Human Life in a Post-AGI World
Talk at Google DeepMind | July 8, 2026
Thank you all for having me in to what is now my third talk at London DeepMind, but I’ve never spoken at DeepMind to so many philosophers before. I’m sure you’re all familiar with the problem of the ship of Theseus. The planks are replaced. At what point is it the same ship? This is a possible starting point for thinking about just the meaning of life in what some of you call a post-AGI world. I’m not convinced the AGI concept is well-defined, but the notion that there’s strong AI that’s very powerful, I’m completely willing to accept. I’m not a Gary Marcus sort of denialist at all. I just think it’s a continuum rather than a discrete point.
So what will lives be like? What will be the sources of meaning? I think in a funny way, there’ll be a surfeit of meaning for quite a long time. We’ll be drenched in meaning. It will exhaust us. So you read accounts of life in a post-AGI world, and somehow we’re sitting around twiddling our thumbs. Maybe you pick up a flute and play it, or you walk over and pet a sheep. Maybe in some very distant future, but that’s not how I imagine it.
I think what we will need to do for quite a long time is in essence rebuild all human institutions. This will be the most ambitious task humanity has ever undertaken. And like the ship of Theseus, you can debate at what point you have a new ship. It’s a common point, you know, Derek Parfit, other philosophers. The molecules in the human body, they turn over something like once every seven years. At the end of the seven years, are you the same person? Well, this is all going to be accelerated.
So the first question I wonder is simply like, how good a job are we going to do at this?
Rebuilding every institution
So I think of the UK, and in particular England, as maybe the most successful country ever, but certainly in the top tier, has had a profound, incredible impact on the world, the Industrial Revolution, amazing novels. I don’t need to go on here. You all know the story. And if you ask yourself the question, well, if you had to rebuild the UK today, you know, could you do as good a job as the people who built what we got to? And if you all feel a bit nervous about this prospect, I fully understand.
If I think of my own country, well, there’s New York City, which I consider to be quite a marvel. But you go to some other towns like Buffalo, New York, maybe none of you have been there. But in Buffalo, New York, any building, say before 1940, is just gorgeous, marvelous, like a nice building in London from the 18th or 19th century. Any building after, like World War II in Buffalo, is just horrible and ugly and terrible, and it’s parking structures and ugly banks. And okay, so people tried to rebuild Buffalo, and they just completely botched it. And this is in the United States, a wealthy country, what we would all consider a successful country, a country, at least then, still somewhat today, is good at building things. You know, it’s not that we failed to do this in Chad, Africa. We failed in the United States.
So the notion that we’re now entering this period where everything needs to be rebuilt, including this company, I assume, right? A lot of the things you do now with humans will be done with AI. Presumably many of them are already.
I was just in Paris with my wife. And as you probably all know, in the 19th century, Haussmann tore down most of the medieval parts of Paris and rebuilt them with these wider boulevards. I’m not here to debate if you prefer that or the elder Paris. The point is, this has happened many times in human existence. I think we know many times it’s gone well, and many times it hasn’t gone well, no matter what your particular aesthetic or practical judgments may be.
We just passed the 250th anniversary of the U.S. breaking away from you all here. We built something pretty amazing. If we had to do it again today, if we had to rewrite our constitution today, I would run away screaming. Today, when constitutions are rewritten, they seem to be 250 pages. Back then, we had a nice short one. Slavery aside, which we did get rid of, has lasted pretty well. So just the dramatic stakes for the world will be so high.
So like I live near Washington, D.C. D.C., as you know, is surrounded by this belt of non-profit institutions. Some are lobbies, some are think tanks, some are charities. I think you’d all agree with me if you work here, those will need to be rebuilt in some way. So I saw Institute for Progress advertising for a job where there’s basically they want to hire one person, and that one person with agents will do sufficient work to be equal to what used to be one whole think tank. Whether they can do that today, you can debate. But if not today, in a year or two, I actually think they can probably do it today. At least once, you know, a few new things come out in a few months, and we’re going to have to rebuild that. Hollywood movie making will probably change a good deal, and so on and so on. You all probably have thought of more examples than I have.
Now, the question of how long it will take us to rebuild everything, I don’t think we have a good answer. There’s people I know at Anthropic who somehow think this all happens in the next year or two. I think that’s very, very, very wrong. It’s not the official stance of the company. But human beings are involved. The human bottlenecks are slow. Even if we could do it in two years, and I’m pretty convinced we cannot, we’re not going to do it in two years.
It took us 40 years to optimally distribute electricity in the United States, which, again, is a pretty fast, pretty productive country. This is now talking about 19th century stretching into the 20th. That took us 40 years. I don’t think incorporating strong AI will take 40 years, but again, it’s not just a year or two.
Drenched in meaning
So for the foreseeable future of the lives of the people in this room, people listening, again, I just think the dramatic stakes will be so high.
The problem will not be how does my life get meaning, but how do I deal with all the meaning my life will have? A kind of exhaustion. And this comes up in the labor supply debates. So again, there’s one point of view like, oh, there’s AGI, there’s going to be mass unemployment. The more moderate, reasonable point of view is not that there’s mass unemployment. Many jobs still require humans. There’s comparative advantage. But total leisure time will go up. I think that’s likely the correct view, but across what time horizon?
If you think about your lives today, like I’m much busier and I’m busier because of AI. I’m working much harder. I don’t have to do that. But the point is my relative wage gradient for working harder today, it’s really quite extreme. And if I were, say, an 18 year old, I would feel I really had to work hard not to fall behind. There’s this new thing coming to the world. All sorts of people will be jumping on it. If I only start looking at it when I’m age 23, I’m behind by X number of years. So I would truly be working hard. At age 64, I don’t have to feel I need to work that hard. I can always just say if I choose to, well, I’m going to run out the clock, as they say, just kind of step back and wait until I die and I’ll be fine. I’m not going to do that.
Every time a new model comes out, I’m still excited. I used to be very excited. But they come out more and more frequently. And now I look at my calendar and I’m like, uh-oh, could you all wait a week, please? Because you want to be ready. You want to play around with it. You want to test it out. You want to talk about it with your friends. It’s a slight bit of an exhaustion. And again, for you all working here, you have access to models that haven’t come out yet or maybe will never come out. But all the time, you have fresh stimuli. And I hope, I think you must all be drenched in meaning. And you’re like, oh, my goodness, someone else tells me, look at this new model. What do I do with that?
So I think sometimes, like, when will the time come when the leisure dividend from AI arrives? No one is forcing you to work harder. But there’s a substitution effect from the higher implicit wage on your future earnings that if you work harder now, it will have a payoff. You’ll at least avoid being behind. But there’s a good chance you’ll be a leader of something. And again, you don’t have to be doing AI research. You could be like a horse trainer, a dog trainer. And if you’ve learned how to use AI, you have a big leg up on the other horse trainers and dog trainers. So it’s not literally true for every job. But when you read about, say, Mercor trying to master NBA basketball analytics using AI or using AI to write better poetry, the more you think about it, you realize it really is for most jobs that you have to get on the AI ball right away, and the sooner the better. So that, again, drenches our lives with meaning, even though somewhere out there is that leisure dividend, which I actually predict I’ll never see again at age 64.
The leisure dividend and the end of smoothing
If I see it, it will be because AI extends my life a lot, which I hope. I’m not really counting on that. But if AI extends my life to 112, I’m pretty sure I will see the leisure dividend. But in the meantime, human lives will have this extreme alternation between intense work right now for X number of years and then eventually a leisure dividend. We’ve done that a bit already with retirement. So in a funny way, it’s continuing a trend that was already in place.
You may all know John Maynard Keynes in the 1930s, who wrote this famous piece, Economic Possibilities for our Grandchildren, and he said, they’re going to work 15 hours a week and enjoy leisure. That was quite wrong. And I don’t just mean wrong at the AI labs, even more normal jobs. There aren’t that many people who work 15 hours a week. Even like in Germany, you work 30, 32 hours a week.
But what we have done is extend retirement. We extend it, both you live longer and you live a healthier life. So you can actually do things and you retire earlier. In the UK, you have like the triple lock pension thing, which is going to extend retirement all the more. Terrible idea in my view, but it seems your voters love it. And rather than 15 hours a week, you have this period when you’re younger, you work pretty intensively. And then when you’re older, you have this big, long period where you’re retired, semi-retired. The world already was moving in that direction. And it seems we’re going to get this big dose of a lot more of that.
Now, for me as an economist, that’s slightly weird. Economists typically will intuitively think in terms of smoothing. You smooth consumption, you smooth your leisure time, eating. Well, you have three meals, something like three meals a day. But you smooth that out too. You don’t want any one of the three to have too much food. And it seems with like meaning and this intensity of meaning, we’re doing the opposite of smoothing. I just wonder in this future, how much of the AI-rich future is the opposite of smoothing? Intense involvement and drama, and maybe stress and tension. And then at some point, it just goes away and you’re looking to pick up that flute, pet the sheep, visit your grandkids, your great-grandkids, depending how long you live. I’m not sure, but I think the point of doing these talks is for me to come here and present some ideas that maybe are not always in the debate.
AI maniacs and rule-followers
I also think there’ll be this special class of people. I meet a lot of them already. I call them AI maniacs. And I mean that in a positive term. Almost by definition, all of you are AI maniacs. But I had a Zoom call with a 15-year-old kid two nights ago, lives in Long Island. He’s like, I’m doing AI agents. Here’s my company. Can you get me a philanthropic grant? And I asked him, I’m like, what are your plans for going to college? And he’s like, college? My grades are so terrible. I haven’t thought about college. And normally, you’d think like, oh my goodness, this kid’s in trouble. We’ll see, but I don’t actually think the kid’s in trouble at all. I think the chance the kid will do great is reasonably high relative to his peers. He’s an AI maniac, and he’s completely absorbed in what he’s doing. Whether his particular business succeeds, I don’t know. But we’re going to see a lot more AI maniacs in just about every field. And they’re the people who will do well.
So I think the people who used to do well, and I’m fully aware that I’m in London when I’m saying this and not in Silicon Valley. There’s a certain class of person, and I meet them a lot in London. I call them people who follow the rules. They come from good families. They go to very good schools. They’re quite smart, may not be brilliant, may not be super ambitious, but they’re definitely ambitious. They want to live in a nice neighborhood in London, which to me is maybe the nicest place in the world you could live. They get some professional whatever. London, all these business services, there’s finance, there’s law, there’s consulting. You know the whole story here better than I do. And they do that, and they follow all the rules, and they went to Oxbridge, and so on and so on. And they just do really well, and they’re kind of masters of the universe. They’re never billionaires, but to think they’re earning like a million and a half, two million pounds a year is entirely normal. And they have these beautiful spouses, and the family, and maybe the dog, and they go to Wimbledon. I think that’s really endangered. I’m not sure how much of that will continue. I guess some of the people like that, they don’t really mind being bored or kind of following orders, playing the game by the rules of the system. I just think the AI maniacs will out-compete them or out-compete their companies. Those people, they’re not going to starve. Again, they’re smart. They’re clearly hard workers. They put in all these extra hours to become law partners. They’ll be fine. In the US, I sometimes joke they’ll be sent to Houston to be mid-level managers in an energy company, and they’ll earn 300k a year, not 1.5 million. Then I reassure them that the taquerias in Houston are better than in New York City. But that, to them, is a horrible outcome. Again, I know plenty of these people. Some of them are my friends. Some are people I don’t really like, whatever. They’re generally impressive. Politically, they kind of rule a lot of countries. I’m not saying they all have the same politics, but you ask who is prominent in a political party, who’s a bundler raising money for candidates, who has influence, who gets to shape the future of business in the country. These people have a major role in that, who might end up in the House of Lords or something.
Say two-thirds of them are toast over a 20, 25-year time window. That’s really a radical change in our countries and in our lives. In some parts of the world, I think if you’re in Houston, it will matter less than if you’re in London, because London is built on those people. You go to the Bay Area, there’s plenty of AI maniacs. I’m not sure how many are in London, but I’m sure the rule followers, there’s lots of them in London. London’s built on them. The idea that we have to reconstruct the main sources of energy for our economies, our polities, who’s backing our political parties, and the AI maniacs are going to be on the rise, for me, very dramatic.
I wouldn’t say I’m pessimistic, but I think with any change that big, you should be broadly agnostic. Things feel pretty good now. If you’re agnostic, you ought to be worried. Things can always get worse. Don’t assume it will go well and how politics will react to that. I can imagine a lot of political movements to try to slow down or stop that. I’m just not sure how well those can succeed, because AI work you can do from other countries. I don’t really think the US will have like a great Chinese firewall. There may be different forms of protectionism, which will slow it down. I expect that, but in some medium to long-term equilibrium, I just don’t see how you keep the more productive technology out of the market. I do think this will happen. That, too, will just be full of drama.
The declining return to being an intellectual
Other effects I’ve been thinking about, like how much will intellectuals matter? This is a sensitive point. I see Rob Reich on the call. He’s a professor at Stanford, highly intellectual, has written some number of excellent books, extremely well read. I’m very glad that Rob exists.
But the rate of return to being you, I think, will go down. I’m not worried about you. You’re like me. You can run at the clock, right? People already care. What does Rob think? If you’re starting out and you want to be the next generation of Rob Reich, I think there’ll be many fewer Rob Reichs, or for that matter, Tyler Cowens, because the AI’s already. You don’t even have to talk about AGI. They’re just better at knowing things. You know, you ask 5.6 whatever, or Fable 5, just don’t ask about bio. I mean, it does very, very well, and of course, it’s only going to do better. And once we train it to do more things, it will do better yet. That’s just a matter of time.
So we’re going to be in this funny future where I think individual intellectuals will be more important, but the ability to become one, I think, might be more difficult. And this is fully granting that this is a wealthier world. I don’t think it’s a world where there’s mass unemployment. So this isn’t a catastrophic scenario. It’s just the rate of return to that kind of smarts may not have that high status. So you’ll have plenty of people who become intellectuals like an espousal capacity. You might have people who are intellectuals who are supported by corporations. If you look at Stripe, which is in South San Francisco, they already have on the payroll some number of people who write intellectual material.
We’re at Google DeepMind. I forget how many of you here identified as philosophers. A whole bunch of you. Like, you’re living this world now. So if you think back, say, to 19th century philosophy, there’s many fewer philosophers. There’s Kierkegaard, right? Hegel. You all know the names here. They’re incredibly important. Even the lesser ones in relative terms have reasonable importance because there aren’t that many philosophers. Because the whole networks of super specialization, I think, will be somewhat attenuated. And if you inject fertility crisis, low enrollments into the mix, which seems already on the way, that makes it all the more so. So universities have way lower status, much less of a path to becoming the next Rob Reich or Tyler Cowen.
As a Tyler Cowen, I’m not thrilled about this. I sort of feel that what Rob and I do, I guess what some of you here have done, is good, and the world should copy it and validate it. But I have to ask myself, is that really correct? Or is that just my kind of greedy, selfish little intuition? Like, oh, I’m so important. There ought to be more like me. I’m not sure. Like, I’ll never be able to address that question very objectively.
Philosophy will feel more real
But if I think about the future of, say, philosophy, since you’re philosophers here, in a funny way, I think philosophy will be more important. You know, someone like Derek Parfit, I believe he did not, in fact, get a philosophy PhD. Plato, obviously, did not get a philosophy PhD. So people will be far less specialized. If you pick up philosophy journals today, most of it doesn’t interest most people. Some of you may know that Agnes Callard essay, where she says, I went back and I read my favorite philosophy journals in the 50s and 60s, and it was so much better than today. I think the comparison is more complicated than that. But that someone could even feel that way, I take to be striking and interesting. So I think philosophy will feel more real. It will be closer to normal life. It will be more like Kierkegaard, supported by companies, religious bodies, wealthy patrons, spouses. And this is philosophy throughout most of its history. Again, there’ll be a lot of pluses and minuses. I’m not pretending to give the complete ledger here. But that’s how I think many things will be.
And in a sense, maybe more dramatic. Like you go to Spotify, there’s so much music. None of it to me feels that dramatic. It’s sort of, I don’t mean the AI music, just even what was there, a lot of slop. And you don’t even know what year something is from. We may leave that world for a world where there’s more dramatic production, more dramatic creation, fewer creators, and they’re more important. And even if the AIs are better at some of these things, I don’t think the AIs will always be better at producing meaning. So you might think, well, the AI will be great at solving problems in logic and analytic philosophy. I definitely agree with that. But there’s something about like Hegel having come from a human, Kant, Adam Smith having come from humans that gives it special meaning. Or Taylor Swift, she sings the song as a human, like a woman who’s in love, just got married at Madison Square Garden. Whatever you think of that, the AI can’t copy that. The stories of Picasso, Plato, and so on.
So the people who are doing philosophy, music, whatever, they’re going to inject it with as much personal meaning as possible to compete with the AIs. It won’t just be, you know, here’s my proof of this conjecture from the Principia. The AIs really will just blow you out of the water if that’s what you’re trying to do. It will be things where people care that it was done by a human. So the philosophers will have these persona, which I think they did more in the 1960s than they do now. Kind of a Susan Sontag, like is she really a philosopher? Well, she’s thinking philosophically. Maybe a pop philosophy department wouldn’t hire her, but people read her to think philosophically. She had a big impact. I think it is, in fact, wonderful philosophy, and philosophy will be more like Susan Sontag. I think a lot of that will be good. Again, hardly a future drained of meaning.
Physical life
Next area, physical life. I think there’ll be a lot more stress on physical life. So the Rob Wrights and Tyler Cowens get sent a bit to the woodshed. So it’ll be harder to market yourself as I have all these great ideas, or I can tell you what you want to know. So, you know, maybe how you look will matter more. And I’m not sure the world will insist that everyone look perfect, whatever that means, like that you’re a 10 out of 10 on some like internet ranking board. But you’ll somehow want to look distinctive, or dress in a way that’s memorable, or make an impression on people, or have charisma. And you go around today, you look at the people who have charisma. I think in general, they’re often not like the best looking people. They just have charisma. Napoleon had charisma. He seems to have been kind of short and ugly, right? But he had charisma. So this future, I think there’ll be a lot more invested in charisma, face-to-face interactions.
You all expected me to do this like on a Google Meet call. No way, like I’m going to be here. Since GPT-4 came out, I have like 2.5x my physical in-person appearances, knowing those are the future terms of competition. I’ve tried to get better at being in-person. I write less, I’m writing fewer books, and I’m doing way more appearances, and I appear whenever I can. That’s my, you know, war against the machine, so to speak. It’s going very well for me, I have to say. But, you know, people have to remember something about you.
And some people are worried this will be this very superficial future, all about plastic surgery, and GLP-1s, or whatever we call their successors. People wearing weird things, like dressing like Agnes Callard, so they’ll be remembered. I think that’s fine, by the way. I like how she dresses. I don’t know, I think it, again, from my point of view, it’ll be fine. If you go back to medieval times, like what was your choice of clothing? Hardly anything. It was what you could make at home. Like, could you wear makeup? I’m not sure, but there can’t have been much choice. It wasn’t that there were these great Korean skin creams, right, in medieval Oxford. What were really your options for getting a good haircut, or good dental care? Like, very, very minimal, at best. So, in today’s world, there’s all kinds of things you can do for your looks. We’re all used to that. We’ve internalized it. All you hear, you’re, like, really good looking, right? Great. You’ve done it. You’re not miserable. You’re not, like, superficial, because you did that. So, if we just go a lot more down that path, just like we’re not superficial compared to our medieval precursors, those who follow us and invest more in charisma and looks, I think that’ll be fine. Maybe, in some ways, it’ll be stupider, because there’ll be fewer people investing and becoming, like, university intellectuals. But it will be much more like the general history of mankind.
And if you get outside a room of intellectuals and academics and just tell normal human beings, life will be much more about how you look, how you talk, touching grass, and less about having your nose in a book. Well, like, 85% of people are going to say, that sounds pretty awesome. Like, they don’t regard this as some terrible dystopian. The academics will. They’re terrified. They’re fighting back. Oh, the students, they’ll cheat with AI, whatever. I’m like, fine. I think big changes are coming. And again, negotiating those will be very dramatic and not at all short of meaning.
Religious life
Religious life, I think one thing we’ll see, and this is, again, pretty soon, it won’t be hard to create your own religion. I’m not sure many people will do this. I don’t think most people will. But they’ll be like accretions to the religions we have now. And I think with Fable 5, you could even do this already. Like, you ever actually try to read through the Hindu sacred texts? They’re pretty naughty, pretty detailed, quite long. Many parts are great and dramatic. I wouldn’t say they’re smoothly or evenly written. Not all of it is well written. They have significant meaning. For some people, a lot of people consume them through stories they’re told with their children. It’s not that every Hindu is like reading through the whole Ramayana. That’s all fine. But if you can sit down with, you know, the latest quad, whatever, and create your own set of sacred books. Again, I think like 2% of people are going to do this. Not most people. People have other interests, other hobbies. A lot of people aren’t religious. But if 2% of people do this, you end up with a lot of new religious accretions. Some of them will be totally new religions. But I think a lot will just be like, here are my sacred books of Christianity, or my add-ons to the Book of Mormon, or my whatever’s. There’ll be this extreme religious diversity. I don’t know, too much, too little. I think it will be quite different.
And you’ll get used to taking religious queries to the AIs. This already is the status quo. I’m pretty sure there’s no priest who knows the Bible as well as the strongest LLMs. So you have a question about the Bible, you know, where should you go? Well, you go to the LLM. Not everyone has internalized this, but a few years from now, it’ll just be understood. Rabbinical analysis, go to the LLM, and so on. So there’ll be this disintermediation of some of the priests, the rabbis, the imams, whoever the sacred intermediary figures are. They’ll be maybe modestly less important. So their ability to kind of crack the whip and say, hey, you can’t create your own religion, that’s not what I told you, that’s going to be a bit weaker, because you’re not going and asking them, oh, pastor, you know, in the Book of John, what does this mean? You’ll just ask your LLM, and the pastor will be some distant figure, or maybe he like sings a hymn at the church on Sunday, but he won’t quite have the force and authority. There may be fewer such people, just like there’s fewer Rob Wrights than Tyler Cowens. And again, that will make it easier for religions to multiply. Again, I’m not convinced it will be good. I think it will be pretty weird, very dramatic. Drama is not always a pleasant thing, but it’s another way to think about what this future will look like. Oh, I have like nine minutes left, two final topics, which are huge, but I’ll mention them.
Jobs: gathering data and running experiments
Jobs. So I mentioned before, I don’t believe in the mass unemployment hypothesis. I do think leisure time will go up. It has been going up for a long time. Two sets of jobs that I think will be prominent. One is just gathering data for the AIs. Even with good humanoid robots, I think actual humans will play the major role here. And good humanoid robots, you can debate when those will come, but there’s so much data in the world that’s not in AIs. And the returns to doing that, that could become maybe after healthcare slash biomedical, the largest sector in our so-called economy. And we might eventually exhaust doing that. But it seems to me the world’s always generating new data. If you imagine really a big chunk of the workforce gathering data, somehow marking it or grading it, organizing it, obviously with big help from AIs, but humans doing this, even just humans rewriting laws so data can be gathered. So right now you want to get data from a company, you need the company’s permission. You could imagine something in the future, a bit more like compulsory licensing, where you want to sing happy birthday, you have to pay a royalty at a fixed fee. There’s even bankrupt companies that have a lot of IP in their data, and it’s just sitting there locked up. No one’s using it. You’re not guarding any precious secret or creating any valuable incentive for someone to innovate with some other company. It’s just sitting there, dead government databases. So it’s kind of restructuring laws, making it possible to get at, process all this data. That’s going to keep us busy for a long time. And that’s what I think a lot of the world will be about, full of meaning.
Running experiments, I think, will be another fantastic growth sector. Right now, the number of humans who are involved in running experiments in the formal sense, it’s pretty low. It’s not, as far as I know, a job classification where you can get a number. There are people who, in essence, for a living, have experiments done on them. I think there’ll be so much more of this. This may depend how much you think AI can test propositions just by simulating. But I think even when AI can simulate, we humans, with our propensity to regulate, will still want, say, drugs, devices, procedures, tested on actual humans, even if we kind of trust the AI simulation. So that will be this other big growth sector. You could say humans as guinea pigs, but you’ll be on one end of the AI process or the other, gathering the data or being the subject of an experiment.
Healthcare sector, biomedical, we may even call it something different, but it will be a lot about gathering data, running experiments, and we’ll come up with all these things that help humans live to year dot, dot, dot. Again, I don’t know the number. It just seems to me people will live a lot longer because we’ll have all this stuff. I don’t at all believe what Dario once said, like, oh, in a few years, we’re going to cure everything. People in biosciences, even who are AI-pilled, don’t agree with that. So this will just go on for a long time.
An interesting question, like all these AI races, there’s you, there’s anthropic, there’s open AI. It’s this incredible race that just seems to accelerate. Like, does that just accelerate forever? What if that at some point ends in asymptote? And I’m not saying improvements stop, but it stops just accelerating, accelerating, accelerating. It’s a technical question. I don’t at all feel qualified to judge. I’m not sure anyone is qualified to judge, but I know I’m not. But your vision of the future, I think that’s an important question for just like, how dramatic is it really going to be? If competition across models and improvements just always accelerates as it’s been doing, like, especially since December, I think you would say, maybe a little bit earlier. That’s, again, kind of too dramatic for my taste, but very dramatic, human life full of meaning. I guess the economist is inclined to think of asymptotes at some point, but don’t take that as a technical judgment that I take seriously. It’s just my core intuition. And we may need that just for life to slow down a bit and not get too sopped up with all this meaning, which will be very stressful. So, you know, I think jobs, there’ll be lots of them.
Spreading AI to the rest of the world
Another job we’ll have, I call this imperialism, but I mean that in a value neutral way. But AI comes to different parts of the world at different speeds. I think the countries where AI changes a lot of things first, there’ll be a very high demand for people from those places, which I’ll think to be the US, possibly UK, to go around the rest of the world and teach people in other places how to integrate AI into what we have. And a lot of those demands won’t be fully rational. They won’t be, oh, give us the best possible AI. They’ll be like, oh, we’re Peruvians. We want to keep things a certain way. You may or may not agree, but we want you to give us a version of AI that helps keep it that way. And that will be the job. And I think Americans in particular, probably Brits as well, huge growth sector will be living in other parts of the world spreading AI. And again, the fact that AI can do it better may or may not be true, but I don’t think it’s what will matter. I think the Peruvians or some analogue will want humans to come and listen to their concerns and assure and persuade them as humans, that’s what they’re going to get. I’m not saying it’s always going to go well, but that will always be, I think, a big job for humans to do.
It’s already a growth sector for Americans to want to live abroad. Like we have all this accumulated wealth. Life in America can be a bit dull. Life in Europe in particular is amazing. Personally, I love life in most parts of Latin America. So it’s already a trend for Americans to live overseas. For another reason, it’s nothing to do with AI. So if there are all these future job opportunities, like full of meaning, like come to Kenya, help Kenya, you can save 73 lives or maybe like 73,000 lives, help them build out their AI in a way that’s acceptable to them. That’ll just be this phenomenally rich inner and outer life. And I think it’ll be a great source of job creation.
AI sovereignty and legitimacy
And just to close in my last like two minutes, the UK, everywhere you go, you hear this phrase, AI sovereignty. I mean, the deeper question is what will sovereignty be in this future? The AI will be so, I don’t know if powerful is the right word. Strong is not the right word. Knowledgeable, I don’t know. I don’t think we have the right word yet. But it will be more something, more simple or more, I don’t know, than governments in some ways. And whether the government can just swallow the AI and become the AI plus the government and the government takes the AI like a pill and becomes a super government. I mean, that’s one vision. I don’t think that’s impossible.
I don’t actually think that’s how it will work. They’ll somehow be competing like the government and the AI and they’ll be symbiotic and they’ll be clashing. And a little bit kind of like with the Pentagon and Anthropic has been, and it’s never quite resolved. And there’s always all this back and forth and no one’s ever happy. And there’s a kind of regulation by threat, but the AI companies are also indispensable. So you can’t just zap them and tell them to do what the government wants. Very messy, not what we’re used to. Not narrowly democratic in the traditional sense. So there’ll be so much more politics in some way. Again, I’m not sure this is good. I think the net picture of strong AI is likely to be good rather than bad.
But the question in this future, like where does legitimacy reside in like the government? And governments have already in many splintered thing. There’s like Washington DC, there’s a Little Rock, Arkansas, there’s the local mayor. Does AI really reduce the power of the local mayor? Because the local mayor doesn’t have the great system that say the executive branch does. I’m not sure, but that seems like quite possible. So it will change government a lot. And to think like a lot of our lives as political agents will be debating, like where will legitimacy arise? And that will be a growth sector. And even if the AIs can debate that better than we can, which I do expect, I don’t think we’re just going to let them do it. We’re going to do it ourselves. And that too, a lot of job creation, just one way or another debating where does legitimacy reside in these institutions and trying to figure that out.
And if AI keeps accelerating, like my goodness, that’s tiring. It’s so dramatic because you figure it out. And then three months later, well, the new models, the new whatever’s have come out and you’ve got to figure it out all over again. Like my goodness, a bunch of funny world.