like I said, I do not see with the amount of debt that has built up at that
00:43:53.740point. I do not consider that to be the most likely scenario personally. Well, and also the
00:43:57.740current administration is unlikely to engage in the necessary kind of regulation, but yeah.
00:44:03.340That's exactly right. You know, what I think is the worst case scenario would be
00:44:10.820basically a lot like 1929. That's why I said party like it's 1929.0.56
00:44:15.740um because what happens when there are all these negative feedback loops right now we are
00:44:23.600experiencing in a lot of these markets what are essentially positive feedback loops the ai
00:44:28.580companies like microsoft and alphabet and nvidia they keep releasing these blockbuster earning
00:44:34.620statements those earning statements give people confidence that they can give more money into the
00:44:40.520companies that are part of the AI bubble, which makes their profits go up even more because
00:44:45.900there's even more money that's like flowing through this system. And on top of that,
00:44:51.420people see their 401, not just their 401ks, but especially rich people see their asset,
00:44:57.700their portfolios go up in value. They can spend more money on all of the things that people spend
00:45:03.780money on, whether it's, you know, consumer goods, whether it's investing in new businesses,
00:45:08.700whether it's in hiring more workers in their current businesses, we're experiencing all those
00:45:13.900positive feedback loops right now. The problem is when those go into reverse, and then those
00:45:19.360positive feedback loops become negative feedback loops. That's what happened in the aftermath of
00:45:24.940the stock market crash of 1929. People saw their portfolio values going down. They had to sell
00:45:31.200their stock sometimes in order to just meet margin calls when their stock went down. They had to
00:45:38.340sell stock in order to pay their bills, in order to meet their debts. The people who'd been hired
00:45:43.980during the previous years, they got laid off. When those laid off people are out of work,
00:45:48.420they don't spend as much money. So the economy starts tanking. And that's kind of the slow,
00:45:54.580you know, it didn't happen all at once after 1929. Over the course of the following three years,
00:46:00.180though, it just became apparent that there wasn't enough substance holding up the rest of the
00:46:05.880economy to keep it going when the bottom got pulled out from the stock market. That's what
00:46:12.860I'm worried about would happen if there's an AI crash, that all of these other weaknesses that we
00:46:18.180see in the economy right now, where there's so much wealth that is just tied up in a relatively
00:46:24.800small number of companies and in a relatively small number of centibillionaires. We've never
00:46:30.860had a situation where, again, this much wealth has been tied up in the stock market before,
00:46:38.000and this much debt has been tied to a single industry before. I really do worry that if
00:46:45.840it goes down, you would see something like the slow deterioration that happened after the 1929
00:46:53.580stock market crash, where the banks don't necessarily go under right away, but over time,
00:46:59.800the economy is revealed as being so fundamentally weak underneath it that there's the recovery that
00:47:07.940people kind of keep holding their breath for doesn't materialize. So that I think is the
00:47:13.860worst case scenario. And I do worry that it is getting more likely every day that goes by
00:47:19.540where we see the S&P 500 hitting another new high, but the underlying revenues that are going to
00:47:28.520need be needed in order to ultimately pay the bills that are mounting up for these companies
00:47:32.340are not actually materializing. You know, you're reminding me of something that I saw a lot of
00:47:39.860people say during the wildest heights of the crypto bubble, which was, uh, and, and shortly
00:47:46.700after it, you know, started coming back down, which was that, you know, the crypto bros were
00:47:51.360speedrunning the history of economic and banking fraud and learning about, you know, why the
00:48:01.180regulations that exist exist. What I'm hearing from you now is sort of like, okay, and now the
00:48:07.060AI bros have found a way to do that same sort of speedrun of history repeating itself this time as
00:48:15.460farce, but for the 1929 crash. And everybody has been calling this the second Gilded Age.
00:48:23.740And one of the things that I've drawn hope from is that the Gilded Age was followed by the0.62
00:48:28.720Progressive Era. And you are sort of making a case for, well, maybe the second Gilded Age is
00:48:34.260going to be followed by something more like a second New Deal era, because we might have a
00:48:39.880second great depression so on that cheerful thought uh although you know a second new deal
00:48:46.220sounds pretty good but let's talk about bailouts there is going to be a call for bailouts there
00:48:53.760was a call for bailouts in 2008 as you note in your paper this is not only a bad idea um you know
00:49:00.960for the world but it was also a bad idea politically it was wildly unpopular for the
00:49:05.140government to give handouts to the largest companies and the wealthiest people in the world.
00:49:11.100What should we do differently this time? When this bubble pops, how do we keep there from being
00:49:17.240bailouts for these companies? Because they're, as you note in your report, they're already starting
00:49:23.380to make noises about there being a bailout, you know, a need for a bailout if something like this
00:49:28.320happens, that, you know, there's going to be a government backstop, that if the government
00:49:32.440doesn't provide that backstop, then we'll lose in the race to AGI with China, which is another
00:49:38.700narrative that I want to just stamp on a whole bunch until it's dead, because it makes no0.96
00:49:45.960goddamn sense. And we're going to have to do an episode about that one at some point. But anyway,0.86
00:49:51.180tell me about how do we stop bailouts? Number one, I think that the energy that has arisen
00:50:00.620around the data center protests, it gives me some hope. I had a line about this in my report.
00:50:08.000I think that I said something to the effect of, if policymakers spend gargantuan amounts of money
00:50:13.620bailing out a deeply unpopular industry because it overspent on deeply unpopular infrastructure,
00:50:19.360it would be deeply unpopular. So the first thing is to just appeal to policymakers' sense of
00:50:25.320self-preservation. I think that if there were a bailout of the AI sector, rather than steps to
00:50:33.280make sure that ordinary people retain their economic security in the wake of a crash,
00:50:38.560that if you thought that the Tea Party and the Occupy Wall Street protests were kind of an
00:50:46.120impressive social movement after the 2008 financial crisis, you ain't seen nothing yet.
00:50:51.960So I think that people need to start making noise about that, both appealing to the it's the right thing to do and also appealing to policymakers' sense of self-preservation.
00:51:03.380But lastly, a bailout would be futile.
00:51:06.840There is not enough money that the U.S. government could print to backstop this bubble if it bursts.
00:51:15.640the tarp bailout in 2008 was about 800 i want to say 800 billion dollars i think was the final
00:51:23.260figure adjusted for inflation call it about a trillion even today i've done some kind of back
00:51:30.700of the napkin math even just accounting for the amount of likely toxic data center debt
00:51:38.320that would go bad in the wake of a that would be left behind in the wake of a crash you're talking
00:51:45.480double that, probably. And that's not even counting, again, the effect of just people
00:51:53.060feeling poorer in the wake of the stock market, the biggest, most overvalued stock market in
00:51:58.160history crashing. Just to backstop the debt, the bad debt, would cost probably double as of today,
00:52:07.360again, back in the napkin math, as it did in 2008. I just don't think that there's enough.
00:52:12.100And as you may recall, when the Iran war started and when the Liberation Day tariffs happened, bond investors were not nearly so eager to buy U.S. dollars and U.S. treasuries in the past year when the U.S. started to hit economic uncertainty as they have been in past crises.
00:52:34.820Even in the COVID pandemic, everybody just piled into buying U.S. treasuries because that was considered the safest asset.
00:52:42.100If there's an AI crash, I don't think that the government will be able to print enough
00:52:47.780dollars and issue enough treasuries that people would want to buy in order to stop the spiral
00:56:04.420I'm going to finish with one last question here
00:56:06.840and you've sort of already answered it in a way,
00:56:10.140but what are you hopeful about these days?
00:56:12.840I read a book recently called The Public and Its Problems by John Dewey. You're nodding your head,
00:56:20.24099.5% of people would be saying who? Okay. The Public and Its Problems was written in 1927,
00:56:26.340so almost exactly a century ago. And it was about how is it that we can create a society where our
00:56:36.300government is more responsive to the people in a time of rapid technological and social change?
00:56:41.100like that was the entire thesis of it and there was one line in particular that stood out to me
00:56:47.820i actually like tried to find it but it was something to the effect of we create all these
00:56:53.660machines and we become beholden to them not because the machines have a will of their own
00:57:01.160but because we do not and the point that he was making and i know that that doesn't sound hopeful
00:57:07.620But the point that he was making and the reason that it, that I do have hope is that I think that people figured out that the technologies of that time in the new deal era, they figured out that the technologies of that time and the businesses that controlled them did not have to control us, that we could use them as a means to better our own lives and better our society.
00:57:34.220And for a time, for about 40 or 50 years, we created a society that is imperfect as it was, at least was not dominated by a few large corporations and billionaires in the same way it had been before.
00:57:52.260and that brought decades of relatively constant economic progress and gradually improving social
00:58:01.920conditions for people as well. And I think that John Dewey is somebody that everybody should read
00:58:08.640because he kind of, in my view, he saw that the fight for building a better society, it's not
00:58:16.180just a marathon, it's an Ironman. It's a long sweep of history. But if you're patient and you
00:58:22.200have faith in ordinary people, in their ability to take agency for themselves, that you end up
00:58:31.080with a better society. So that's what I'm taking hope from. Well, I don't think we can find a
00:58:35.480better place to end it than that. So Matt, thank you for making the time. This has been a great
00:58:40.600conversation and we'll have to have you back sometime to talk about Deep Space Nine. Or our