How to Stop “Ensh*ttification” Before It Kills the Internet - ft. Cory Doctorow

There's a kind of caricature neoliberal version of this where you say, actually, no, this is just reveal preferences. People don't just like Facebook. If they did, they'd have left, right? In the same way that like, if you sell your kidney for rent, you ever reveal preference for one kidney. I'm Bethany McLean. Did you ever have a moment of doubt about capitalism and whether greed's a good idea? And I'm Lujas and Gales. We have socialism for the very rich, rugged individualism for the poor. And this is capitalism to a podcast about what is working in capitalism. First of all, tell me, is there some society you know that doesn't run on greed? And most importantly, what isn't? We ought to do better by the people that get left behind. I don't think we should have killed the capital system in the process. Since our show discusses what's working in capitalism and what isn't, today we are going to talk about inshitification. And what? And shittification, a term coined by Cory Docterov to describe the increasingly common customer experience on the internet today. It was named the 2024 word of the year by Australia's McCory dictionary. And it inspired the 2025 season of Black Mirror. What makes this world so special? Isn't this just funny way to say the things are getting worse? Well, the way Dr. O'Rourke uses it, it's far more than a word. It's actually a diagnosis of why we as customers are having a worse and worse experience on the internet. He outlines these steps. First, platforms are good to their users. Then they abuse their users to make things better for their business customers. Next, they abuse those business customers to claw back all the value for themselves, or most of the value. Finally, they have become a giant pile of shit. Even Paul Cardman adopted this term and wrote a little model about it in his sub-stack. OK, OK. But shitta sighed, and I'm sorry for this language, but it's new about it. Have you bought a printer recently? Go on Amazon. And you will find a color HP dash jack for only $59.99. Even better. It includes a three-month ink trial, because when the ink trial appeared, and you discover that the ink cartridge may be made of gold, because it's cost as much. At that point, you feel and I'll let you complete the world. In the language of the economy, it's called customer lock-in. Don't you find this experience extremely frustrating? Yes, of course. What I'm telling you is that that's part of the problem, is that is the customer lock-in. And so, rational customer wants to be paid to voluntarily lock themselves up. They are paying the form of a better up front deal. Once they enjoy the up front deal, they regret having done it. It's like the credit card that you love when you can spend it, but then you are left with a debt and you hate it. Is it also end-shortification? Now I see what Richard Taylor was referring to when we talk to him at the Stigler Center. You economists are still locked in pun intended, the rational framework, normal customers, not your super rational ones, by printers without expecting the high cost of cartridges, and then they feel, I can't use this term on air. Definitely. You're right, but the evidence seems to be on my side. If customer were completely oblivious of a lock-in effect, the printer seller will not need to discount the printer to induce the original sale, but they do. Maybe not enough, not all customers are rational, but some. The problem, and there is a famous academic paper on this, is that the system generates last distortions. I end up printing much less than I would like because I'm overcharged for the cartridge, and so my marginal cost of printing is too high, we respect to what it will be efficient. I think you might be missing the point, or at least Doctor of's point. He talks about an inshidogenic. He has great at making upwards. He talks about an inshidogenic environment, an environment that favors the lock-in through regulation. The environment effectively rewards bad behavior. Do you mean he has rediscovered Stigler's regulatory capture? Of course, you're an economist, an unbearable one, and you have to have gotten there first, but I think we better bring Cory Doctrov into defend himself. He's a very accomplished science fiction novelist, journalist, and technology activist. He's a special advisor to the Electronic Frontier Foundation, a non-profit civil liberties group that defends freedom and technology, law, policy, standards, and treaties, and he's the author of several books, including his new book. So a far less catchy phrase for inshidification would be a race to the bottom, and I'm always curious broadly speaking, why this happens? I mean, if you believe in the beautiful version of capitalism, it's supposed to lead to a race to the top, where companies compete to do better and to make better products in order to do better in the market. And it seems in your book, and in your telling of this, that these were deliberate strategies taken by companies in order to allow the process of inshidification. And if it was deliberate, why is it just easier that way than it is to compete? And what fundamental forces enable the race to the bottom? Yeah, so I actually don't think it was internal forces. I don't think it's endogenous. I think it's exogenous. I think endogenously the pressure on firms is to lower the amount of money you spend on products, lower the amount of money you give to workers and suppliers, and maximize the amount of money retained by the firm. But under in a state of nature, a firm cannot charge infinity dollars and pay zero dollars to its workforce unless it's an academic publisher, in which case that is how it operates. The thing that prevents firms from doing this is the fear of a competitor, but of course we stopped enforcing anti-trust law. We stopped preventing horizontal mergers. But we also didn't want to enforce vertical mergers either. We subscribed to industrial organization theory that said vertical mergers were uniquely safe. The only time the state should intervene in a merger is if we know for sure that the merger is intended to be anti-competitive. And of course, if you don't want to ever prevent a merger, that's a great standard because how will you ever find out what the intention of a tech boss is? Unless of course that tech boss is Mark Zuckerberg, who never had an illegal impulse. He didn't commit to email that was later discovered by a prosecutor. And so Mark Zuckerberg sent an email at like 230 in the morning to his CFO that was like, yeah, we are buying Instagram. The reason we're spending a billion dollars on a company with 12 employees, which you seem to think is crazy, is because people are leaving Facebook and going to Instagram and not coming back. So we'll recapture those users here. This is like sending a memo that says like, Bob, that guy we were talking about killing just so you know, it's definitely gonna be a murder. And I was just thinking about it, I'm probably pre-meditating it right now, right? And yet the Clinton administration waved that merger through, right? Obama, administration, Obama. Obama, Obama, I beg your pardon, you're right. No, the Obama administration. Yes, it would be amazing if the Clinton administration waved that merger through. And so this is like, we had just no competition law enforcement and we got monopolies. And you know, you folks working for an economics podcast might appreciate just how frustrating it is for someone like me to hear economists describe the cartilization of every sector in our economy as being unrelated to the pro monopoly policies of economics, right? It's like, we used to not have a rat problem and we were putting down rat poison. And then like these guys came along and said, you know, you probably don't need the rat poison. You should stop putting it down and now rats are eating our faces off. And they're like, well, why are you blaming me? Right, maybe it's like sunspots that created the time of the rat and made them more fukun than in any time in history. And of course, it is true who the rats did by the rat poison factors and shut them down. But so what we're not using, rat poison, that's like an economically rational outcome. This is like perito optimal. You know, and so we stop enforcing competition and then like we lose another source of discipline, which is regulation. And you know, we think of these as separate domains, but you know, the collective action problem that 100 SMEs have goes away when it's a cartel of five companies. They can decide on what they're gonna tell their regulator. They're not engaged in what Peter Teele calls a wasteful competition. So they have lots of free cash flow that they can devote to policy entrepreneurship. And they can make their policy preferences felt. And so you know, like why are so many of the tech monopolies making so much money from spying on us? Well, maybe it's because we haven't passed a new privacy law since 1988 when Ronald Reagan got so angry about Robert Bork's video rental history being leaked to the newspapers that he signed a law that bans video rental history disclosure without permission. And that's the last privacy violation that Congress has sought seen fit to prohibit. Right? So like regulatory capture, you know, we had other forms of regulation that firms didn't avoid, but rather mobilize like IP law. When they cartelized, they were able to weaponize IP law in order to protect their cartels. And then they had to worry about their workforce because their workforce is very strong because they were valuable. But then when you have a cartel, they can gang up on their workers too and eventually supply meets demand. You see half a million tech layoffs in the last three years. This endless, like, disgustingly horny talk about AI replacing programmers. AI doesn't do software engineering. It just writes little chunks of code. Software engineering is a thing that requires a lot of context, which is the thing that AI cannot do. You cannot get a large language model to think really hard about what happened before and what's going to happen afterwards and what's happening adjacently because that is geometrically more computationally expensive than just writing one routine. And so, you know, it's not that they think they're going to replace programmers with AI. It's that they think they're going to frighten programmers by talking about replacing them with AI. And so the power programmers gets broken too. And then we just end up in a situation where the guys who get up every morning go to the C-suite, grab the giant lever on the wall, mark it in shittification and pull it as hard as they can, no longer have it gummed up by competitors or regulators or interoperability or an empowered workforce. It just moves very freely. It's been totally lubricated by policy choices. And then we get in shittification. Of course, you're right. The incentives are to pay the least and to sell the highest price possible, et cetera. And the big insight about us meet is that competition fixes these problems and delivers desirable outcome. And however, the competition that Adda's myth is talking about is competition on prices or on quality. They often competition takes different forms. And those don't deliver the same good outcome that Adda's myth predicted. So if we are competing in trying to make it more difficult to leave the jail in which we are the customer in or if we are competing trying to make it difficult for you to leave and choose somebody else, that's a terrible competition is very destructive. And this is where I think the antitrust authority should intervene massively because this is clearly incredibly well for destroying when Google is making its product worse in order to make more money. This is like unbelievable, right? Yeah. The problem is that I think economists have not recognized and of course, policy makers, even more so, how devastating this dimension is. Yeah, I think that's well put. And I do think that this is an area where like that Neombrand Dysian sometimes have a blind spot where competition for its own sake gets reified. So I remember there was an op-ed in the sling, which is an excellent publication written by this guy who works for one of the advocacy groups, like add industry groups in the UK who was talking about attribution. So this is a thing that the ad industry is totally obsessed with. It is the idea that after I show you an ad programmatically, I follow you around in the physical world. I look at all of your credit card transactions and look at everything you do on the internet until I figure whether you bought the thing that I showed you the ad for. And this is a thing that Google and Meta are uniquely suited to do. It's one of those things you do get returns to scale from. And smaller firms cannot. And so they can't compete. And so the point of this industry lobbyist who also managed to sneak something into one of the competition market authorities reports the one on ad tech is that this is an unfair competitive advantage and we should address it by making it easier for smaller firms to do attribution. And his answer to this is assigning every Britain a permanent advertising identifier at birth that can be used to track all their actions. And it's the most insane goddamn thing. If you believe in competition as an end instead of as a means, I think of competition as embodying John Sherman's idea that if we would not have a king, we shouldn't have an autocrat of trade. But what we want to do is not allow in bread merged cartel that's screwed so many of each other that they've got a corporate Habsburg jaw to just take over our sector, structure our markets in our lives. And I want competition to turn them into a rabble who are at each other's throats so that we can have effective regulation. And I say this to my libertarian friends all the time. Look, even if you are so Elon Musk adult that your copy of the fountain head can no longer be read because the pages are all stuck together. And the only thing that you think that a regulator should do is enforce contracts. They still have to be powerful enough to enforce contracts. They have to be more powerful than the entities they're enforcing the contracts among. The referee has to be able to boss the players around. Otherwise, they won't play by the rules. And so the smallest government you're ever going to get is going to be determined by the largest corporation you're willing to tolerate. And so even if you're a small government person, I think you should want sectors organized around SMEs and not around large dominant firms. I mean, IBM was in antitrust litigation with the DOJ for 12 consecutive years. And every year it outspent the entire DOJ on outside counsel to fight the DOJ antitrust division. Right? When a corporation can outspend the federal government for 12 consecutive years, it's going to be really hard to stop it from cheating. So another version of this question, and I think your views on this might have changed over time is do insidified firms die or when do they die? Yeah. Is it simply the speed of insidification like the proverbial frog in the boiling water? Reading your work makes me think that you've become more cynical over time. That you once maybe believed that insidified firms did die and now you think they might be vampiric and they might have life forever. Am I right about that? And is there any line or any sort of set up characteristics you can draw to say when they might meet their deserved fate? So I call them zombies and vampires, but otherwise I agree. I think that they're well passed the time we should have double tapped them and stuff them in a shallow grave, but that we struggle too, we're trapped in their rotting carcasses. And I guess when I wrote initially in that first blog post and then they die, that it was more wish casting than anything, they're not bad at finding the minimum quantum of value that will keep us stuck there, at least until they don't, right? At least until something exogenous happens that tips the equilibrium from, I hate this place, but I can't stop using it and I hate this place and I'm not gonna use it anymore. So far they've been able to often recover from that, Facebook has had several of these livestream mass shootings and other things that have provoked bolts for the exits and sometimes it's provoked like weird strategy changes at Facebook, like Mark Zuckerberg announcing that the future wouldn't consist of arguing with your most racist uncle using the primitive text interface that he created in his Harvard dorm to non-consensually rate how bangable fellow undergrad were, but rather that he was gonna convert you and everyone you love into a legless sexless low polygon heavily surveil cartoon character so he can imprison you in a virtual world and then he can actually stole from a 25 year old cyberpunk novel they call the Metaverse, but for the most part these firms have hung on despite these mass missteps and I think that should alarm us, right? I think that there's a kind of caricature or neoliberal version of this where you say, actually no, this is just revealed preferences. People don't dislike Facebook if they did, they'd have left, in the same way that like if you sell your kidney for rent, you ever revealed preference for one kidney, right? But if you acknowledge that people sometimes make hard choices and those hard choices arise out of like policy frameworks, right? Like if you have no tendency protections then people might have to sell their kidneys for rent and if you allow firms to exclude new market entrants, purchase their competitors, engage in a raft of sleazy conduct, spy on their users and so on, then they might be able to lock their users in a long pass the time when they would have left. And the thing that I worry about now is not whether these firms will die because I think they will eventually. It's the harm they'll do on the way and the harm they'll do when they go because the thing that you really want, especially in these social platforms, but also in other platforms like banks and so on, is you want to evacuate them. You don't want them to just collapse, right? Like there are lots of very bad banks. One of the reasons we step in when those banks are failing rather than allowing them to just fail is that the depositors suffer substantial harms if we just allow nature to take its course. And in the same way, people on Facebook are there because they have really important social relationships with each other. It's why they're still there. In fact, it's the great tragedy of Facebook is they have weaponized the things that matter most to us and they are always seeking an equilibrium where the harms from losing everything Facebook gives you are just slightly outweighed by the harms Facebook is choosing to inflict on you to benefit its shareholders. And that's why I think this is a policy matter and not a matter of us shopping incorrectly or even the wickedness of Facebook's management. So Bethany and I had a debate about what exactly knew behind this idea of an insurgification? Because some of the ideas are around an economist, what is special about it? Well, as you say, I think that a lot of this exists in the economics literature. I would say that I do identify sectoral characteristics of tech that I don't think are well explored elsewhere. Those would be the unique composition of the tech labor force in that they are quite powerful despite not being unionized because of the scarcity and because of their productivity. There's an NBR paper that estimated that the average tech worker in Silicon Valley was adding a million dollars a year to the firm's bottom line. So when you have a worker who's very scarce and is worth a million bucks a year, you treat them very well and those workers themselves, often they're path into tech involved having a kind of revelatory life experience with tech where they discovered their community, they discovered their politics, they discovered a way out of what seemed like a predestined life. And so they often want to give that to everyone else so they cared about their users. And then there's this one other characteristic that is distinct to this that I think has been very under theorized in the economics literature, which is the role that interoperability plays specifically as it relates to digital computers, which have a kind of interoperability on steroids and as much as the only computer we know how to make is the touring complete universal von Neumann machine that literally can run every program. And not only is that what all our computers are, but we do not know how to make a computer that only runs some programs, which means that every 10 foot pile of shit in an app you use has an 11 foot ladder that can go straight over it. Historically, the fact that interoperability would allow new market entry, a very low capitalization to alienate the affections of customers, right? You block their party ink. I make the thing that lets you install third party ink and that that can lead to new market entry at low capitalization that just scoops up the business of incumbent firms. You block third party ink in your printer and someone else comes along and offers me the tool that lets me unlock that and put in their generic ink, you know, the cost of materials for ink is a penny a gallon. The retail price of ink is $10,000 a gallon. That's a lot of margin that you can play with, even if you have to spend some of it on like keeping up to date with jail breaking every make and model of every printer as it gets updated. You know, I might just buy my next printer from that ink company. It created a great deal of dynamism and in the arguments about the expansion of black letter IP law and its enforcement and interpretation, arguments about how that would foreclose on market entry were not taken very seriously by policy makers. They just had this idea that stronger IP laws necessarily created more capital formation and that that capital formation would result in innovation. And the fact that it would allow incumbent firms to block interoperability, this engine of dynamism in technology was just completely ignored by economists and by politicians alike. One of the parts I love of your book is the story about Bruce Lehman and how we got Sado to use it with the IP law we have now. Yeah. So Bruce Lehman, he was, I don't know if it's Lehman or Lehman. I've always said Lehman like Lehman brothers, but maybe it was Lehman, maybe it was Lehman brothers. Bruce Lehman was the IP's R for Bill Clinton and Bruce Lehman went to him and said, we're going to need a new IP law for the internet. And that IP law should ban all modification of devices or services unless the manufacturer approves of them even if you're doing it for a purpose that is otherwise lawful. So if you're modifying your car that has been designed so that you can only interpret the check engine light if you're paying the manufacturer for a check engine light interpreting tool, which is a thing that every mechanic today is paying for, $10,000 per manufacturer per year. And if you modify that car so that it just takes any check engine light modification tool, that should be illegal, right? Modification without the consent of the manufacturer should be illegal. And Al Gore thought to his credit, this was a terrible idea, and he told Bruce Lehman to go pound sand. So Bruce Lehman went to the World Intellectual Property Organization, WIPO, Ingenie Vitts, the specialized UN agency that makes copyright laws, they have the relationship to bad copyright law that Mordor has to evil in middle earth, right? Every bad law emanates from them. And they passed two laws, the WIPO copyright treaty and the WIPO performers and phonograms treaty, which are collectively known as the internet treaties. And these contained this provision. And then Bruce Lehman went back to Congress and he said, well, now it's an obligation, right? This is no longer a proposal. It's a requirement in 1998, Bill Clinton signed the Digital Money and Copyright Act, which embodies these proposals that had been rejected by the Administrative Branch. And Bruce Lehman subsequently gave many speeches where he described this as, quote, doing an end run around Congress. So that's how we got these first anti-circumvention laws, and you can see why the US would want them, right? Because they were going to both have firms that domestically produced technologies that could extract rent by having these anti-circumvention tools and therefore return more money to their shareholders, but also they would export these around the world. And these tools would become a system for America to extract rent from its trading partners. But it's harder to see why those trading partners wanted it, right? Why sign up to this? And the answer is the US Trade Representative said, if you don't do it, we're going to hit you with tariffs. And so everywhere you look, basically every country that has advanced technical capabilities has signed up to make it illegal to modify things. And there are some European firms that have taken advantage of this to extract their own rent. It's not good, though, right? Like one of the reasons diesel gate happened is because anti-circumvention creates a zone of legal jeopardy for security auditing. Because if you open up the software in like a car engine to find out how it works, and there's like a digital lock that you have to break to get there, well, then that's a crime. And so no one wanted to audit a Mercedes or a Volkswagen engine to see whether they were gaming their diesel emissions outcomes. This killed thousands of Europeans a year. Today, Mercedes uses it. They sell you your accelerator pedal. So you buy a top of the line Mercedes, but it doesn't go as fast as the engine can, unless you buy a monthly subscription. There's another European firm that uses this. It's a company called MedTronic. They're the largest MedTech company in the world. They're nominally Irish. They're really American. They did the largest tax inversion in history, but not before buying every other MedTech company. So they make all the ventilators. And they use the thing. They're actually minnesotan in fact, so they're close to Canadian, which might actually be a lot better anyway. That might be that. Yeah, yeah, yeah. This is the definition of minnesota nice. So they broke the ventilators so that if you put a part in a ventilator, a good part, the ventilator won't recognize it until you pay for a service call out and an unlocks step. And it's against a lot of bypass that. And during the lockdowns, hospitals were nursing their ventilators along and doing things like taking a donor screen out of a dead ventilator and putting it in a working ventilator with a dead screen. But it wouldn't work because metronic uses these software locks that are against a lot of bypass. My favorite example, there's a Polish train manufacturer called newVog and EWAG. And they sabotage their locomotives. So the locomotives are continuously checking to see where they're parked. And if they find that they've been parked in a service yard belonging to a rival train servicing company, the locomotive immobilizes itself and cannot be started again. And you call up newVog and you say, my locomotive is broken. And they say, well, let us do some remote diagnostics. And then they say, oh, we found your problem. That will be 20,000 euros to reset it. And it turns out that it was just a software bomb in the train. And the security researchers who were hired by a Polish train operator to find out why their trains kept doing this, when they published this result and made this public, newVog threatened them with liability under section Article 6 of the Copyright Directive because they had to bypass the software locks to find the sabotage in the locomotive. So getting rid of this in Europe should be a no-brainer, because it's not like any Europeans are happy about the firms that in Europe, the handful of firms that are using it. They're all using it in terrible ways. There are no good ways to use a prohibition on the owners of devices deciding how they work. That's just not a legitimate thing. So you said something super interesting that I never heard before, which types were threatened unless you agree on the IPL. So Trump was not the first one to threaten types for political reasons. Oh, this is a new thing. The Trump is doing. I mean, I think that the biggest impediment to a sovereign, safe, European, digital sphere has been the fear of having tariffs imposed on them. And when I was European director of VFF, I worked in more than 30 countries. Everywhere I went, everyone I spoke to said, your policy prescriptions sound very good. And we have no doubt that they would be beneficial to our economy and to our people's thriving. But they wouldn't be beneficial enough to offset the costs of having tariffs imposed on us by the US. I used to teach, sometimes in the summer, at CU and Budapest, in an international relations program. And a lot of my doctoral candidates were former government officials from around the world. And one of my students was someone who had been the information minister for a Central American country during the Central American free trade agreement negotiations with the US. And they described getting a call from their negotiator saying, I know you told us not to cave on anti-circumvention with the trade rep, but they will not take our agricultural exports if we do not sign up to anti-circumvention. And so now they're already not able to send their agricultural exports. And they have an excellent technical university in the capital. And if they were to pick just one line of business where the US is extracting tons of rent, like automotive diagnostics, right? And made a tool that was 25 bucks a month that every mechanic in the world could buy that would be costless to export. The United Economics are incredible, right? It's just a piece of software you're exporting. Your ongoing cap ex is just like, you know, you buy every car as it comes off the line, you install every software update when it's pushed out and then you jailbreak it. You could make billions, like billions of dollars in a software export market. And you know, being a Canadian, I remember what it meant when we had one company that was successful, right? We had research and motion. And it spun out an entire regional tech sector. You know, look at what happened to Finland when it's stopping the sick man of Europe and it's leading rubber boot manufacturer Nokia started making phones instead, right? These are gigantic opportunities for countries around the world. And, you know, Americans that I'm at present time living in America, I have American citizenship now. Americans won't be left out by this. You know, in the same way that we can buy reasonably priced pharmaceuticals from Canada, we could buy those digital tools. So long as we had a payment method and an internet connection, we could be unshackling our phones and our cars and our printers, our games consoles, our ventilators, our tractors from control by American monopolists because, you know, when you look at how much stock everyone who's not in the top desial owns, right? And therefore, a number of Americans who benefit from the stock gains, the superior cash position of firms that do this rent extraction, it's like, it's nothing, right? Almost all of us would benefit from cutting off these rent seeking arrangements. So you are an activist who is very attentive to the so-called theory of change. So this phase of instantification actually creates some opportunities for change because at the beginning, when most customers were getting all this product for free and were novel and were happy, it was very hard to sell them on, oh, we should limit Google, we should limit Facebook. But now that they start to see the real nature of this big tech giants, I think there is a real opportunity for a popular movement of antitrust. Do you agree with this? I think there is a real moment for antitrust. I think it's already happening. I actually think this is one of the great under theorized areas of political science in this moment because you've probably seen the studies that look at empirical data on policy outcomes and compare them to attitudes among different groups in the population and conclude bluntly that things that billionaires want happen and things that billionaires don't want don't happen. And the exception to this is antitrust and it's not just happening here and not just happening under Biden. It started under Trump one. It's happened obviously in the European Union but also in the competition markets authority in the UK and not under like a radical labor. It wasn't Jeremy Corbin who kicked off their antitrust enforcement. It was a string of Tory governments each more shambolic than the last, including the one that did not last ahead of lettuce. And yet they did more work in five years of the competition markets authority than they had under in the previous 50. Canada had the weakest competition regulator in the world. Our competition bureau had only attempted to stop three mergers and had never succeeded in stopping a merger. And then in 2024, Justin Trudeau whipped his caucus to give them the most far reaching enforcement powers of any competition regulator in the world and he gave them a budget to use those new powers. We've also seen big antitrust swings in Australia, Singapore, Japan, South Korea but also in China, right? Where the cyberspace directive prohibits Chinese tech giants from blocking interoperability by new market entrants in China because contrao everything Nick Clegg said, China doesn't use its tech companies as a way to project power around the world. China views those companies as a competitor for power at home and internationally and wants to chop those firms down to size. So this is quite remarkable as a political science matter. It is as though like gravity was repealed, water started flowing upwards and the skies filled with flying political science pigs and nobody noticed, right? Like we're just like, oh, that's quite a curiosity. Oh, it's Neil Brandisians. Somehow Tim Wu has changed everyone else's mind. Tim and I went to elementary school together as I've known him all my life. He is very persuasive, but he's not that persuasive. You know, I don't think anyone understands it. I don't claim to understand it. It is a remarkable new fact on the ground. I think that you don't even have to understand why it's happening to understand that when there is a tailwind this strong, it's time to unfurl your sales. It's funny. The first question I meant to ask you at the beginning of the podcast was about Tim. And if there was something in the water, where the two of you grew up, that shaped the way you think. And also, if you're thinking does divert on any points because when reading his book in your book and certification, the age of extraction, they're similar themes in some ways. Are there places when you guys sit down together over a drink of brown liquor or do you fight? Are there places you disagree? Or do you find yourself pretty sympathetic? I mean, I would say broadly, Tim views market economies as an end to themselves or at least as something optimal or optimizable. And I think of markets, I don't know if you know the economist, John T. Harvey, but he talks about markets as a tool. And he says, you know, be very weird if you met a carpenter who said, no, I only use screwdrivers. There's no way I'm going to use a hammer, right? That's just immoral. What are you, a communist? I think of markets as a tool. I think the case for markets comes and goes based on historic contingency, but I don't think they're the most important tool we have. And you know, I, I, I don't know, there's a, there's a very good book called The People's Republic of Walmart that revisits the socialist calculation debate in Vienna in the 20s where you had, you know, Hayek and, and funmizes saying, well, like you could just never build a computer big enough to, you know, build a command economy, the size of the Soviet Union. And the point of the authors is that Walmart is an economy the size of the Soviet Union and it's a command economy. And so at the very least, like we can say that, like as a technical matter, you can run a command economy that is the right order of magnitude to be a powerful nation state, whether or not you want to is a separate matter, but like part of the argument against planning in industrial policy has been that it's just a, a technical nonsense. And yet, you know, you have Walmart. And, and not just Walmart, right? But Walmart is such an important buyer that they tightly integrate with the supply chains of their key suppliers. So like when Walmart orders a bunch of toothpaste from Colgate, palm olive, they don't send them a PO. They schedule their production line, right? They reach into their production scheduling. This is like what they tried to do with the Soviet shipyards and steel mills was to actually have like a joint management so that when the shipyard needed more steel, they didn't place an order, they coordinated production. And that is in fact what Walmart does. They don't use price signals, right? To turn on more Colgate, they reach into their scheduling app and they summon their workers to the Colgate factory when they need toothpaste. So I'd say that that's our biggest difference that like I'm at least willing to entertain maybe that, maybe this is a difference between being a science fiction writer and an academic, is that I'm willing to entertain the possibility of a fascinating world where markets are less important and planning is more central. Although I also, I'm quite interested in how we can pluralize management of these things. I like the idea of municipal fiber, but I don't like the idea of Eric Adam being my ISP. There are lots of ways that you can hybridize that. And I think this is maybe where Tim and I agree where like you should have municipal fiber in municipal conduit and anyone should be able to rent space in the municipal conduit to run their own fiber. And then the municipal data centers should have essential facility sharing so anyone can rent a shelf in the data center. If they don't want to run their own fiber, they do want to be an ISP where they can offer ISP services from the shelf in the municipal data center. And that way you have a kind of best of all worlds where it's not just Eric Adams is your ISP or Victor Orban is your ISP, but rather you have the stuff that markets can't do, which is find an efficient way to clear the costs of digging up all the streets in your city and putting conduit in the ground, being handled by the state, and then a public option that acts as a check on the private sector and then private options, including cooperative or nonprofit options that act as a check on the state as well. So you're a very prolific writer and you write a basically a book a year. And you're coming up next to books a year, so. Yeah. So next year, one of the two books that you're coming up with is about AI. Can you give us a sneak preview to view about AI? Yeah, it's called the Reverse Centaur's Guide to Life after AI. And the title refers to two, I think, important phenomena, one of which has to do with automation and labor economics. So an automation theory, a centaur is someone who's assisted by a machine and a reverse centaur is someone who is conscripted to assist the machine, always at a pace that exceeds normal human tolerances and always over time scales that exceed normal human endurance. And so machines don't just use humans, they use them up. So this is the story of an Amazon driver, Amazon warehouse worker, a therapist who's been conscripted to be the back end of 25 simultaneous conversations with a chatbot that's doing AI therapy and has to be monitoring all of them and waiting for any of them to throw a flag and say, oh, this person has gotten into water that the chatbot can't handle. The reason the title says life after AI is that the economics of AI do not pencil out and can't be made to pencil out. That the economics of AI are grounded in the idea that you can display flavor. That you can just substitute algorithms for wages and that if you add up all the wage bills that AI can zero out, it makes up for the $700 billion in CAPEX that the firms have already committed to and the hundreds of billions, or that they've already expended and the hundreds of billions that they've already, that they've committed to besides, it's a number that Bane estimates as a $3 trillion a year by 2030. I don't think we get there. I think when you look at the jobs AI can do, they're not high-waged jobs, their jobs where it doesn't matter. People make mistakes like AI porn and election disinformation, the wage bills for AI porn that AI porn and election disinformation disrupts are like round to zero. The big demos that you can do with AI, like commercial illustration, which I also don't think is very good, but even if you like it, commercial illustrators are the most precaratized workers in America. If you zero out all their wages and give them to Stamm Altman, it won't pay for the kombucha bill for the technicians overseeing a single training run of the next model. It's effectively zero. Meanwhile, all the things that we want AI to do, like teaching or driving or what have you, they only work in these reverse centaur configurations where you fire most of the people to do it, you put chapots in charge of it, the chapots make mistakes at a pace that the human can't possibly catch that is being put in charge of babysitting the AI, but the human gets the blame, right? The human is the accountability sink. It's a sector that's spent $700 billion so far that claims to be making $60 billion a year, sectorally. That number is grossly inflated, right? It includes open AI getting $10 billion worth of tokens for Microsoft for their cloud and then giving them back to Microsoft, which Microsoft books is revenue, right? That's like you, you know, walking past a Starbucks and they hand you a coupon for a $7 latte and you redeem it and they book $7 in revenue, only it's $10 billion out of $60. They are depreciating their assets on a five year time scale. The GPUs are burning out in two to three years. The data centers are super specialized and optimized, it's not even clear whether they can accommodate the next generation of GPUs, much less general purpose computers that can do other kinds of compute work. And they're passing around the same $100 billion IOU over and over again and pretending that it's new revenue and seven of those companies are 32% of the S&P 500. And so I think we are headed for a crash. I think that crash is gonna be so big that 2008 will look like a utopia by comparison. And I think when it's over in addition to whatever great political wreckage that creates because those crashes are often attended by rises in authoritarian movements, I think we will have some productive residue that we can salvage. There'll be a lot of GPUs, a 10 cents on the dollar. There'll be a lot of unemployed applied mathematicians and statisticians and there'll be a lot of open source models that are gonna outlive the companies that flip them over the wall as a kind of loss leader and calling card. And they'll have barely been optimized because the big AI companies, the way that they signal to the market that they're serious is by increasing their capex. They're basically making an investor proposition where they exploit this Byzantine premium where no one understands what they do. And so all they can understand is like the magnitude of what they're doing and they're hoping that investors go, well, Opala, shit, this big, has gotta have a pony under it somewhere. And so there's just this opportunity to optimize these models that they have thus far had no interest in optimizing and getting them to run a commodity hardware on premises to do interesting and useful things that are not going to result in $3 trillion worth of wage displacement and certainly aren't gonna make like a super intelligence, teaching the word guessing program more words will not make it intelligent any more than breeding horses to run faster and faster will cause one of them to give birth to a locomotive. They're just not the same thing. So as I was the last question for you, as I was thinking about insinification and it's not quite the same process but I couldn't help thinking of Isaac Osomov and he in the foundation, he writes, why they don't even understand their own colossal any longer. The machines work from generation to generation automatically and the caretakers are a hereditary cast who would be helpless of a single detube and all that structure burnt out. So are you an Osomov fan? And is that the inevitable and result of insinification? I do like Asimov. My first short story was published in Asimov Science Fiction Magazine. So I'm a fan of his. But I don't think that's what's going on. So one of the things I'm at great pains to do in this book is recover the, what I think of as kind of a suppressed history of policy choices that led to this place. I don't think it is mysterious. I don't think it's that the machines are black boxes that we don't understand. I think we said let's not enforce anti-trust law. We got monopolies. I think we didn't pass privacy law. We got privacy invasions. I think tech workers had an opportunity to unionize when they were strong and they said, no, I'm a temporarily embarrassed entrepreneur. Why would I need a union? And now they're like working for bosses who don't fear them anymore and whose treatment for the workers they don't fear involves apple factories with suicide nets around them. And Amazon vans with pea bottles in them. And there's nothing about working in, you know, with a computer that says that you can have a pea bottle next to it, you know, for a guy with his kid and with his penis shape rocket. Jeff Bezos really does not like our kidneys. Like these are all choices. And, you know, Bruce Layman's choice to create anti-circumvention law. The US trade representative's choice to spread it around the world. They're not inevitable. They didn't come down off amount of no-to-stone tablets. We can choose otherwise, right? We aren't prisoners of a black box. We're not, we're not trying to figure out how to like embalm a pharaoh or build a pyramid without power tools. Right? We're just like saying, hey, that thing that we stop doing, we could do it again. And that thing that we do do, we could stop doing. And for that reason, I'm actually very hopeful. If you're enjoying the discussions we're having on this show, then there's another University of Chicago podcast network show you should check out. It's called the Chicago Booth Review Podcast. What's the best way to deliver a negative feedback? How can you use AI to improve your business strategy? And why is achieving a soft landing so very hard? The Chicago Booth Review Podcast addresses the big questions in business, policy, and markets with insights from the world's leading academic researchers. It's groundbreaking research delivered in a clear and straightforward way. Find the Chicago Booth Review Podcast wherever you get your podcasts. It was fun, no? It was really fun. Where do you agree with him about regulation? And where do you disagree? Do you think he puts too much weight on the failure of regulation in the failure of anti-trust law or does all of that 100% resonate with you? So there is a failure jointly of regulation and anti-trust law. These platforms do bring new problems. I think new problems require some new thinking or at least some clear thinking. And instead, we try to apply a framework that maybe it was not even working before, but clearly doesn't work very well in the new world. I love his emphasis on interoperability. I think that that is, to me, the best solution because he's a technical solution and he's a solution that favors freedom and competition. And I think there are limited ability to induce regulatory capture. Remember, we had the discussion with Tim Hu on this issue and he's very skeptical about regulation. And I share some of that skepticism except that if you're really trying to enforce a common standard and interoperability, I think that I see as the cost being much less and the benefits being enormous. What Kori said is very interesting is that we is a general sympathy to our anti-trust. He only, in my view, overestimated the impact that this had in the United States because while we have seen Jordan Cantor and Lena Can being quite aggressive on the exact side of the branch, we saw no progress whatsoever on the legislative side of the equation. This might really be linked to what Tim Hu told us is at the end of the day, the presidential election is the least corrupt by money ironically because it's the most visible one. And so it's the one where probably you can get somebody like Biden making a decision to appoint people like Lena Can and Jordan Cantor. While in Congress, all Congress people and all senators are really dependent on money and they don't wanna alienate the Silicon Valley industry. I'm gonna say perhaps the least cynical thing I've ever said, but I do wonder if there's another component to it, other than money, in the sense that this movement toward understanding the failures of anti-trust in the fact that it is enabling capitalists to not behave like capitalists actually, any more or what Tim Hu said is so dramatic. It obviously did happen slowly, but the moment in which we realized it happened, happened all of a sudden, maybe some of our legislatures just actually haven't caught up. They're still trapped in a world in which American competition, enabling American greatness, enabling American companies to do whatever they wanted to because those companies inevitably we're gonna operate for the good of society and for the world. And I think there might be a more fundamental human nature thing going on than just money. I love this sort of known cynicism of yours, but I will reply with Afton Sinclair. It's very hard to convince people when their money depends on that not being convinced. So I think that is to say, you are right that there is an ideology or a leftover an ideology that makes it more comfortable to sort of defend a status quo. However, when they have the right incentives, they learn super fast. And when they don't have the right incentives, they never learn. I thought his ideas in contrast or in conjunction with WUS were also really interesting. What did you think their areas of disagreement were? Or what, or maybe what did you think their areas of particular resonance were? My impression is that they, on the substance, they disagree less than they claim the disagree. I think that they come from two very different perspectives. My impression is that, and I'm kind of a speak out of term because I don't know all their intellectual history, but my impression of the intellectual history is that Cory is a leftist activist that has always been kind of more on the anarchic side. And as come to this dislike for Big Tech, because of his fundamental saying libertarian now sounds very political, but you see what I mean. Law for freedom and independence. And Tim Wu, I think, has more of a traditional, if you want neoliberal formation. He was in the Obama administration, basically at the time, where the Google deal that Cory was criticizing went through. And he did not raise any issue with that. And only later it's changed. But I actually give a lot of credit to Tim to have changed. And maybe their starting points are very different, but their ending point these days are not very different. Remember, basically, very strong endorsement that Tim Wu gave of Zora Mamdani. They end up in a pretty similar place. You know, that's really interesting, because I had thought, until you mentioned Tim's endorsement of Mamdani, I had thought that they actually were coming from pretty radically different places. In the sense that I see Tim as more of a believer in capitalism. He wants to save it from the capitalist, per your language, but he's a believer at his core in capitalism. At least in something clear to its ideal state, whereas I think Cory very clearly is much more of a socialist-slash communist and doesn't even necessarily believe in the power of markets. Or he says markets can be a tool, but just a tool and not even. And so I think in that way, they are very, very different. But maybe I'm conflating some of Tim's prior views with his positions now, because I had forgotten until you mentioned it about his endorsement of Mamdani. And maybe his mind is fake that I don't see sort of a call you as such a communist. This is what? Really? Oh, that's interesting. Didn't you think his example about how, which I wanted to get to, because I'm not sure he was right, his example about how you could do run a command economy, because Walmart runs a command economy. So criticisms of a communist style of running an economy aren't right, because Walmart can run a command economy. I thought that very much betrayed, not betrayed, just showed up a position. I disagree. I always say that basically a command economy is a firm. The firm, but a cause will agree with that. The example of a command economy is a firm. The fact that you can run large firms that suggest you can run large economy. The big difference is that you have competition among those firms. You don't have competition when you have one command economy in your own country. I think that's the big difference. I did not take, and maybe the way I read, I did not take this as an endorsement of a Soviet economy is more that the standard criticism that people make is superficial, because there is a good counter-example. I did take it as more of an endorsement, so, but I might have to go back and listen to what he actually said, and see, we can decide. Our listeners can decide who's right. I guess I heard it in the context of his overall skepticism about markets, and he did say they could be a tool, but definitely not the only tool. But I will subscribe to that view. I will subscribe to that view. Markets are a very important tool. You don't want the market everywhere. No, who does? I think that's been some of our powerful episodes, and have shaped my thinking on that, that you don't want the market everywhere, for sure. It should stay in its cracks fear. So, okay, all right. All right, we should have asked him that question. Is he outright, are you a communist? Would you prefer that way of... I'm not that of running an economy. But do you think I actually was interested in that? Do you think broadly that that comparison is right, that because you can run a firm like Walmart as in a command way, you can run an economy that way? I'm not sure the two are comparable, not just from the standpoint of competition, but I'm not sure that analogy holds. And I guess one reason I would think that it doesn't hold is that the number of inputs required to run an economy in a command way are exponentially more than the number of inputs required to run a firm in a command way. So, I wasn't sure I was thinking as he was saying it, I was trying to decide if I agreed with that analogy or not. Do you agree with it? I actually think that, of course, depends on what country and which firm, but between the country of Bhutan and Boeing, I think Boeing is infinitely more complex than the country of Bhutan. The supply chain of Boeing is an nightmare. And the level of sophistication, I think this is actually one of the favorite questions I asked my students. I give them as an assignment to read the famous piece by IAC, the use of knowledge in society. And then I asked them, I said, look, higher growth in 1945, but of course there were no computers, or at least no computers at the level we have today. And so, to what extent the existence of computers changes is reasoning. And to now, I said, what extent the surveillance economy we have and also the AI might change the answer. And I think actually, the cost of giving away the answer, but at this point, the judge, the GPT does the same, is the computers don't, does not really change the equation very much because IAC was very careful in saying, it's not just a computational issue, okay? If it's only a computational issue, that can be resolved by computers. In his view is a transfer of knowledge issue because the knowledge is diffuse, there is an incentive problem. But actually, if you really read, and I'm sorry to say, I spent more time and I cared for a minute on that piece by IAC, but if you really read carefully, he basically said that the knowledge is not transferable. So almost by construction cannot be transferred. And so his idea is you need to delegate the actions to the individuals because the individuals are the only one with the information because that information is by definition not transferable. And so I do you coordinate them, you need prices to coordinate, okay? So it's not, most people say, IAC say the price, actually transfer information, no, price only coordinated the action of people who have a lot of information at the local level and they cannot transfer that information. That is the advantage of a market economy according to Hayek, which is to some extent, not solved by the Walmart problem because at the end of the day, the Walmart employee that has to make this, the Walnut employee can now make decision based on his local knowledge or is not fully because he's forced by this command and control. Interesting. Sorry, I went overboard with my left ring. I felt that totally fascinating. See, I can learn all the economics I never knew from you. Capitalism is a podcast from the University of Chicago Podcast Network and the Steegler Center in collaboration with the Chicago Booth Review. The show is produced by me, Matt Hodeb and Leah C's Reign with production assistants for Uthgondi, Matt Lucky, Sebastian Berka, Andy Shee, and Brooke Fox. Don't forget to subscribe and leave a review wherever you get your podcasts. And if you'd like to take our conversation further, also check out promarket.org, a publication of the Steegler Center and subscribe to our newsletter. Sign up at Chicago Booth.edu slash Steegler to discover exciting new content, events, and events.