Capitalisn’t: Can Liberalism Survive This Kind of Capitalism?

Episode Transcript

Daron Acemoglu: The most powerful way of going towards human complementarity would be if the tremendous amount of talented people who go into tech, rather than waking up every morning and thinking, "I'm going to build artificial simulated intelligence today," they thought, "I'm going to build more human complementary AI." Then most of that battle would be won already.

Bethany McLean: For most of the last century, the deal in a liberal democracy was simple. The economy grows, everyone's boat rises, work hard, you get a fair shot. Somewhere along the way, a lot of people stop believing in that deal. They think the system is rigged and that no one in charge is listening.

Luigi Zingales: I'm afraid they might not be entirely wrong. What if liberalism itself, the system that gave us more freedom and prosperity than any in history, actually broke a promise?

Bethany: That's a provocative way to put it. It's basically our guest's argument. He's not some fringe provocateur. This is Daron Acemoglu, professor at MIT, one of the three economists who won the 2024 Nobel Prize.

Luigi: As many of you know, this podcast is part of the Stigler Center at the University of Chicago, which is embarking on new project investigating whether capitalism can be made popular again. We spent many episodes diagnosing what's broken. Now we want solutions.

Bethany: We may find them, or some of them, in Acemoglu's new book, which is entitled What Happened to Liberal Democracy?. The argument, in a nutshell, liberalism worked brilliantly as the underdog, fighting kings and tyrants, but once it became the establishment, it forgot its own promise of shared prosperity, and a class of college-educated elites took it over and started using it to remake everyone else's cultures and values.

Luigi: It's a familiar story. It's a story that Wright has been telling for years. The coastal elite looks down on you. It's generally surprising to hear it from a coastal elite economist from MIT who, I'm going to guess, has never voted Republican.

Bethany: That tension is why I think this podcast is worth people's time. Acemoglu isn't saying burn it all down. He's actually trying to save liberalism from itself, and he's trying to save capitalism. He has a name for the replacement, working-class liberalism.

Luigi: To the benefit of our listeners, I would like you to start with what you mean by traditional liberalist, and tell us when you first realized that this form of liberalist was in a crisis.

Daron: I'm not really sure [chuckles] what all of these terms mean, even though I write about them. [laughs] They're very confusing. What I am associating with liberalism are the forces that have really shaped a lot of institutional, social, and economic "progress", always progress in quotation marks, of course, around the world. Also, liberalism hit some roadblocks. If you look at the United States, where some of these things are most clear, the kinds of bargains that liberalism had offered don't seem to work as well.

In the United States, we've had pretty much consistent and steady increase in inequality in every aspect of inequality from around 1980 to about the mid-2010s. That's really striking, but at the same time, also much greater levels of dissatisfaction with the political process, people thinking they don't have voice, they really don't have meaningful self-government, and complaining a lot about public services that they cared about, both in their communities and at the national level.

More at the perception level, I think that a lot of people started arguing, sometimes subject to propaganda from talk show hosts or political leaders or cultural entrepreneurs, that liberalism wasn't working and liberal values weren't working. I think that's also related that there was some weakness in terms of the tangible things that liberalism was promising and delivered for a while, but then failed to deliver over the last four decades or so.

Luigi: One thing that was a surprise is that in the book, you seem overly critical of part of the liberal elite for what we call in jargon, you don't use this term, but wokies. You blame wokies for basically the political backlash for the left abandoning the working class and so on, so forth. One could say you're blaming your colleagues in humanities, but economists have made no mistake. What mistakes did economists do?

Daron: [laughs] I certainly did not mean that. Why did the Roman Empire collapse? Of course, the barbarians or the so-called barbarians invaded Roman Empire and triggered the collapse, but it was the internal problems of the Roman Empire that enabled that. That's my view of liberalism as well, both in terms of not adapting to the post-industrial age, but also the social changes that happened.

What you called wokism, what I call social engineering, is part of a bigger whole, and it is the excessive empowerment of people like us, so economists, humanities, engineers included, the educated segments of population, which became both socially more powerful and influential, politically more assertive, and developed their own culture or subculture and values, and stopped empathizing and working enough for the common good while their political power was in the ascendancy.

Now, there are parts of it where economists were more active, which is, for example, encouraging a view of the market and technology where if you left everything to the power of the tech companies or the financial giants, everything would work out well.

Bethany: Well, thank goodness you weren't letting economists off the hook.

Daron: [laughs] You wouldn't [inaudible 00:06:12], Bethany, right?

Bethany: To continue that analogy, if liberalism is the internal problems of the Roman Empire, then what or who are the barbarians?

Daron: Oh, the talk show hosts that started taking the mistakes that liberalism and liberal elites were making and blowing them out of proportion.

Luigi: At the cost of overplaying your analogy there, which I love, I think I'm more interested not in the barbarians, but in the power struggle, because you call them power struggle. Now, stretching my memories of Roman history, the beginning of this power struggle was an economic struggle between a plebeian lower class that used to be compensated with land whenever there was a new conquest-

Daron: 100%.

Luigi: -and the patricians that were trying to keep this plebeian on the side. Not to play too much the economists here, but economic forces were really much at play here. My interpretation of the wokies is that this was the only thing that a left completely captured by money was allowed to do it, because to be very concrete, if I am Kamala Harris and I'm running for office and I campaign on defending Lina Khan, I get all my funding cut off, but if I campaign on being more woke, the people that have money are very happy to give me money.

It's not that these stupid intellectuals went into this rabbit hole by choice. They were pushed into this rabbit hole by people who are smarter than them and richer than them.

Daron: I'm not completely disagreeing with that narrative, but I think the way you put it has too much of a conspiracy theory feel to it. The main place where I start from is the transformation of the industrial economy into the post-industrial economy. It was that process that empowered the educated groups. It was that process that started breaking down trade unions and other community organizations that were so important for working class voice.

It was that process that increased the social influence and numbers of college-educated post-graduates and other groups that then also became more concentrated in urban areas, and therefore, their culture became more insular. You're absolutely right, there are some themes where economics and culture met, but the way you put it, I think, sounds to me a little more conspiratorial than I would.

I think if you look at, for example, support for progressive causes in terms of donations, even during Donald Trump's last election, that money was there and was very powerful. Some of that was articulating progressive causes. It's not like if you look at the New Democratic Party, especially after the-- [crosstalk]

Luigi: Sorry, economic or social progressive?

Daron: Economic. There were-- [crosstalk]

Luigi: Can you give me an example?

Daron: Yes, I was just going to do that. Even if you look at the Democratic Party after the ascendancy of Atari Democrats and the Watergate Babies that turned their back against antitrust ideas and trade unions, they were very committed to redistribution. Democrats always wanted higher taxes and higher transfers, which isn't what Koch brothers would ever want.

It's just that Democrats wanted economic tools that were more in line with a particular ideology, and here economics plays a role, and were more in line with their economic interests, which if you're a very highly paid engineer, you really don't want labor regulations, et cetera, slowing down your company, but redistribution in some general form, you may have the values to support it. In my mind, it's a more complex story, but as I've emphasized, the root of it in terms of economic changes, the rise of post-industrial technology and dynamics, was critical.

Luigi: One could have a different interpretation of recent history and say, look,-

Daron: I'm sure you will have one, Luigi.

Luigi: -the reason why, at some point, liberalism gave prosperity to everybody is because there was a pressure from an alternative that was communism, and a lot of the benefits that were shared with the majority were under a threat of a Soviet revolution. When that threat disappeared, liberalism shows its true original intent, which was it doesn't care, and went back to what it was before the Soviet revolution was around.

Daron: I think that's a very serious thesis that needs to be taken seriously. However, if you look at the pathways for shared prosperity, meaning the gains from economic growth that became widely shared, most comprehensively throughout the Western world in post-war decades, is that wages grew. Wage growth had something to do with trade unions and with democratic governance, and both of those were indirectly impacted by geopolitical competitions and defensive moves against communist threats or communist ideas, but they were most important in laying the foundations for a labor market in which the main engine of shared prosperity was wage growth.

To a first approximation, and I'm generalizing, and, of course, in such a context, over-generalizing, but I don't know of any example in history of sustained shared prosperity which works through anything other than sustained wage growth. You increase employment, you increase wage. Everything else is background factors enabling that. Wage growth, as most economists would instinctively think, is not something you can that easily engineer.

Of course, national bargains, et cetera, play a role, but at the end of the day, firms must be willing to pay those wages and must be willing to hire workers at those wages for those kind of wage growth to have a meaningful impact. That's exactly what happened in the United States starting in early decades of the 20th century, so much so that many companies, as I talk about a little bit in the book, were sending their agents all around the 50 states to find people that could work, and most of them, for example, relocate to places like Detroit because there was such a hunger for labor.

It was that hunger for labor that was at the root of shared prosperity. It's the disappearance of that hunger due to technology, organizational changes, and other economic changes, again, bolstered by the social environment and the political environment, but it was those economic changes that brought down shared prosperity.

Luigi: You mentioned, and there are beautiful passages in the book discussing this moment in which companies are desperately looking for workers. Why? In part, it's because they had a superior technology that was not, at the time, easily transferable, but in the last 40 years with globalization, we've seen the companies actually go where workers are to the detriment of American workers. Globalization and massive increase in immigration have made the American workers less scarce, and as a result, they've heavily penalized their wage growth.

You only emphasize technology, but political choices like globalization and immigrations are very important. Are you in favor of reducing globalization and immigration to promote American wage growth?

Daron: First of all, in your account, one should also add, and if the new investments are of the sort that are not labor-intensive, that don't require labor, same thing. That way of saying it includes the Trinity, technologies that create labor demand, globalization's effects, and immigration. Now, my own research suggests that globalization has had, especially trade with China, has had a negative effect on job creation, but on the whole, about less than half of the effects of automation technologies in the period since 1980.

Both of them are important, but technology is more important, and I emphasize technology even more for two reasons. First of all, its societal effects, I think, have been larger, and we can talk about those, and second, it was technology that enabled much of that globalization and offshoring. Now, immigration, I spend a lot of time on immigration, especially on cultural aspects of it, but the evidence, at least, that is most dominant within economics doesn't find that the direct impact of immigration has been as large.

Some of that may need to be, perhaps, reevaluated, but it's a fairly uniform set of findings across periods, so that's why I put much more discussion on immigration's cultural effects than on its economic effects, but obviously, I think there's much more to be learned and discussed there.

Bethany: Aren't we, in all of this, ignoring, really, another actor, which is the market? As I understand it, the share of profits that has been going to capital versus labor has grown extraordinarily, and we're now at an all-time peak. No matter what talk we have of reforming institutions to put labor at the center, give labor primacy, if the market is rewarding corporations for reducing their labor costs as much as possible and keeping their profits, doesn't everything else fall by the wayside?

Then does the question become, how do we reform or change the market to reward something different than it does today, and is that ever even doable?

Daron: I do absolutely believe, I mentioned in the book, that monopoly is a problem. However, my assessment is that monopoly has worked mostly in the non-manufacturing sector, which is very important, so it's not unimportant, and it has worked mostly, but not mainly, via its impact on technology choices and globalization, so it's a bundle. For instance, the share of capital in industry value added has significantly increased in manufacturing, much more so than in any other sector, broad sector, in the US economy.

My estimates, and I have three very different papers on this, one of them still unreleased, is that a very, very large fraction of that, 80-plus percent, is about automation. Monopolies might still play a small role, but it's not the main thing. Now, of course, if you look at some other sectors, such as retail, the picture is very different, and I have not written a paper on retail, and I'm willing to entertain that monopoly has been more important in retail.

Certainly, monopoly in digital tech industry is a huge problem, but when you look at the change in national labor markets until about the mid-2010s, manufacturing changes are very, very important, and those are not just about monopoly, or they are only tangentially about monopoly. They're much more about technology, and then secondarily about globalization. I completely agree with you, moving forward, we have to really deal with monopoly, and part of it is we have all these amazing antitrust tools which we haven't used for 40 years.

Bethany: Now you and Luigi agree. [chuckles]

Luigi: Daron, I have to admit, I'm shocked to know that there is a topic where you have not written a paper.

Daron: [laughs]

Luigi: I love this idea that technology should be more complementary to human skills. However, you know better than I do that technology is the result of endogenous choices, so you need to transform the incentives of people-

Daron: 100%.

Luigi: -to shape-- How do you do it. Is it through taxes? Is it through interfering in the government choices? How do you plan to do it?

Daron: Wonderful. Thank you so much, Luigi, for asking that and asking it in that very, very polite and nice way.

Luigi: [laughs]

Daron: First of all, I want to convince people that the menu ahead of us with digital technology, and especially with AI, has these complementarities in it. These complementarities are not pie in the sky. They are real complementarities that are feasible. That thesis doesn't imply that we know how to get there. I do not believe for a second that the government can, as an entrepreneur or as a bureaucratic planner, can tell you, "This is the technology we're going to adopt. This is the technological path we're going to go." That would be disastrous.

I also believe, and I think history shows, that government policy and also general social pressures do set the stage and the agenda within which entrepreneurial energies and incentives are shaped. For example, in an area like energy, it was those societal pressures and some government policies, some starting in Germany, some starting in Europe, some starting in the United States, many implemented in China, that led to a huge improvement in renewable energy sources.

So much so that from starting from levels where they were more than 10 times as expensive as fossil fuels, today they are cost-competitive with fossil fuel. That is a technological redirection. Not every aspect of that can be translated into the direction of AI between automation and human complementarity, but it is a very powerful proof of concept.

I think the most powerful way of going towards human complementarity would be if the tremendous amount of talented people who go into tech, rather than waking up every morning and thinking, "I'm going to build artificial civil intelligence today," they thought, "I'm going to build more human complementary AI." Then most of that battle would be won already. Now, how do you get those minds to be changed? It's about communication, it's about aspirations, it's about societal objectives, and some policies could be useful as well.

Luigi: I appreciate the fact that innovators have a role. However, when we look at Silicon Valley today, there are hundreds of billions of dollars that are invested with a very specific goal in mind, which is to make money. How? Mostly by replacing existing workforce. You need to create an equally attractive impulse for people to do it. How do you do it?

Daron: 100%. The overwhelming impulse in much of technology today is to develop what I call tools of automation that would enable companies to save on labor costs by cutting wages or by cutting workforce, and enabling those processes and tasks to be taken over by algorithms or capital. That makes shared prosperity much harder, and I would argue that if it goes excessive as it has done, it also makes the objective of gaining productivity from these technologies much harder, and therein is the opening.

Luigi: Sorry, I'm a little bit confused because I thought that productivity was precisely substituting labor. The reason why agriculture has become so productive is because machines have replaced men, mostly men, that were plowing, seeding, and so on, so forth. No?

Daron: 100% that automation is and can be a powerful source of productivity growth. If automation enables you to use cheaper machinery for more expensive labor, it increases productivity. However, if it goes excessive, what I've called source automation, you get a lot of replacement, a lot of displacement, but at the margin, capital is not doing so much better. You get things like dysfunctional customer service, which doesn't do its job. It replaces workers, but it doesn't deliver value to consumers or value to firms.

That's the opening where now, if the right kind of technology was produced by tech companies and understood and demanded by CEOs, it could be human complementary and productivity-enhancing, and the market could support it. There is no law that the market would never support human complementary technologies. That being said, there are a number of biases against human complementary technologies. Some of them are in policy.

For example, our tax code massively subsidizes capital and taxes labor. For two equivalent technologies, you would choose the one that's less labor-intensive, more capital-intensive because of the tax incentive. Managers who hate trade unions would love to have automation tools because it reduces their dependence on trade unions or organized labor.

There are other biases that may create a marginal preference for automation technologies, but some of those are avoidable, such as the tax code bias, and some of those would not be important enough if you really have very promising human complementary technologies. One of the things that I've emphasized a lot is that at a time like ours where we need a lot of new goods and services because of aging populations, changing global balances and economic relations, climate change, human ingenuity, and therefore human complementary tools that are innovative and adaptive, are really valuable.

Any company that can correctly apply these tools can actually make money and increase its productivity. That's why I think productivity-wise as well, there may be $100 bills left on the sidewalk, but it takes two to tango. If I'm a company and I'm really decided, I'm really convinced as a manager that I want to use my human resources better, I want to empower my workers with the right kind of technology, still my hands are tied if the tech sector is not producing these new technologies. It takes two to tango, but both the tech sector needs to produce these tools and the companies need to demand and deploy them correctly.

Luigi: Let's try to understand this because you have the same example that Danny Roderick about the solution is to have a nurse who, thanks to technology, is going to be able to do more of the jobs that now are done by doctors, and this will bring wage growth, prosperity, all the great things. If every nurse can do the same with AI, the relative scarcity has not changed. My prediction as an economist is that the price of those skills is going to go down and the nurse is going to make the same amount.

Now, if only a few nurses can do that, then you're going to generate a lot of inequality. What are you really telling us-- [crosstalk]

Daron: Thank you, [unintelligible 00:26:52]. That's another excellent question, which is central and part of the reason why I think a lot of discussion of "augmenting technology" in the media or even in economics is actually confused. You have to distinguish between workers becoming more productive in the tasks that they are already performing versus workers performing new tasks. Let's first start with the former.

London cabbies used to have very, very useful skills navigating London streets, which are quite complex. GPS-assisted navigation technologies come. Now anybody can become a cab driver or an Uber driver. What that does is that it commoditizes that specific skill because what that technology is doing is that it's making people better in already existing tasks. That is no way to create sustained wage growth because, as you've intuited, it creates two opposing forces. It makes you more productive, but it also reduces the price of the services that you provide to the market.

In most cases, those two effects more or less cancel out, so there isn't a huge wage gain for you. When you do new things, that second effect isn't there because you're not creating price pressures down on other people who are already doing that thing because you're doing something new, and that new thing is also adding to productivity. In the nurse's case, you can see that there is nuance there because some of the new things that nurses are doing, could be doing with AI, are things that previously doctors were doing. Now there is more complicated things.

I can give you examples from electricians, for example, or from creative processes from teachers where these are really new things, so they don't run into the tasks that other people were doing before. It illustrates that there is a big difference between what, as economists would call, labor-augmenting technologies and new tasks where you start doing new things that require new expertise, new capabilities, and add to the production process. It's the latter that would be, in my argument, the savior of labor.

Bethany: Has there been a time in history where we have been able to move to those new things without going through the stage, the ugly stage, of the labor-augmenting period first? I mean, can you have one without the other? Do you have to-- [crosstalk]

Daron: The two are often going to be intertwined. Both my very crude estimates in an early paper with Pascual Restrepo, but much better estimates by David Autor and co-authors, finds that in the United States today, more than 50% of tasks that labor performs are new tasks that did not exist in the 1940s. Those new tasks was really crucial for the very steady and powerful wage growth in the late 1940s, '50s, '60s, '70s. That came, of course, after a big shock, World War II and the Great Depression. Now, were they related? We can tell stories where they are related. We can tell stories where those links are weak, et cetera.

It is certainly possible for the market economy to invest a lot in new tasks and those new tasks to act as an engine of wage growth. That's why I don't think the alternative to high inequalities and socialism or government ownership of means of production or means of computation, but it's a better functioning market economy with the right technological investments and with the right democratic and regulatory framework.

Luigi: Sorry, your story of the new task is dependent on the fact these new tasks are not commoditized. Imagine that the nurses are now able to tell immediately you are prescription for something, and maybe something that doctors cannot even do now, but just because AI enabled them to do that. It means that this is a commoditized feature, so it will not necessarily carry a premium. It will not increase their wages.

Daron: No, no, that's why I said in passing, but it has to be new tasks that require new expertise. If it's a new task that electricians-- [crosstalk]

Luigi: Only the nurses can have.

Daron: Or only the right kind of training. For instance, one of my examples is electricians now using AI tools to do much more advanced work with electrical equipment, deal with troubleshooting and problems much better. You still need electrical engineering information and hence skills and experience as an electrician, but you don't need to be the most seasoned electrician with 30 years of experience and the best in your class to be able to do that.

It's a continuous race. That's why the first paper that laid out this framework that Pascual Restrepo and I wrote was called The Race Between Machine and Man or something like that, where you have these new tasks and over time they're going to start becoming commoditized or automated, so you need to keep on creating them.

Bethany: I think it's oversimplifying your point of view to say that you've argued that AI is not contributing to productivity, or rather at least that any gains to productivity from AI aren't showing up yet. What is the more nuanced version of what I just said and perhaps the more accurate one? If it is true that it isn't yet showing up in productivity statistics, why does it matter? If it's not a human augmenting tool in the end, if it's just nothing, why are we even spending all this time talking about it?

Daron: I certainly don't think that AI is a nothing burger. AI is already having major effects, and it's going to have much bigger effects. Your summary is pretty accurate. I think that AI's productivity impacts so far aren't large enough to show up in national statistics. What will happen in the next several years is the key. Even there, I am not optimistic that we're going to see an immediate, rapid productivity boom because these models are getting ready, but are not yet ready for prime time, and they're going to spread slowly.

Importantly, their displacement effects often can come ahead or far ahead of any productivity gain. You can have a lot of workers displaced from their jobs even without huge productivity effects. There is something to worry about even when the productivity effects are not materializing. The big reason to be talking about AI is that whatever its short-term impact, at least the agenda of artificial general intelligence, as both you, Bethany, and Luigi at different points in this conversation have emphasized, is going towards more and more labor cutting. It will have major effects.

Moreover, there is an opportunity cost in terms of failing to use AI to do the useful human complementary things. There is a lot to focus on so that we don't waste this opportunity.

Luigi: Speaking of opportunity, you are the leading economist of my generation. Imagine that in November 2028, Jon Ossoff is elected president and appoints you economic czar. You can pick whatever you want. What are the three concrete legislative proposals that you want to present to Congress in-- Hopefully, there is a Congress that still is voting-- in the first 100 days?

Daron: So many jokes in this statement, Luigi. Me being the leading economist, Jon Ossoff getting president, and Congress existing. All right. I wouldn't want the power of a czar, but let me tell you three policy ideas. Number one, clean up the US tax code so that it doesn't have this huge asymmetry between capital and labor, and in the process, contributing to both inequality and tax evasion and lots of things. I favor a tax code where all income is taxed symmetrically.

I would not want to go there right away because as every policy idea, there are risks. In an ideal world, we would experiment with, say, for example, increasing taxes on capital by getting rid of many of the giveaways, amortizements, and loopholes, and reduce taxes on labor, for example, by reducing payroll taxes to some extent, especially on low-paid workers, and then see what the effects on investment are. If they are not disastrous, then go to where I'm suggesting. That's number one.

Number two, this might come in as out of the left field, and you might think this is a minor one, but I think it's quite important to set the infrastructure for data markets in the United States, or perhaps globally, so that people can invest in high-quality data and other data that they produce are not expropriated and stolen by tech companies, which will have not just distributional effects, much more important, it would actually create opportunities for better development of technology because a lot of the more positive directions of technology that I'm talking about really require high-quality data.

If you want to build a tool that will help electricians to do the kinds of new tasks that I'm talking about, you need the best electricians spend hours and hours demonstrating how they troubleshoot very complex problems. That kind of data right now cannot be created because if it were created, it's going to be stolen. That's because data, by everybody's assessment, is going to become much more important than land for our future economy, but we have no data markets.

Data is a tragedy of the commons, so that cannot be a good thing, and I think we have to rectify that by building the infrastructure, laws, property rights on data, et cetera, which require some nuance because if I have a land, I should be able to sell that individually. If you have data, you may want to sell that individually, but there may actually be benefits from having collective ownership of data, so you need to have different laws for data.

Then finally, I think the US would benefit a lot from having an AI agency with perhaps some inspiration from natural institutes of health where a lot of expertise is concentrated. There are best practice ideas and also support for some under-researched parts of the AI landscape, as well as high-quality, objective, expert opinion on safety and other regulations.

Bethany: I really like the optimism in what you say [chuckles] and what you write about, but I guess my cynical brain worries that it is misplaced optimism and that when we look at the history of Silicon Valley in recent years, at least in the sweep of the last few decades, it's been a movement from optimism and human-first to exactly the opposite, profit-first and at the expense of humans. What happens if it does turn out that a lot of the utility in AI is replacing humans?

When you take into account both the lack of morality demanding anything bigger plus the profit incentive to do this, how do you fight against that?

Daron: Well, on this one, I have to slightly, slightly disagree with you, Bethany, that I'm not actually that optimistic. I spend-

Bethany: Maybe not good.

Daron: -many sleepless nights worrying about where AI is going as well as where liberal democracy is going. My emphasis is that left to its own devices, the current tech ideology and incentive system would really get us astray. I also think that, technically, there is a different direction that could be much more beneficial. That's the germ of optimism. That is a very qualified optimism for reasons that you and Luigi have also articulated.

Even if that's true, convincing people of that and finding the policies and the narratives to support that isn't straightforward. The government cannot legislate or dictate that the right direction or the more socially beneficial direction of AI should be taken. There are many pitfalls in there. The ideology of artificial general intelligence, artificial superintelligence, and the power of a handful of companies and their founders are overwhelming. There are many, many difficulties, but I think that path is still open. We have no option but try. I don't think that's overly optimistic, but there is just a glimmer of hope.

Luigi: There is no doubt that Daron is super smart and super broad, so very good at answering any question on the spot. This said, I am sympathetic to what he's saying. I'm not so sure I know how to proceed. I was intrigued by his three proposals. Except for the first one, which needs to be articulated much more, they're not really related very much with the theme of the book, how to make wage growth more stable for everybody, and how to make technology, being more labor-enhancing, how to create the incentives to do so.

Yes, there might be some psychological reason that you don't want to deal with the unions, but by and large, are economic incentives that make people substitute labor. If you want to create something else, you need to create some different incentives. Maybe I'm deaf, and I would like to, actually, if I was, please correct me, but I didn't hear any concrete proposal to make that change.

Bethany: Neither did I, and that's why I pushed with that question on whether it was even doable in the face of this elephant in the room, which is the market and the incentives that the market provides to do exactly the opposite of what he asked. On this whole question of the social structure, I guess I tend to think if wages are going up and people can live in a place they like and afford a home and believe that their children are going to have a better life and be able to provide for their children, that some of the social stuff goes by the wayside.

In other words, I think the social stuff is, in some ways, an outgrowth of economic instability. I'm not even sure it's a chicken and egg question. I think the economy comes first, and that's the part that's most important. I think if we don't fix that and don't fix wage growth, that all the other stuff is a little bit incidental. Maybe I'd like to hear you argue that if you don't agree with me.

Luigi: No. Unfortunately, I agree too much, so I cannot argue. [laughs]

Bethany: Yes, sorry. [laughs] Anyway, on the other part, I didn't hear anything either, and I guess I tend to be really, really cynical about that because it seems like everything is pushing in the other direction. The market, while a construction of human beings, also has a life of its own. For a lot of years, I've been saying the market is like Frankenstein, right? It's something that was created by humans and therefore, we should be able to control, but which has completely stepped outside of our control.

I'm not sure now, given the market incentives, meaning stock market incentives for companies to treat workers in exactly the opposite way that he advocates, I'm just not sure there's any fixing this. I don't know. It sounds well and good to have policies that would enable us to be human complementary, but I don't see a way that they rise organically. I mean, if you look at the pronouncements out of companies like Meta, it's exactly the opposite of that. Any policy I can think of that would be human complementary would come at the cost of the stock market, and I just don't see that happening.

Luigi: I think that's one problem, but the second problem, because I thought that our discussion on tax specificity and wages is very interesting and echoes the discussion we had a few years back with Besson, I don't know if you remember, but he was very much in that spirit too, which is to say, at the end of the day, what guarantees you a wage premium is that you have some unique skill that is not completely commoditized. The fact that nurses can do new things is not going to be the solution unless these new things can only be done by nurses with a particular training which are in short supply.

If everybody can do the same stuff, that's commoditized and not sell at a premium. I don't see, maybe it's my lack of imagination, but it seems to me that what AI is doing is actually commoditizing people and particularly making some human skills less specific. One colleague of Daron Acemoglu at MIT, Danieli, has this evidence, for example, that when you apply AI in even simple tasks like call centers, the lower-quality people are doing better.

In that particular case, you close the gap, but as a result, everybody becomes more productive, but the labor is commoditized because their high-skill labor is now worth the low-skill labor, so will be paid at the low-skill labor. I think that the pressure is there.

Bethany: He did say that there are plenty of examples throughout history where this has happened, where technology has been human complementary and has been value-added. I just have trouble seeing how it can and will be so in the age of AI. I also, which we didn't get into with him and it's not really part of his argument, but also that AI can be destructive in ways, too, that still create lower costs. One example is in medicine, where a lot of medicine to be done well still does need to be human, but the temptation is going to be to do it through AI.

Luigi: What is interesting, that if you take his argument about technology really, really at heart, in some sense, you go straight towards some form of socialist. Socialist in this initial Soviet spirit. You know the Soviet means in the factories in the Soviet Union, there were these councils of workers that were running the factories. Soviet is the council of the factory. Why? Because in the traditional Marx and Leninist tradition, the workers should take over technology.

As you know, Marx was not against technology. Marx was very progressive and embraced technology, just thought that technology needed to be controlled by workers in order to make society better off. It seems to me that without saying that's what he wants, which is fine if that's what he wants, but he should be open then to say, how do you get the control? Then you should have the worker-controlling factory or the union having some say, at the very minimum, some form of German co-determination.

Because it is true that in Germany, now Germany is going down the tube, so it's not obvious that long-term this was great, but Germany was less fast in replacing workers because unions had a seat at the table. They control 50% of the supervisory boards of all publicly listed companies.

Bethany: Isn't he saying something a little bit different than that, or am I missing his argument? In that I think he's not saying worker co-determination over technology, he's saying worker co-determination over setting the rules under which technology will be used. He doesn't want the government to be explicitly involved in saying this will happen here and this will happen here. Rather, he wants the government to be involved in setting the rules that will make it happen in a way that will be better for workers. There's a layer in there that I think allows some practical illusion, if that makes sense.

Luigi: You may be right, except the fact that he does not specify those rules. If he cannot specify those rules, let alone the government or the working class specify those rules, I actually believe more in residual right of control. If he said the workers can see when they see it. I think that in German factories, workers were favoring some technological innovation and delaying some other technological innovation. For example, they completely delay the innovation of electric cars because this will destroy a lot of jobs.

The reason why Germany now is in a funk is because they had to fire-- By the way, Volkswagen is the prototype of a company that is run by workers. Volkswagen has to fire I don't know how many thousands of workers because electric cars are becoming important and they need much fewer workers.

Bethany: I think he might almost be, although this might be unfair, but making a circular argument of sorts, which is that he wouldn't want that to have to happen and for the workers to be able to do that because the rules would be set in such a way that the workers didn't feel that they had to do that because it would all be working out, at the risk of overusing the word work. Right? [chuckles]

It's a bit of a duck in the end because it's the argument that if the rules were set in the right way, then we wouldn't need worker involvement in decisions at the factory floor because the rules would have been set in the right way. Then if you don't know what those rules are and how to set them, then you're right, it becomes circular very quickly. I was actually wondering. I had a slightly different question for you. I was thinking about working-class capitalism. Is it anything more complicated than wage growth? Can you just sum it up by calling it wage growth?

Luigi: The part that I was trying to get at with my first question is when he had this enlightenment of the crisis of liberalism because, unfortunately, I see a lot of my colleagues who woke up to the problem only when the working class started to vote for Donald Trump. Because up to that point, they were not creating any issue, and everything was great. What he wants is them not to vote for Donald Trump again.

He's talking about wage growth, et cetera, exactly, by the way, in the same way in which I was saying, the liberal society before was giving opportunity to workers to defeat communism. Now, you want to give money to workers to defeat Trumpists. It's not that you want it because that's the right thing to do. It's only because otherwise they go in the wrong directions. [laughs]

Bethany: So cynical, Luigi. So cynical, Luigi. Actually, I was thinking while you were talking that by one measurement, we have come a long way in the past decade, becauae I remember when Trump was first elected. I did not know Daron then, so I don't know if he was in this camp, so it's not fair for me to assume anything, but at least in the quarters in which I travel, the idea that economics were in any way to blame for Trump getting elected was something you could not say in polite society. The only reason that Trump was elected was because people were racist.

The idea that economics was to blame was you letting people off the hook for being racist, and therefore, it was something that you weren't allowed to say. The fact that in the past decade, now that's a pretty mainstream topic of conversation, that economic anxiety is at least-- some people may still want to blame it on racism, but at least that economic insecurity is a really real thing and a problem, that's actually enormous progress, I think, and enormous progress toward rationality and enormous progress toward a solution because you can't have a solution if you don't acknowledge what the problem really is.

I think I'll take that positive out of our conversation. What do you guys think?

Luigi: Oh, I completely agree. Now, this did not come without cost, at least for me, because I was an early believer. In fact, in my book in 2012, I pointed out the problem even before Trump was elected, and people dismissed it. When Trump was elected, I was trying to tell my colleagues that the problem was economic, so much so that I wanted to invite Steve Bannon, and I did invite Steve Bannon at the University of Chicago,-

Bethany: I remember this.

Luigi: -and I had to go with the security guard for a while because I was under threat. That was the reaction. Eventually, actually, Steve Bannon did not come because he was afraid of a real debate, but my purpose for inviting him was precisely to expose this economic anxiety because Steve Bannon has always been a proponent of the fact that people voted for Trump because of economic anxiety, and I thought he was right. He was wrong in a lot of other things, but in this dimension, was right, and so it took 10 years, but eventually people got it.