Carl Benedikt Frey, "How Progress Ends: Technology, Innovation, and the Fate of Nations" (Princeton UP, 2025)

Welcome to the New Books Network.

I'm Caleb Zakrin, editor of the New Books Network. Today I'm speaking with Carl Benedikt Frey about his recently published book, How Progress Ends: Technology, Innovation, and the Fate of Nations. Carl is the Dieter Schwarz Associate Professor of AI and Work at the Oxford Internet Institute at Oxford University. How Progress Ends is a truly epic and riveting work. The book takes readers on a tour of history's largest economies to untangle a puzzle that has eluded any simple, static formulation. What's more effective for economic growth, a powerful centralized planned economy or decentralized free market? Depending on your ideological persuasion, you may find yourself instinctually edging towards one formulation over another. You might think of these 2 poles as geared towards different economic goals. The planned economy is geared towards static efficiency, boosting output of predefined products. Meanwhile, the decentralized economy is geared towards innovation and the serendipitous improvements that come from allowing new entrants to outcompete entrenched incumbents. After the collapse of the Soviet Union, the question of centralization versus decentralization seemed settled. However, the rapid growth of China demonstrates that the top-down approach remains a viable option for nations, especially autocratic ones seeking to develop. As we head further into the computer age, with breakneck investment in AI. Which economic system will prove most effective for delivering economic progress? To help us answer this weighty question, I'm pleased today to have Carl Benedikt Frey on the Princeton University Press Ideas Podcast. Carl, thanks for joining me today on the New Books Network.

It's good to be with you.

This is just such a fun book to read. I felt in many ways like it came at the right time because I kept reading books about America's economy, China's economy, various economic questions about how, you know, what's the best way to spur development and progress. And I feel like this book really is like the best overview that I've read on the topic that's recently come out. So I'm glad you wrote it because I feel like you just do this incredible tour through different bureaucracies, different markets, and really try to, you know, untangle the question like what systems work for which goals and what will that mean for the future of progress? But before we even jump into the book, I was wondering if you just tell us a little bit about yourself and your background.

Sure. Well, first of all, thank you very much for those very kind and generous remarks. Now I guess I have to try to live up to the expectations, so pressure's on. So I'm an economist by background. I read actually Joe McCure's book in secondary school. That was what got me interested in technology and economic history, big questions about why some places are poor, others rich to begin with. And so I was particularly thrilled to see Göran Nobel this year. And I think my book tries to speak some, build a little bit and draws upon much of his work as well. And so, yeah, so having read that, I decided to pursue a degree in economics. And after that, I I actually got my first job in academia at Oxford Martyrs' School, which is a multidisciplinary institute. And at the time, there was a lot of anecdotes going around about technology taking jobs and AI and robots doing things that were previously inconceivable. And so I teamed up with a computer scientist to try to shed some light on that question. And we wrote a paper which was essentially an assessment of which type of jobs that are most exposed to developments in technology. I think it's fair to say that that paper was somewhat misunderstood because we have a section where we sort of describe that, you know, what's technologically feasible is not the same thing as, you know, actual adoption. And so my subsequent work has been dealing a lot with, you know, what are the sort of political and regulatory and cultural constraints on progress and growth? My most recent book, as you kindly mentioned, is called How Progress Ends, and it makes clear that we shouldn't take technology and innovation for granted. If there was something inevitable about progress, it wouldn't have taken us 200,000 years to have an industrial revolution. If progress was inevitable, every country around the world would be rich and prosperous today. If progress was inevitable, once a former technology leader, Britain would not have seen 2 decades of productivity stagnation. The book is sort of pushing back a little bit against that.

Right. And you do a tour through many different economies. You look at free market economies, you look at planned economies. Could you give a little insight into what your research was like for this book, what you were reading, who you were talking to, and how your ideas developed over the course of the project?

Well, that's a very good question. So research is a messy process. And so, you know, it's hard to give a neat summary of how it works, but In the early stages, as I've outlined in the book, progress is about exploration. And so writing a book is no different in that regard. You're out there, you're talking to people, you have lunch, coffees, try to talk to some leading scholars in the various fields that you're engaged in. And then at some point you sort of try to crystallize out what it is you want to do, and then you need to sit down and actually write it down and execute. Then there's a lot of revision happening. during that stage as well. And so the book has been 5, 6 years in the making. I've been working on various other projects on the side as well. So it's always been jumping a little bit back and forth. A key theme of the book, which then pulls together various strands of research is trying to explain why some societies grow very rapidly only to stagnate and in some cases to collapse. And so if you take one bucket of explanations which tend to center around geography, geography is fixed and so it could not explain why the Soviet Union grew very rapidly in the '50s and '60s only to collapse. in the '90s. And so the geography was the same. Another strand of research centers around culture, but you would probably be hard-pressed to argue that the culture of the Soviet Union changed markedly. And if it did, if anything became more open because nothing could prevent really the radio and television waves from Western Europe from traveling right through the Iron Curtain. And so Soviet citizens became more exposed to Western cultures and ideas. And then obviously Gorbachev introduced a number of reforms, which if anything made scientific debate and discourse more open. And the other bucket that economists and scholars of development tend to Point 2, being institutions. Well, the Soviet Union did certainly not have anything like secure private property rights throughout. If anything though, the economy became more decentralized during the perestroika years. Firms were granted a little bit more autonomy and still it collapsed. And so The point I'm trying to make in the book is that all these buckets are basically incomplete for understanding these patterns. So we need to look at how institutions and technology interact at different stages of the technology lifecycle.

The book spends a lot of time looking at the 19th and 20th century, various 19th and 20th century economies, how they developed. But you also begin by looking at China and the development of one of the first modern state bureaucracies. And then you also look at the Roman Empire as well and the consequences of the collapse. I was wondering if you could put these 2 sort of systems into conversation and their consequences, looking at why and how the first Chinese major state bureaucracy that we would sort of think of as a modern state bureaucracy developed. What was the purpose of it? And then we can go and we'll talk a little about the Roman Empire and its collapse and the kind of impact it had on a different alternative set of systems that developed in Europe. So could you just begin by talking a little about Chinese state bureaucracy?

Sure. So Chinese state bureaucracy didn't emerge from out of nowhere, right? And so building bureaucracy is not an easy thing. Building a strong state is not an easy thing. And one thing that you do need is to be able to raise tax revenue. And to be able to do that, you need to have some idea of what people are producing in order to tax them accordingly. And so technologies like soil mapping, where China took an early lead relative to Europe, was essential for achieving that. A unified writing system was essential to communicate across this vast territory. An empire. The state in turn used its capacity to build infrastructure, to build elaborate irrigation systems, which in turn spurred technological progress by integrating markets, which made it easier for ideas to travel and people to travel. And so at the height of the Song Dynasty, China has a fairly integrated market system and a strong center, a strong state. The downside to this system though was it was not very dynamic. And so had Galileo been born in China, he would probably become a bureaucrat, rather than a scientist. And so the Chinese state bureaucracy drew in a lot of talent who were competing through the civil service examination system for jobs in the bureaucracy. In addition to that, if you had ideas that were not supported by the political authorities or if they went against the wishes of the ruling classes, you could get into trouble. And that was certainly particularly true later on during the literary Inquisition, where people like Wang, who developed a new dictionary but didn't pay, give enough credit to relatives of the emperor who developed a predecessor dictionary. Well, both he And his family suffered the consequences from that. And that's quite different from Europe where there was no political elite that could really clamp down on public intellectuals. And even skilled craftsmen could move around as Protestantism was made illegal in France, for example, the Huguenots took their skills with them to Germany, to the Netherlands, to Britain. And so skills and ideas proliferated because of European fragmentation, as Joel McCure has pointed out. But at the same time, you had sufficient degree of cultural unity for the communication of ideas to continue across Europe. So you had this combination of political fragmentation with a sort of market space for ideas. And this all depended on the Roman Empire collapsing, which it did in large part because it didn't have the technologies to build a sophisticated state bureaucracy the way that China had. And so rather than integrating the communities it conquered into centralized bureaucracy, it largely left local structures intact. And so at some point as the Roman Empire expanded, something would have to give. It would either fragment or it would centralize. Unfortunately for Europe, it fragmented.

This type of fragmentation where there's, you know, ends up being, you know, many different, I guess you could say, sort of sovereigns and different rules in different areas, it could breed different cultures. And I think part of what's interesting in the book is there are many different economic systems that develop in Europe in the 19th century, we see different paths taken. There's obviously the Industrial Revolution. It begins in Britain and then it goes, it also develops in the United States. But you also look at how various centralized states, autocratic states like the Prussian bureaucracy, they were still able to grow and develop. And I was wondering if you could just Explain a little bit about how the Prussian bureaucracy developed and how they were able to expedite the rise of big business and the consequences of this.

So a key distinction, which I borrow from Michael Mann in the book, is between despotic power and infrastructural power, which I mostly refer to as bureaucratic state capacity in the book. And the latter means that you still essentially have an institutionalized state which is able to carry out tasks, whether it's building railroads, developing an educational system, and carrying out the wishes of the ruling elites, or you can have a more Despotic power, which essentially just depends on society as being divided and even weaker relative to the ruler. And the thing about Prussia is that it was more institutionalized if you compare it to, let's say, Tsarist Russia. And so you had a bureaucratic legacy going back. Several centuries that enabled Prussia as the Industrial Revolution took off in Britain to play a more significant role in industrial catch-up. And it did so by investing heavily early on in schooling. It did so by creating or supporting new institutions like technical universities. It did so by supporting the rollout of railroad infrastructure. It did so also indirectly in many ways because civil servants often went on and took on jobs at some of the leading industrial enterprises in Germany, like Siemens, for example. And so it aided the rise of the integrated industrial firm in numerous ways through just providing pools of skilled managers as well, not to mention— efforts by people like Friedrich Althoff to develop numerous educational institutions from what we call the Max Planck Institutes today, which are still a pinnacle of scientific research in Germany, but also efforts to set up business schools that educated this new managerial elite.

Yeah, it often seems when people today talk about the roots of modern academia, they look at Germany, they look at 19th century Germany and that, you know, many of the developments of the university system there as being roots, which I find quite interesting. And what I also find interesting too about this sort of comparisons that you're making between these different 19th centuries economy, I mean, there's vast difference and I wonder when you're describing the American state in the 19th century, it's obviously post-Civil War, the American state, it still has, obviously they fought maybe one of the first modern wars, but they have relatively weak state capacity. And how did this weak state capacity actually give rise to just unbelievable explosive economic growth? to America? And then we'll get to the response to it because obviously there is a response because the American state does develop eventually. But what was the sort of 19th century sweet spot that allowed for such incredible growth in America?

So what the founding fathers did early on was creating an institutional framework that essentially banned internal tariffs that created a large internal market and laid the foundations for a patent system which would become the envy of the world. And so if you look at the United States in the 19th century, and if you look even at the growth during the post-war period, which is the longest streak of over 2% productivity growth that the United States has ever seen. The foundations for that growth were laid in the 19th century, mostly by independent inventors. So the railroad companies did not invent the railroads. The telegraphic companies did not invent the telegraph. Large firms like DuPont and AT&T Their R&D departments basically didn't do internal R&D. They were screening the market for patents, and it was a very liquid market for trade in patents during this period of time. And so much of that laid the foundation for much of the exploration that led to the Second Industrial Revolution. What the American state didn't have was what you call state capacity. So the civil service was very slim. It wasn't meritocratic, at least not until the Pendleton Act. And the solution to this problem, because America was left with a vast continent but no, very little infrastructure, to connect it. And so early railroads were built entirely through private initiatives, but, you know, that led to a fragmented network which did by no means link up the entire country. And so the Railroad Act of 1861 was very important to the end of linking up the country. And through public-private partnerships by which the state induced the private sector and created a framework where it granted railroad companies land to build on, it overcame its state capacity deficit. So to speak. But it also, along the way, created very powerful industrial enterprises that were by many measures larger than the American bureaucratic state itself. And you begin to see the consequences of this in a range of corruption scandals that surfaced during the Gilded Age. That eventually led to a public backlash, which eventually led to the Pendleton Act, which introduced a meritocratic civil service. And then the various acts from the Sherman Antitrust Act to acts that essentially created the FCC to begin to regulate these large-scale industrial enterprises to seek to force competition, make sure that they couldn't block entry. And that framework is basically still intact today.

Yeah. What's also interesting too, I feel like about the growth of American bureaucracy is like in certain instances there was, you know, quite an attempt to keep it decentralized. You know, if you like look at the, the You know, the Federal Reserve banks, they're scattered around, or obviously the court system, it's decentralized as well. You know, it does seem that even in America's quest at centralization, there still was this decentralized aspect in its DNA, which I find interesting just in terms of thinking about ultimately, like, I feel like where our conversation is going to get to is kind of thinking about America versus China, the debate that a lot of people are having about which system You know, which system will win out eventually? It's kind of the new Cold War in a way. You turn then to the 20th century and you look at innovation in the Great Wars. And I oftentimes think that people think of the Great Wars as times of accelerating innovation. Like, obviously there was accelerating innovation in defensive capacity or in killing capacity. You know, World War I begins with cavalry and it ends And World War II ends with a nuclear bomb being dropped. But you say this isn't the case. It was actually stagnating, a period of stagnating innovation. So could you talk a little bit about World War II in particular and this notion of whether or not it was a period of innovation?

Well, as you said, it's easy to cite these examples. And you said during the war we advanced radar technology and we developed the bomb. And so you can give all of these examples of technologies that came out of the But obviously what you don't see directly is the developments that weren't made because resources were reallocated towards the war effort. And so one thing is seen, the other thing is unseen. And as a result of that, you know, we tend to focus on the things that are seen. But if you look at the data, I mean, patenting basically plummets during the war. I mean, there's a very sharp decline. Yeah. Decline in patenting. And you can make the case, well, some patents are more important than others, but if you look at various measures of breakthrough patents, you see the same pattern. If you look at productivity, Alexander Field has documented this in some detail. World War II was a productivity disaster. The best thing, the best case you can make is that maybe it created some institutions that aided post-war growth, like the productivity program, for example, and some efforts to standardize various activities which may have helped mass production. And then obviously the GI Bill drew in a lot of talent and gave them education and skills. And people like Henry Kissinger benefited from that. But by any stretch of the imagination, World War II was not the productivity miracle. World War II was not an innovation miracle. It reduced innovation by any measure. And so yeah, it's easy to focus on the scene, but if you actually look at the data, you'll get a sense for the unseen too.

Right. And that argument that these wars were accelerating innovation is oftentimes a point in favor of centralized economies because the United States in World War II, like the Soviet Union, was essentially becoming a planned economy at that point. I don't know what percentage of GDP the military was responsible for, but I think it was probably a massive amount in the United States during war directing all sorts of production and, you know, commandeering factories of all sorts. After the war, you know, there's a, you know, indisputably America is in this fantastic position. It hasn't been ravaged like Europe or Japan. But we see, you know, soon afterwards, especially with, you know, with the help of the United States, that both Japan and Western Europe are able to, you know, catch up pretty extraordinarily, able to develop and What were some of the sources of Japan and Western Europe's development, especially in the context of how you, of this kind of question of centralized planning versus a more free market approach?

Yeah. So I think, you know, it's fair to say that, you know, the Rad Lab advanced radar technology, the Manhattan Project, you know, delivered the bomb later on, the Apollo Project put the man on the moon. And those are all noteworthy achievements. But during the war, time is scarce, you're under pressure, and so you can't set aside a decade to explore some completely new trajectory. You tend to pursue the low-hanging fruit, and most of the scientific advances needed for the bomb were essentially at hand. Same for radar technology. And so coming out of the war, United States is a more centralized economy than it's previously been. It was certainly not the command economy during the war either, and the best that the Roosevelt administration could do, what to issue various directives to incentivize businesses to work for the government. But it is, you know, there is certainly this view that America was a more managed economy and that delivered sort of the postwar miracle. But, you know, I think if you again go back and look at the data, so the productivity upsurge begins in the 1920s, continues through the '30s, even the Great Depression, then collapses during the Second World War, and then rebounds during the post-war period. So the foundations for that post-war growth was already there before. And if you look at some of the examples of what supposedly came out of this initiative, I think the most transformative of them is clearly ARPANET, which was the predecessor for the internet. But that doesn't show in the productivity statistics until around 1995. So that had certainly nothing to do with the post-war miracle. And it was certainly no plan at any point in time here to develop the modern internet, right? And so one of the virtues of ARPA was that that it was highly decentralized. And so you had program managers like Licklider who came in and could realize their ideas, and they had a decentralized network of universities that they could use to draw, to find talent. And so that was essentially, you know, the framework that actually delivered some innovation In this period. Now, most people tend to agree that growth in Europe and Japan during this period is not due to innovation, right? It's through catch-up. And unlike previous times, the key difference here is that, well, the United States was concerned about the specter of communism spreading its wings over Europe, and it tried to do what it could to rebuild Europe as defense against that threat. And so Marshall aid contained a significant amount of technology transfer and also transfer of management know-how. to Europe. And it is basically the introduction of the American system of mass production to one industry after another in Europe that gives rise to the post-war miracle. Same thing in Japan, same thing a bit later on in South Korea as well. And so Europe, which has already once caught up to Britain in the 19th century, has a lot of the institutional ingredients in place to do this. And so the technology is American. The institutions that permit catch-up, they are basically European. But the returns to the mass production system then eventually peters out in the 1970s. Then something new is needed for growth. That new thing is the computer revolution, which is from then on basically going to be an American affair.

Right. And even while this is happening in the post-war period, you know, the Soviet Union, obviously there's, you know, there's Stalinism is ongoing, you know, one of the most brutal centralized approaches to government ever devised. Though simultaneously, you know, you have American economists observing the Soviet Union and saying the Soviet Union is still growing. It's very possible, you know, they thought that it was very possible, very likely that the Soviet Union would eventually overtake the United States. This was not an era of triumphalism for free marketers. This was an era still of concern and fear that the centralized model might work better. Can you dig into that a little bit?

Yeah, and it's not just the Soviet Union, right? So coordinated capitalism, if you like, is the theme in Europe. And so in France you have the Monet Plan, which is basically trying to create this concerted economy and coordinates between industrial partners and coordinates investment across industries. In Italy and Spain, you have state holding companies, huge industrial enterprises that coordinate investment. And obviously the Soviet Union is the most extreme example of this. And so if you do catch up, well, what, what you need to do is basically take technologies invented elsewhere and introduce them, uh, in your domestic economy. And the Soviet Union was pretty good at doing this in some priority sectors and What you then need is to be able to monitor performance. And when something is standardized, when you have benchmarks, it's much easier to hold factory managers accountable. And so this was something that the Soviet authorities could do respectively well. But when the technology is new, What you benchmark against. And so as the computer revolution begins to take off, it actually makes planning harder, not easier, paradoxically. And needless to say, Soviet contributions to the computer revolution were essentially none, right? If you were, let's say, an aircraft engineer in the Soviet Union, and you wanted funding for your, uh, technology, you could go to the Red Army and ask for funding. If they decline, well, maybe you had 2 or 3 other options. If they declined, your idea would die with you, right? That's quite different from the US system where, uh, Bessemer Venture famously declined to invest, uh, in Google back in, uh, 1999. But that did not prevent the rise of Google because others took the opportunity to invest. But it also illustrates that, you know, Google's success was by no means a foregone conclusion at the time, right? So AltaVista and Yahoo were dominating search. And so to know if something will actually catch on, somebody needs to take the risk and actually invest in it. And so in a decentralized system, you can explore more technological trajectories. And so, you know, to different degree, what the Soviet Union, what Western Europe, and what Japan as well had in common is that either large industrial conglomerates, state-owned companies, or, you know, uh, the authorities, uh, themselves created very significant barriers to entry for new firms, for new players to enter and explore. They were essentially geared towards industrial catch-up. And if you compare, you know, the Soviet Union to Western Europe, there's no doubt Western Europe did better in that regard as well. So Soviet Union managed to do some spectacular firsts like Sputnik. That was basically in the military domain where the Soviet system was exposed to very significant competition. But what's maybe even more remarkable than Sputnik itself is that no telecommunications industry worthwhile mentioning emerges in the Soviet Union subsequently to that, whereas in both Europe and the United States it does. And so Even in cases where it had some spectacular, spectacular firsts, it didn't really spill over on the domestic economy and didn't really improve the lives of Soviet citizens, sadly.

Right. There's obviously this problem of commercial applications. I think as you, you have a line where, you know, there is no, there is no significant commercial innovation that has come from a you know, a top-down planned economy because it's just not geared towards thinking in the sense of commercial viability. So they might launch a satellite, but then they don't think about— then there's no one there to think, oh, well, if we have this capability to launch satellites, then what can we develop from that? So I find this tension, you know, extremely interesting as you explore it through the book. And one of the stories too is Is essentially about you know China faring better than Russia, than the Soviet Union. China eventually you know coming to to be the alternative that actually is more competitive with the United States. Obviously you know there's Chimerica. There's this this kind of fusion of economic interests that China and the United States begin to develop with the opening of China. So I was wondering if you could just introduce a little bit China's. contemporary economic development, that story, why China was essentially able to replace Soviet Union as the United States' main competitor?

Yeah. So a key difference between China and the Soviet Union is that China was already in the post-war period much more decentralized. So in the Soviet Union, basically everything from steel to railroads were managed from Moscow through vertical line ministries that managed every industry on the national level. In China, provinces have much more autonomy, and what happens with the grand opening from 1978 is that gradually, uh, local authorities are permitted to experiment more. And so most top-down industrial policy early on still centers on Beijing, whereas in the south you have experiments with special economic zones that open up to, uh, trade. And what you see is that firms then which were set up in Beijing actually begin to move down south to integrate into global value chains. And Deng Xiaoping's southern tour in '92 then made abundantly clear that this was working out. Extraordinarily well. But what the Communist Party retains a monopoly on is promotion. And so what you get is essentially local provinces and local leaders competing for positions inside the Communist Party based on meeting growth targets. And so the incentives this creates are not that different actually from a market economy. And so local authorities are in some ways acting more like firms than they are acting as political entities. At the same time though, this way of monitoring performance and rewarding growth becomes harder and harder as you approach the technological frontier, because once you start to need— start to explore, It can take years or even decades until it becomes clear whether investing in a particular technology was a success or not. And so what we have seen, for example, is that as China has begun to introduce targets for innovation like patents, you see a lot of low-quality patents being filed. And they are of such quality that authorities can't even be bothered to pay maintenance fees for them, which tells you that they don't really are related to valuable technologies. In addition to that, what we've seen over the past decade is China becoming more centralized again. Pursuing industrial policy more on the national level, more heavily, and shifting priorities more towards common prosperity, technological self-reliance, and national security concerns rather than growth. And in China, the drivers of productivity and innovation have been private firms. And foreign invested companies. And in particular, what we're seeing now is that because these firms are very hard to instrumentalize for political objectives for which they have little to gain, uh, state-owned enterprises are again becoming more prominent in the Chinese industrial landscape, which has again been a drag on productivity. And so for all the talk about the Chinese miracle, since the Great Recession, productivity in China has been flat or even negative by some measures. And China is already— in China, growth is already petering out. at much lower levels of income than you saw in Japan or in Taiwan or in South Korea for that matter.

Right. And that's one of the things that you, maybe a conclusion of sorts, is that while there is this notion that both the United States and China, especially with the coming, you know, the AI age that we're already in, but the coming prospect of, you know, greater productivity that might come from AI, you know, you argue that they're both stagnating or they're both at risk of stagnating. So I was wondering if you could bring in that sort of AI story and how you see both their prospects, because obviously there are other factors that determine economic growth besides just new technology. You know, there's also aging populations, demographic issues, all sorts of other problems. So could you kind of bring us up to the present and talk about both the prospects for both economies with the AI age and how you see it maybe altering, you know, the more decentralized American economy versus the centralized, though with some decentralized characteristics of the Chinese economy?

So I would say that the key feature of American progress has been its relative dynamism. So if you take, you know, the 5 largest firms by market cap in the United States, they're around 39 years old. They are all in tech. If you do the same exercise for Germany, they're 100 years plus. They're mostly firms of the Second Industrial Revolution. If you do the same thing for France, they're mostly in luxury goods. And, but, you know, there's nothing inevitable about that dynamism continuing, right? And so we've seen a significant slowdown in dynamism, whether it's, you know, reshuffling of market share or exit and entry of firms. And we've seen productivity growth slow in tandem. And this is linked to, you know, some anti-competitive practices. We've seen killer acquisitions, for example, whereby incumbents buy up promising startups and promising new technologies just to shut them down. We're starting to see more of a revolving door between the US Patent and Trademark Office and incumbents, where we see patent examiners grant incumbents low-quality patents and then go on and work for these companies. And that makes it harder again for new firms to enter the market. And we've seen a doubling in spending on lobbying since the 1990s by incumbent firms as well. And all of this is before the tariffs and all of these tariff exemptions and golden shares in key strategic companies. And so, Back in the early 2000s, people thought that China would become more like the US. Now it looks like the US is becoming more like China. And so the point is not that growth is inevitably going to stall in the US or in China for that matter, but it cannot be taken for granted. And so artificial intelligence, I think it's the best bet that we have to restart growth. But, you know, realizing the promise of that technology will require competition and entry and exploration and new ideas. And so if the path forward in AI was just scaling, throwing more compute and more data at the problem, then consolidation might Makes sense. But what we're seeing is that these scaling laws are beginning to peter out. We don't know exactly what the future of AI is going to be. Is it large language models? Is it small language models? Reinforcement learning? Is it symbolic AI? Is it something else? Is it some recombination of some of these different approaches? We don't know. And so we need different firms to take different bets and we need more exploration to that end. And so something that we do see during the early days of the computer revolution is that basically every firm from Apple and Microsoft to Google and Amazon, they IPO, they enlist in the stock market and become firms in its own right. Since then, we've seen WhatsApp, Instagram, YouTube, uh, exiting by acquisition. And I think we need to go back more to the IPO trajectory Also for the sake of harnessing artificial intelligence.

One of the things I like about your book's title is that it's How Progress Ends. And I feel like recently, especially books that talk about AI or talk about economic growth and development, you know, they oftentimes paint this like overly optimistic, you know, picture where then they, it feels like a lot of the book is just like this kind of speculation on all of the great things that AI is going to do and how it's going to, you know, deliver us to a, you know, post-scarcity, fully automated luxury AI communism society. And I'm wondering if there's any way that you could just direct listeners when we're reading the news, when we're trying to think about these questions of economic growth and development, both in the United States, both in China, but also in other countries as well, is there anything that you would caution Readers, caution listeners. You know, here's how how you should think about these various stories about economic growth because I oftentimes find myself when I'm reading stories about China that it just is making me question all of my assumptions about what is required for a good for for a good economic prospect. So you know when you're reading about China or reading about you know these various various people that are arguing that America. You know, needs to centralize. America needs to follow China's model if we want to be able to scale production. You know, what is it for you that, that, that makes you give pause to this idea and, and think that actually, you know, we need to double down on decentralization, we need to double down on, you know, on, on various aspects that have led to innovation in the past?

So I think the first thing to say is that just extrapolating growth rates from current trend doesn't make much sense. And so, If you want to understand growth spurts and collapse, you need to understand how technology and institutions interact. So that's point number 1. Point number 2, we hear often a lot of stories about AI did this, AI did that, robots can now do this or that. And so there are a lot of demos, But we don't really care about demos and we don't really care about, you know, some human benchmark. We care about them doing things in the real world. And so the key benchmark is productivity. And so have a think about, you know, what's the productivity contribution to this technology and how, you know, what share of the economy is it actually having an impact on, right? If it's just a really small section of the economy, Then the effect on the aggregate and the larger scheme of things is not going to be that big. And then finally consider, you know, if AI does this, you know, what are some of the institutional bottlenecks that still exist, right? If AI supercharges medical discovery that many people expect, well, still we have clinical trials and that's tremendously expensive and there's a lot of regulations around that. And that means that essentially anybody needs to partner with a large pharmaceutical company to go through that process. And so unless we streamline that, it's still going to be a major bottleneck.

This book, I feel like, is such a fantastic overview of economic history, economic progress, really in the past 200 years. But even more than that, I feel like it's just a very fantastic overview. And it was really wonderful to get the chance to speak with you about this book, How Progress Ends. There is a lot that we were not able to touch on. And even the things that we were able to touch on, you know, only, only really very briefly. So I really recommend that people go and pick up the book and, and read it. I think like no matter what your area of study is or what your area of interest is, there's going to be a chapter in here that you will find interesting. And I think, you know, I'm just, you know, really looking forward to seeing, you know, what you write next. And, you know, I'm grateful that you wrote this book. So thank you so much for being a guest on, on the New Books Network, on the Princeton University Press Ideas Podcast. Carl, it was really wonderful to speak with you. Thank you.