← Conversations with Coleman
Conversations with Coleman
Humanity in a Thousand Years with Will MacAskill (S3 Ep.29)
01:03:38
My guest today is Will MacAskill. Will is an associate professor of philosophy at Oxford University. He is the co-founder and president of the Centre for Effective Altruism. Will is also the director of the Forethought Foundation for Global Priorities Research. In this episode, we discuss his new book "What We Owe the Future". We talk about whether we have a moral obligation to the billions of humans that will be born in the next several 1000 years, and how to weigh those obligations against those of living humans. We discuss population ethics in general, and Derek Parfit's Repugnant Conclusion thought experiment. We discuss the role of economic growth in humanity's long-term future and how to weigh that against present-day wealth inequality. We talk about the ethics of abortion, and the notion of moral progress. We also discuss the possible AI futures that lie ahead of us and much more. -Get 20% Off and Free Shipping with the code [20COLEMAN] at Manscaped.com. That’s 20% off wit...
Loading summary
Transcript70 lines
- [00:00]
A
SA.
- [00:31]
B
Welcome to another episode of Conversations with Coleman. If you're hearing this, then you're on the public feed, which means you'll get episodes a week after they come out and you'll hear advertisements. You can get access to the subscriber feed by going to ColemanHughes.org and becoming a supporter. This means you'll have access to episodes a week early, you'll never hear ads, and you'll get access to bonus Q and A episodes. You can also support me by liking and subscribing on YouTube and sharing the show with friends and family. As always, thank you so much for your support. Welcome to another episode of Conversations with Coleman. My guest today is Will McCaskill. Will is an associate professor of philosophy at Oxford University. He's the co founder and president of the center for Effective Altruism and he's the director of the Forethought foundation for Global Priorities Research. In this episode we discuss his new book, what We Owe the Future. We talk about whether we have a moral obligation to the billions of humans that will be born in the next several thousand years and how to weigh those obligations against those of living humans. We discuss population ethics in general and Derek Parfit's repugnant conclusion thought experiment. We discussed the role of economic growth in humanity's long term future and how to weigh that against present day wealth inequality. We discuss the ethics of abortion. We talk about the notion of moral progress. We discuss the possible AI futures that lie ahead of us and much more. So without Further ado, Will McCaskill. Okay, Will McCaskill, thanks so much for coming on my show.
- [02:10]
A
Thanks so much for having me. I'm a big fan.
- [02:12]
B
Yeah, I'm a big fan as well. I've been aware of you for a long time and you work with effective altruism, your work as a philosopher, and I imagine a lot of listeners to this podcast will be aware of you too. And if podcast listeners remember my episode with Peter Singer from well over a year ago, I think they may know that you are. I don't know if you would agree, but I would say you are kind of the inheritor of his sort of philosophical and other legacy as a thinker and as an influencer in the world.
- [02:42]
A
Well, he's got big shoes to fill. I don't know if I fill them, but I'm certainly extremely inspired by Peter.
- [02:48]
B
Yeah. So we're here to talk about your new book, what We Owe the Future, correct?
- [02:53]
A
That's exactly right. What We Owe the Future.
- [02:56]
B
So my first question is in the movie Tenet by Christopher Nolan, which came out two years ago. Basically, the future wages war on the past by trying to invert the flow of time. And the heroes of the movie have to prevent this, prevent an apocalypse of. Of the future, attacking the past or the present. And they do this in my favorite line from the movie, Kenneth Branagh says the future is angry because their oceans rose and their rivers ran dry. And it's this sort of dramatic moment. And basically the future has legitimate reasons to be angry with us for not having prevented calamities like climate change and other things. So is it fair to say that the impetus for your new book is to prevent the future from hating us and therefore making you a sort of philosophical member of the Tenet team?
- [03:49]
A
I would love to say I was inspired by Tenet. I have watched the movie. I admit I thought it was pretty absurd. I'm not sure that made sense, but I am inspired by the thought, yeah, what will future generations think of us? What will they think of how we act now? Will they think of us as moral pioneers who really pushed forward kind of model progress, or were far sighted and took prudent actions to safeguard a brighter future for the centuries to come? Or will they see us as, you know, myopic, selfish, not caring about our descendants, where when we look back, we see it as a mixed bag? There are some ways in which people in the past have bestowed on us amazing gifts. Medical, technology, science, the institutions we live under that have evolved over centuries, the legal systems, the moral world in which we inhabit, that is the product in part, of countless the forms that activists campaigned for. But at the same time, there's also negative legacies. We have things that are broken about the world, prejudice. We live under the threat of nuclear weapons. And so when we look to the future, yeah, both things could be in play. We could leave a future that is absolutely wonderful for our grandkids and their grandkids and their grandkids. Or we could leave an absolute dystopia, remains of a collapsed civilization or global totalitarian state or nothing at all. We could wipe ourselves out entirely. And it seems like when we realize that these really are the paths, that the humanity's future could go down and that we can, as individuals and as a collective, actually make a difference to which of those paths we go down, then, yeah, the suggestion, okay, we should join Tenet and work together to try to make the future appreciate us, thank us rather than hate us. Seems pretty compelling to me. Yeah.
- [05:42]
B
So one of the key points about the philosophy of thinking about the long term future is that there are going to be or there are likely to be many, many more people alive in the future than there are alive today than there have ever been alive. And depending on your philosophical views, that may or may not matter. If you're a utilitarian and you think broadly that two lives is twice as important as one, right, that it's twice as bad to kill two people as it is to kill one person, then that intuition would lead to thinking that the future might just matter a great deal more than the present and the past has ever mattered by sheer numbers. So what do you make of that video view? Do you subscribe to that view or do you somehow adjust for the fact that people don't exist yet broadly subscribe.
- [06:32]
A
To that view, where I think it's not just the utilitarian view that says two people matter twice as much as one person. I think that's just absolutely common sense where if you can save the lives of 100 people or you can save the lives of 10 people, I think it's just absolutely common sense that you should save the hundreds rather than the 10. And that's precisely because everyone matters equally. And then, as you say, it's this striking fact how early we seem to be in history, where there are 100 billion people who have lived in the past, 8 billion people alive today. But when we think at how many people are yet to come, well, I think it's at least trillions, perhaps trillions of trillions, where even if we just live as long as a typical mammal species, we've got 700,000 years to go, Earth will remain habitable for hundreds of years. The stars will keep burning for hundreds of trillions of years. And if we don't cause our own extinction in the near term, then there's no reason at all why we couldn't survive for such long time periods. And if we do, then we are really at the beginning of history and the number of people in the future just are really vast compared to the number of people alive today. And so that does mean that I think, and this is the argument of the book, that when we take a moral perspective, when we think about what's really of greatest importance today or in our lifetimes, what's of greatest importance are those things that will impact the entire course of the future. Because there's just so much at stake, so many people who will be affected, even if they're not very salient to us today. Does that mean that we should be absolutely fanatically? Focused on the future, no matter how small a difference we can make to it. Give no concern at all to the present? I think not. I think there are reasons for giving special concern to the present. And I'm not a big fan of fanaticism in general. But nonetheless, at the moment, how much of the world's resources do we spend really trying to make the world good, not just for our children, but for our grandchildren and their grandchildren and so on? And there's targeted way it's less than 0.1%. Maybe it's 0.01%. Should that be more? Should that be 1% or 10%? Seems pretty plausible to me. And that's what I'm really trying to push for.
- [08:27]
B
So as I was thinking about this topic, I often think it's sometimes fun to switch up space and time as dimensions. Right. So we would all accept, at least modern human beings would all accept that it's wrong to steal wealth from another place to make your country wealthy. That's a kind of theft. But if we think of ourselves as stealing wealth from the future, that maybe we may recognize in the abstract that that's wrong, but it doesn't really move us as much as stealing wealth from someone in a different location. But if you just consider that space and time are not all that different as dimensions, is there a reason why our being selfish today at the expense of people in 200 years is any less evil than, you know, a European nation being selfish at the expense of a Latin American nation?
- [09:19]
A
Yeah, I think fundamentally there's no difference there. So in what we are the future, they give this thought experiment of if you leave some glass that's shattered on a path you're walking on and you just don't clear it up. And in a hundred years time, someone cuts their foot on it and is harmed, seems like it just really doesn't matter. The fact that that occurred in the future and harm someone that maybe wasn't even born at the time that you dropped the glass. It seems just like harm is harm no matter when it occurs. And so I think the fundamental explanation of why now we start to appreciate that harming people on the other side of the world is seriously morally wrong, but not really thinking about harms that we might be inflicting into the future. The key thing there, I think seems to me to be salience. It's just that now television and the Internet and so on, people on the other side of the world can actually make the case for themselves. They can advocate people in the future, can't they're very literally voiceless in the world today. They're disenfranchised. And that means that there's kind of nothing else you can do apart from have altruistically minded people try and think about what are the ways we're impacting future people and how can we stand up for them.
- [10:26]
B
Ironically, it's just occurring to me that you're completely right that people in the future are voiceless. We never hear their concerns, just by definition, but people in the past actually have quite a big voice. You can spend your whole life reading the memoirs of dead people that suffered the worst atrocities imaginable, and actually you can actually be moved to rectify them, even if everyone hurt by them is dead. And obviously we memorialize things like the Holocaust and slavery and all kinds of other wars and atrocities partly because it makes us feel like we are. We stand in good relation to seeing how monumental those horrors were. But it. All of that is a kind of. It's sort of the dead having a voice and a say in a way, because we have all. Everything they wrote and everything they recorded and all the rest. And there's something a little strange about that, considering that the dead are precisely those who can no longer suffer or flourish, whereas the future contains almost everyone who is going to suffer and flourish.
- [11:29]
A
I mean, people are often skeptical of like, well, how much can you really influence the future? But if we look to the past, people in the past influence today all the time. And you mentioned by lighting in particular. So we're constantly looking at the works of literature or philosophy or science of people in past generations, and certainly we're very often building on them too, and via that, having an impact. Those people in the past who are long dead are having an impact today. And there's some case for thinking just the earlier you are in history, the more influence you can have over the future to come, because you are helping set the contours of what that subsequent generation do and are, what they believe, what their values are. And that continues for those generations too. And if that's right, then we are enormously influential out of basically everyone who will ever live, because we come so early. And yeah, we're at this surprisingly early time, you know, the first 10,000 years of civilization out of what might be hundreds of thousands, millions or billions of years.
- [12:31]
B
Right. And we're also in a Goldilocks zone where before civilization and before writing, if you were just a hunter gatherer, you actually can't really influence the future that much, despite the fact that it's like you're an early investor in Facebook, but you're too early and you got out too soon. So we're in that Goldilocks zone, where we can really shape the future.
- [12:52]
A
And in particular, where we have an understanding of the future as well, where hunter gatherers did have a big impact on today. So in what we are the future, they talk about the extinctions of the megafauna. So there used to be these great diversity of large animals across the planet, not just in Africa and the Indian subcontinent, but there were armadillos the size of cars roaming around South America. Wow. And humans. And, you know, there were sloths the size of African elephants, and loads of them died. Almost all of them died off at just after the time that humans came on the scene. So it actually seems like there was this enormous extinction event of large animals caused by humans overhunting and just changes to the environment from humans. And that is an impact that we still see today. I would like to see a glyptodont. I will not be able to, because humanity made them extinct about 10, 15,000 years ago. However, they didn't understand what they were doing. And that's a way in which the present era, say, the last hundred years, is very different from previous periods in history, where we now have a good enough scientific understanding that we can see that the extinction of a species is an irrevocable event, that the extinction of humanity itself would not be something we could come back from, and I think can see other ways, such as the transmission of information or impact on values that we can influence far, far into the future. So it's not only that we can have an impact, but also we can have a foreseeable impact.
- [14:15]
B
So I want to come back to this intuition that two lives is twice as important as one, which is. It's an intuition I share. I think it's worse when the Las Vegas shooter kills 50 people and shoots 500. I feel even worse. He's done even more harm than a mass shooter that kills five people. And I think everyone gets that at the same time. There's the famous repugnant conclusion of Derek Parfit, which basically says, if you take that line of reasoning to its logical endpoint, which is you could get two possible worlds, one where there's billions and billions and billions of people living just barely at subsistence, like lives that are barely worse, barely better than being dead.
- [14:59]
A
But.
- [14:59]
B
But that would have more sort of utility, more happiness. It would be a better world than a life where there were just a Couple people living really truly happy lives. And so what do you make of the problems with carrying the philosophical problems with carrying that intuition to its logical endpoints?
- [15:19]
A
Great question and always happy to talk about profit and population ethics. So I have a chapter of this in the book, chapter eight, and I think it's the first time that these issues have really been discussed for the general audience. And it was real challenge to try and get some very heavy material, technical material, presented in a way that was understandable. But the first thing I want to distinguish is between being scope sensitive, that is just caring about the stakes and the magnitude of the stakes, distinguish between that and population ethics, which is about changes to the size of a population. So I can have the view that it's more important to save 10 lives than five lives, and it's twice as important to save 10 lives as five lives. I can have that view and have any view I want about whether it's good or bad to bring new people into existence, which is the issue that the repugnant conclusion concerns. So even being a utilitarian, for example, that's just about what you ought to do when you've got a kind of fixed population of people. So you've got the same people in either decision. But then there's this field of population ethics, which is, okay, well, now suppose we're changing the size of our population. We're creating new people too. Under what conditions is that good or bad? And yeah, for the audience, I'll just restate the issue of a repugnant conclusion. So, yeah, let's say you start off with 10 trillion people in the best possible bliss. Seems like that's a really pretty good world. Now suppose you can make the world have twice as many people and they're just slightly worse off. You might think, if you think, oh, it's good to have extra people as long as their lives are very happy, okay, that makes the world better. So it's now 20 trillion people have lives of just slightly worse bliss. And I keep making that argument over and over again. I double the size of the population and just slightly decrease the people's wellbeing. If you thought it was better those first few times, and you keep saying, look, you're getting more total happiness by doubling the size of the population, but having each people's average wellbeing a little bit less, then as you say, we end up at this world where you've got this enormously large population of people whose lives are just barely worth living. And that seems like not a world that we should think is better than the world with 10 trillion people with extremely good lives. And that's why Derek Parfit called this the repugnant conclusion. And I think there are actually really quite strong arguments for wanting to ultimately endorse the repugnant conclusion. Where the issue of population ethics is one of the hardest areas of model philosophy. And you can show, in fact, and I do in the book, that any view that you have has devastatingly unintuitive consequences. So it's a kind of. Well, it's within an area of morality that where we just have to face a paradox. And maybe I can explain that if you want.
- [18:03]
B
Yeah, no, I'm curious, what are the. So I mean, I know what the devastating thing is. You have to accept if you think more people is better at a slight price in average well being, what's the opposite, if you don't accept that view, what do you have to accept?
- [18:16]
A
Okay, so let's start off with this 10 trillion people of incredibly blissful lives. And let's call that the A world. And now I'm going to change that world by just a little bit. I'm going to add another 10 trillion people. And let's say that the people in that A world, their bliss lives, that's like +100, let's say. And now I'm going to add another 10 trillion people and they're gonna have extremely good lives as well, but not quite as good. And let's say their well being will be like 98 rather than 100 but still extremely high. And I'm gonna take the original people and make their well being a little bit better. So they're now 101. So you've made life better for the existing people and you've added another chunk of people with lives that are extremely good. Now do you think that makes the world better? Like if you can switch from one to the other?
- [19:04]
B
Yeah, I would want to say yes intuitively.
- [19:08]
A
Okay, that makes tons of sense. So that principle that if you're not making anyone's life worse and you're adding good lives, then the world is at least not worse. But let's say it's better for now that's called dominance addition. These all have quite technical names. And that's world A. We're now going to move to world B, which is where you don't change the size of the population at all. But we take this slightly unequal world because we've got 10 trillion people at well being 101 and 10 trillion people at well being 98, and we're just going to average it out. So we're going to make everyone equally well off and we're going to slightly increase everyone's well being too. Like on average. So rather than a world of 10 trillion people at 101, 10 trillion people at 98, we're going to have a world where everyone has well being, let's say 99.6. So there's more total well being, there's more average well being, and it's more equally distributed. So the question is, is the world B better than this world A?
- [20:05]
B
That one is a little bit less obvious to me only because you've gotten rid of a certain level of well being. Like if 101 was to be like the Dalai Lama, then no Dalai Lamas exist in this world B. And it's not quite so obvious to me that that's better.
- [20:22]
A
Okay, terrific. So maybe that's the one that you would most want to reject. But that principle of if you can move from one world to another, where the second world is more egalitarian, like as in has a more equal distribution of well being, has greater total well being, greater average, that's called, it's a really catchy name of non anti egalitarianism. If you accept that, then notice what's happened with A moving to B. We have doubled the size of the population and slightly increased, decreased the average well being. And we've concluded that if you think A plus is better than A and B is better than A, or at least is not worse than A, then you have to conclude that B is better than A. But if we've done this one step of saying okay, a population that's twice as big but little lower well being is better, well then you can keep iterating this process over and over and over again and you move from the A world which has 10 trillion people of blessed lives to the repugnant conclusion world. And so what you have to reject, if you reject the repugnant conclusion, then you have to reject one of non anti egalitarianism, dominance, addition or this other principle, transitivity, which is that if B is better than A and A is better than A, then B is better than A in just the same way as like if three is bigger than two and two is bigger than one, then three is bigger than one.
- [21:38]
B
Yeah.
- [21:39]
A
And in my view these other principles are just extremely strong indeed. And it's therefore very hard to escape the repugnant conclusion. You know, we've got this like strong argument for that. We have to reject one of these principles. It seems like the one that has the like least is the least bad in terms of how unintuitive it is. But different people can differ. It's a very hard alien of marvel philosophy, as I say.
- [22:03]
B
Yeah, no, I'm not really sure what I think of it. I've never been sure. It's been a long time since I read Parfit, but I know. Is there some way in which this whole thought experiment has a bit of the continuum fallacy to it? It's like the. You know what the classic example I always heard in philosophy class was like, you know, if you just take one hair off your head at a time, there's really no single point at which we can decide whether you're bald. And so on that basis you could say, well, there's really no difference. If you accept that taking one hair off my head doesn't make me bald and taking the next hair doesn't make me bald, then I never become bald. It's like at some point you do, even if we can't pinpoint where that might be. And so I wonder, is there some way in which reducing the average wellbeing from 100 to 99.6 is like taking the first hair off of the head? And there's. I guess there's no decimal point at which. At which life becomes too grim for this, for this intuition to still hold that it's no big deal to pull the next hair off, but the sum total of it. We've actually created a. We've created a world where we would reject whatever intuition got us here. There really is a difference between bald and not bald. Does that make any sense?
- [23:18]
A
It makes total sense. And yeah, this is an area of work by moral philosophers. One of my colleagues at this institute I helped to set up, the Global Priorities Institute, Cheruji Thomas. He has done work on exactly this question of whether you should. Whether the best way to understand this argument I gave in favour of a repugnant conclusion should be rejected for the same reasons that we reject the argument for, why we can take all of the hairs off someone's head and say that they are not bald.
- [23:48]
B
Oh, I'm not the first person to make that exact analogy to hit.
- [23:51]
A
No, no, not at all. That makes me feel you're hitting on work. That's already a line in moral philosophy. And yeah, ultimately I think it doesn't work. I think the underlying argument still does go through and it's an importantly different situation. But it is something you could say where for one of these two principles, either the dominance addition or non anti Galatavianism. Yes. When you're at the point of 10 trillion lives and we're making these slight changes, like I said, from the A to A plus world, A plus to B, yes, that is making things better. But at some point along the continuum, it's no longer making things better, the principle no longer holds. And it's not that we can say exactly where, but at some point then it breaks. And I think that is something along those lines. That's among the more plausible ways of rejecting the repugnant conclusion. So one view you could have is just that it's good to have an existing another person in existence, but only if their life is sufficiently good. So there's like the zero life at which a life is just has a little bit more good than bad in it. Perhaps that's not actually makes the world neither better nor worse, even makes it worse. But if a life is above that level, it's just like actually no, it's like very good, then it's still a good thing. If you had that view, then you would endorse these principles in particular dominance addition, you would endorse it at the very high level that we started with. But once you got along this kind of continuum and you are now talking about more mediocre lives, is it good to add more of them in order to benefit someone, benefit existing people by a little bit? You would actually say no, a life has to be sufficiently good to make it a good thing to add someone to the world.
- [25:30]
B
Okay, so another topic here that I think it crossed my mind because of the intersection of this topic and everything that's going on in the news in America right now, you know, with the ethics of abortion and how our intuitions about future lives as opposed to present lives interact with the issue of abortion. Right. Because a fetus is not a person yet I would say it has, doesn't have beliefs or emotions or pretty much anything you would associate with full throated personhood, which is I guess a more of a philosophical concept than a layman's concept. But you know, this is not to say a fetus isn't human, it's to say it isn't. It doesn't have all the attributes of a person, interests and desires and expectations for itself and so forth. But the way that you're thinking about future people and valuing future people would suggest a kind of really strong concern for people that don't exist yet, people that we actually don't even know for sure are going to exist, but we just think are likely to exist and really taking their interests seriously despite them not being around to advocate for themselves yet. And I'm curious how that general attitude interacts with whatever your position is on the ethics of abortion.
- [26:52]
A
Yeah, so I'm certainly pro choice and don't think a government should interfere with a woman's ability to terminate her pregnancy. And the crucial thing is that there's a difference between killing a person and preventing the existence of a future person where, you know, I think it's much more wrong, seriously wrong to kill somebody. And so if I thought that early stage foetuses were persons, I would think it would be wrong to kill them. And that would be very obvious. But that's a very different thing from failing to bring into existence a happy life. And now I do think it's good for like, I do in general think that, like, actually having kids is a way to have to kind of do good in the world. So there's this common meme of it's actually bad thinking it's bad to have kids because of climate change, because of the carbon impact of carbon emissions a kid will have. But I see that as only kind of looking at one side of the ledger where a kid, if you raise them well, will go on to innovate, perhaps create technology. They'll pay the taxes, they'll contribute to society through infrastructure. If you bring them up well, they'll be moral change makers. And if they have a sufficiently good life, I think it's a benefit to them as well. In virtue of their good life, the world is a little bit better. And so I do think that means that if we can certainly, as a cultural matter, try and encourage people to have kids, certainly not be opposed to it on moral grounds, I think that's quite compatible with the state not getting involved, certainly. And people being able to have choices over when and how they have kids, such as fire being able to terminate.
- [28:22]
B
Yeah, no, I guess my, my position on it is so I have for a long time bought into the notion that lives which don't exist yet matter and ought to enter our moral calculus more than they do. And that's always pushed me towards the position that everything else held equal. To terminate a life, to terminate a fetus is a bad thing. That's separate from whether it should be legal or illegal. I mean, there are lots of things that I think are not good things which, which I strongly believe should nevertheless be legal to do just because in practice so many psychologically Normal people are going to want them for totally understandable reasons. And, and to ban them is a. Is just highly impractical to, to sort of legislate morality in every ways. And we could talk about, you know, whether it's drugs or prostitution or gambling, all, all sorts of things, what might one might think are bad that are nevertheless so tempting and so easy to fall into until the end of time for normal people that to penalize them with jail time creates a totally dysfunctional circumstance. So I've always felt that I'm pretty much just, I've always been legally pro choice quite strongly, but I also feel there is. That it is an abortion is not a good thing. Right. It's not something to be celebrated. It's basically, it's a failure of contraception and planning on everyone's part and an emergency, basically. And it's no fun. And I guess that's how I would reconcile my, the position of strongly caring about future persons which don't exist yet, with my pro choice position.
- [30:02]
A
Yeah, I mean, you're certainly.
- [30:04]
B
But I do think on the face that there sort of is a tension there. Right. Like there is, which is to say much of the pro choice movement would simply deny that. They would simply stick to the line that I only care about it if it's currently a person. Right. Which I think is not the right way to argue for the pro choice position.
- [30:20]
A
Yeah. So I think it's certainly just widespread. My guess is just the mainstream view among people who are pro choice that it is just, you know, there's a real loss in the course of having an abortion and like, of people I know who've had abortions, that's how they felt. I'm curious, just as a philosopher digging into your view though, whether you would also think it's a shame. Suppose there's just a. There's not yet an embryo, but there's a sperm and an egg and you know they're going to meet and you could prevent that them from meeting. Like, would that also be a shame in your perspective? Because that is if you prevent the sperm and egg meeting, then you are preventing a kind of future life just as much as if you terminate an early stage, very early stage fetus or destroy an embryo. And it seems to me that like we have different intuitions about that case than about the case of abortion. But yet from the perspective of oh, is it bad to fail to bring into existence a new person? They are comparable because again, there's like an identifiable person who will not exist.
- [31:22]
B
Right.
- [31:23]
A
In both cases.
- [31:24]
B
So the way I mean, so the problem with this is when I can think about it from one direction or I can think about it from the other direction.
- [31:31]
A
Right.
- [31:31]
B
If I think about it from one direction, it's what you just said, right? It's like, how is it so different to terminate a zygote as opposed to simply preventing, you know, doing what the, the plan B pill does? Right. It's like, where do you draw the line here? Because you sound crazy to think that to say that preventing two cells from meeting is some, is a bad thing at all. Right. It seems like totally neutral. So viewed from that way, you can collapse early term abortions into your intuitions about basically contraception essentially. At that level you could be saying a condom is bad because it's preventing the sperm. But. But then you can collapse it from the other direct. And you know, everyone would agree that a baby that just came out the womb is something you can't just kill for no reason. And then you just run it back one second at a time to the moment before it's delivered, and then the week before it's delivered. And then you could collapse your entire moral intuitions about fetuses, even early fetuses, to your intuitions about a baby. And I'm not sure either of those ways of thinking are the right way to think about it. It's actually another kind of case of the continuum fallacies. I'm not really sure where the line is to be drawn. Truthfully, I have no strong opinion on that. But it seems like most, most of the European countries are fine 12 to 14 weeks to be a reasonable compromise between people with different views. But yeah, I suppose I try not to collapse my intuitions into either it's always a baby or it's always two cells about to meet.
- [33:01]
A
Yeah. And I think that's actually, you know, from the theoretical philosophy perspective, a position that's actually kind of hard to maintain. I mean, you could think that it's just that over the course of the development of the foetus, there's an increasing risk that it's a person and so increasing risk of doing harm. But I think that's not really how people are thinking about it. I think instead it's like the feelings of loss, the feelings of this being a moral issue, I think tend to just come from a more psychological place of just like building up attachment to a being and a future being, that which is kind of independent of the morality of it per se. And then in terms of the issue of the future life, then if you do have the view that. Yeah, I mean, I guess I do have the view that under some conditions at least it's like, yeah, it's a good thing to have kids and good people, happy people with flourishing lives into the world who will make the world better. But it seems like that's just like a different set of issues than the issues normally being discussed in the ethics of abortion.
- [33:58]
B
Okay, so I wanted to ask you about economic growth and where that fits into this picture. I think it should be obvious to everyone that climate change really tugs on our obligations to the future. Like we want to burn fossil fuels now and that is quite good for us in the, in the short run if gas prices are soaring and if we suddenly stopped using fossil fuels, that would be. Continue to be like a, you know, medium term reality for us and it would make us all poorer. But that, that is balanced against a future where we are, you know, as the movie Tenet said, the oceans will rise and the rivers will run dry and at least to some extent, and that will, you know, force people to move, it will decimate, just slowly get rid of entire spaces of the habitable planet. And no doubt we'll adjust in some ways, but we will have superstorms and, and it will all be much, probably much worse for everyone 100 years from now than it is for us today. The superstorms will get worse, in other words. So there is some trade off between our and the interests of the future. But I think maybe that's the more obvious case. The less obvious case is something like economic growth. And I was very influenced by Tyler Cowen's book Stubborn Attachments, which he wrote several years ago. I'm curious, what do you think we owe to the future in terms of having a high rate of GDP today? Do we owe the future that they be much wealthier than we are?
- [35:26]
A
Great. Yeah. Lots to say on this. So the first point is kind of agreeing with you. Where most people don't appreciate, I think, the benefits of technological progress and economically the part of economic growth that comes from technological progress. And how, why do I have a good life today and we have good lives where I have anesthetic and I can travel and I have running water and clean air, that is because of technological progress and economic growth of the past. So that's like a systematic way in which I think we can benefit the future. However, I actually disagree with Tyler on this being a top priority for the reasons he gives, because it's got to run out at some point. There's just we're growing fairly fast by historical terms like 2, 3% per year globally. That simply cannot continue for thousands or tens of thousands of years. And to see this, just imagine, let's suppose the world does grow at 2% per year for the next 10,000 years. That would mean that we have 10 to the power, 87 times as much economic output as we have today, which would be well over a trillion times as much economic output of the world today per Atom, within 10,000 years. So we just must live a very unusually high period of economic growth that will, at some point in the future kind of plateau. And that means that from a very long term perspective, if you advance economic growth, you're not making a long term difference to the trajectory of civilization. You're kind of speeding up how quickly we get to the destination. Maybe, you know, you do some great invention and it brings economic growth forward by a year. Well then we've gotten to the destination a year faster. And the time period over which we've got to this destination point of technological plateauing, that's a little bit better. For that entire period after that point, you've not made a difference. And so that's where they kind of disagree with Tyler. I'll make one final comment though, which is that there's a difference, I think from a long term perspective between faster versus slower economic growth, which I think over the long run is not an enormous priority, and the issue of technological stagnation, where it's not merely that we go at 1% per year versus 2% per year, but perhaps we hit a point where we technologically stagnate even though we could have gone faster, even though we could have kept going, this could be much worse. I think from a long term perspective, where imagine if we technologically stagnated in the 1920s, then what would have happened? Well, we would have had the technology to use fossil fuels and burn fossil fuels, but we would not have had the technological ability to create clean energy. And so over the long run, it would have taken a long time, but we would have burned through all of the available fossil fuels and we would have created kind of a climate crisis. And similarly, I think we are, you know, we currently have nuclear weapons. We will soon be developing extremely powerful biological weapons. If we get stuck at that level of technological development, then we're stuck at kind of unsustainable level of tech development, such that if we stayed in that level, it would only be a matter of time before there was some enormous catastrophe. So you can think of us as like we're climbing the summit of a sheer face cliff face without any ropes and getting tired along the way. And any stopping along the way is explicit. It's unclear, like how fast you want to go up this phase. Maybe fast is better, maybe slower is better. If you stop and just stay there at a particular kind of level of tech progress, then at some point you will fall. And so that's why I think the best thing is ensuring that tech progress kind of keeps going and then exactly how fast we want to go. I'm actually not sure.
- [39:02]
B
That's very interesting. So I want to dig into that a little bit. So your argument here, if I understand it, is that. So Tyler essentially says the rate of gdp, you know, like a small difference in the GDP today can make a big difference tomorrow, right? The difference between a 1% and a 2% GDP may not matter much to you or I today, but in a hundred years you get exponential growth where it's 1% times 1% times 1%. And it's like the, the famous parable of the emperor who puts a grain of rice on the first square of the chessboard and doubles it on every. And then by the 64th square he has like more grains of rice than there are atoms in the universe or something like that. So his argument is basically that the GDP rate really matters today if you care about this long termist view that you and he would both share. And your rejoinder to that is? Well, actually growth has to stop at some point because if you keep doing the math, by the time we get to the 64th square, there is some end point we're going to reach where the world's economic output would be something. Basically it would hit some actual physical limit where we would have to be outputting wealth per atom. That can't be possible, or at least can't be something we understand today. So how quickly do we get to that sort of economic endpoint? I guess you could call it, you could call it like the economic end of history. How long do we take to get there at a current rate of growth?
- [40:32]
A
So I love the term economic end of history. On the question of where do we hit the limit? Very, very hard to know. Kind of at given rates in order to know that, I'd kind of have to know where's the frontier of tech progress, where's the limit? But I think it would have to be on the order of hundreds to thousands of years, which is actually a remarkably short interval when you think about any sort of either civilizational level or species level. Or certainly geological level time span where humans have been around for 300,000 years. The idea that we're living through a period that can last at most thousands of years more. Wow, that's pretty surprising. That shows we're at a very unusual time. And then when you think about the fact that through the course of life, I mean, life is in neurons, animals have been around for 600 million years, have at least several hundred million years to go, just given the habitability of the Earth, the fact that we're at this period of time that can only last for thousands of years. Wow, that's very special indeed. Yeah. So one way of characterizing the difference between Tyler and me on this perhaps is that Tyler's thinking medium term, but not truly long term.
- [41:36]
B
I see. So if we do reach this sort of economic end of history and we keep surviving long after that, then it matters less how quickly we get there, is essentially your point, because we're still going to have much more time and much more people after that endpoint. And so one of the upshots of Tyler's argument was that this basically takes growth, most growth reducing policies off the table unless they're crucial for some, some sort of other reason, like a, like a wealth tax or something like that. The general kind of Bernie Sanders type policy, which would redistribute but probably slow the economy slightly, slow growth. His argument was that that would be, you know, that would actually be bad for future persons. And so I'm curious what your attitude towards the trade off between redistribution and growth is.
- [42:28]
A
Yes. So I do think Tyler has a point where let's go back 200 years and imagine the government could either put some investment into tech progress, better steam engine or better electricity, which shifts the entire curve of kind of tech progress forwards just a little bit. It could either do that or it could do some sort of redistribution. So kind of giving arms to poor at the time, a surprising and unintuitive implication of taking future generations seriously is that, oh, actually no, the tech looks a lot better because you're not just benefiting the people alive today, you're benefiting many hundreds of years. Until we get to this point of you're making a benefit for that whole period. Until we get to this point of kind of technological plateauing. What does that imply for the world today? Well, I want to be extremely cautious about any kind of simple like, oh, go pro growth rather than pro redistribution or cautiousness about markets for the following reason, which is just that there are some things which don't merely have an impact until we get to this economic end of history, but would make an impact across the entirety of the future. So the clearest is risks to alone extinction. Risks of humanity's extinction, where humanity might live for hundreds of millions, billions or even trillions of years. And if we were to go extinct in the next few centuries, then that entire future is cut off. So that's an impact that would have genuinely indefinitely long effects. A second sort of fairly long term effect, I think can be influencing the values that society or humanity ultimately converges on. Where I give in what we are, the future and the work. I give these some arguments for thinking why actually the course of the very long term future could depend quite sensitively on the values that are promoted today. And again, that would be important for the entire future without this economic plateau. But are we using the fabulous wealth that we have in the year 3000, are we using that to promote free and flourishing lives for everyone? Or are we using it to construct pointless monuments, using slaves? Or are we actually indifferent to suffering and all power is just concentrated in the hands of a small number of people? There's lots of ways in which the future could go, depending on how those future people choose to use what will be enormous wealth that they'll have available to themselves. And so what does that mean in terms of policies today? Well, there's some things that are fairly clear, such as taking strong action to reduce the risk of extinction. We should be very worried about outbreak of third World War. We should really try to reduce the risk of pandemics. And then also carefully governing kind of technologies or political developments that could lead to the end of model progress, the kind of moral end of history, where in the book I talk about AI in particular as a risk here. And then you ask the question, okay, for the kind of classic left right arguments, like economic arguments, does that mean greater redistribution? Does that mean faster growth or something? And I'm like, man, it's all getting determined actually by the flow through effects of these things onto the risk of human extinction, the risk of AI takeover, the risk of entrenchment of values by political elites, or a perpetual dictatorship. And there it's like, okay, it's very non obvious what the conclusion is going to be at any given point in time.
- [45:50]
B
So when you talk about moral progress, you're talking about presumably the fact that our values today are massively different than the values of typical people 300, 400 years ago, 500 years ago, it was difficult to find a single writer that said slavery is bad, full Stop in all cases.
- [46:08]
A
Right.
- [46:09]
B
And today that is seen as so obvious that it's impossible, it's actually impossible for most people to imagine themselves into the position of our ancestors that didn't hold that belief. So when you say you fear the end of moral progress, what do you mean by that?
- [46:25]
A
So I think there's a few ways this could go. I mean, the most dramatic and kind of clearest would be if there's a group with a certain ideology that is then able, like a single country that is able to take over the world, implement that ideology globally, and then via technology, ensure that ideology persists forever. That would be the end of model progress, it would be the end of model change, and we wouldn't have model change, model improvement over time. That is an extreme example, but we've had, in relevant case senses, close calls. I mean, the Nazis and Stalin wanted to form a world government. They had this very strong ideology. They wanted to impose it as far as they could. The difference I think is future technology could enable much greater entrenchment of an ideology like that. In particular AI if you think about why do dictatorships fall? Well, one is coups or rebellions. The army just doesn't like what the state does and takes over. If your army is automated so it's AI agents that are doing the fighting, an AI police force, then that worry goes away. A second reason is just that political leaders die. But if again in the future people move from not being flesh and blood, but now having artificial beings, digital beings, in control, well, at that point in time, and this starts sounding kind of sci fi, but I think when we're thinking about longer term issues, we really should take this seriously because it affects what we should do now. But as soon as you got the point in time where people have put artificial beings in control, well, they don't need to die. Like an artificially intelligent being that was controlling the world could just live forever. Particular pieces of hardware would run out, but it could replicate itself. So that's a way in which you could really get stability forever. There are ways in which you could get it for like intermediate lengths of time as well. So many companies at the moment are working on what's called radical life extension, where you prevent the process of aging at the molecular level. And if that was successful, and maybe it could well be successful in a lifetime, then human lifespans would be in the thousands of years. And again, you could have political leadership in positions of power for much, much longer time spans. So that's the kind of more dramatic ways like third World War, conquest by the winning power, implementation of a non exploratory ideology. But I think honestly, there's just other ways too. You could imagine just gradual homogeneity of culture over time where this is already happening globally, the entire world is different countries, they're all moving more towards Western values at different paces, but that's happening. If you look at like as an example of how much homogeneity is there already in the book, I give this example of the COVID 19 response where, yes, there are almost 200 countries, but did a single one do human challenge trials where willing volunteers get infected with the virus in order to test fairly quickly whether a vaccine works? It's like, no, not a single one. Did any country make vaccines available in the flea market? And as soon as they were there before kind of testing, and it's like, no, not a single one. If either of those things had happened, then we would have ended the pandemic months earlier. But because almost all countries had broadly a similar response, we lost an enormous amount of value there. And so one other way in which we could see the end of model progress is just like kind of, you got to start off with different colours of paint and then they all mix together and you get this kind of muddy brown. And then after that you can't undo it again. Perhaps the entire world then just converges onto a particular culture, a particular set of moral views. And after that point it's like, well, why would we change now? These are the moral values we have. And especially if that culture is not morally exploratory, it's not trying to make moral progress, then it becomes a little hard to see how we get out of that. And if those moral views are wrong, well, that would be very bad indeed. Where to pump your kind of intuitions about this? Just imagine if the world of the future, you know, imagine if moral progress had stagnated with the Roman Empire, where slavery was just utterly accepted, where you would watch people get tortured in the Colosseum, which was, you know, extremely patriarchal. That would be an enormous loss of value, be a loss of almost all value that we could have achieved.
- [50:33]
B
How does your view about the homogenizing of world culture interact with the existence of China, which is like something one sixth of the world, and which to my eye, even in its pandemic response, its ability to lock down with no concern about individual liberties, because there's not a strong tradition of independent liberty against state action. You know, obviously there's the concentration camps filled with uyghur Muslims and other minorities and so forth. So how do you see a civilization like China interacting with the Westernization of most of the rest of the world?
- [51:12]
A
Yeah, I mean, so, yeah, employ like experimentation of cultures and so like having a world that has many distinct cultures trying out different forms of life and learning kind of empirically over time, like what is the best way, you know, structuring a society. That being a way of kind of getting to model progress. But I'm also a big believer in individual rights and liberty. So I think that sort of worldview kind of only makes sense if you could have free immigration between these different cultures and societies. So perhaps, yeah, you could have. In this world of like many, like that's a melting pot of different forms of, you know, societal structure, you could have some that have in general kind of very little liberty for the people that are part of that society. But only if you can opt out of the experiment kind of as a whole, only if you could move elsewhere. And so in the world as it is today, yeah, I'm kind of very much in favour of countries that have kind of liberal attitudes, especially with respect to free speech and discussion of important moral ideas and ability to challenge the status quo. Because if you don't have that, well, that's one pathway to losing out on moral progress.
- [52:20]
B
So moral pro. I think I have a bit of a different view on moral progress in that I can't, I guess, other than the AI scenarios you laid out, I can't see any circumstance in which. I can't see any circumstance in which people don't try to innovate on the morality of the past generation. Maybe that's myopic, but it seems like people, especially elites, will always have an incentive to come up with something new that goes counter to the prevailing norms of the time. And that will be a status incentive always of being the next, having the next idea, saying whatever the ideas that we were raised with are not the right ones here are something new. And so that will always, you know, whether that's progress or regress is another question. But it will never, I think, stagnate unless we get into those spooky AI scenarios.
- [53:15]
A
Great. So I think, in a sense I agree with you. It's just that unless we get into those AI scenarios played a really big role in my thought. Where at what point do we get AI that has human level or superhuman level abilities? There's a big debate on this issue. Many extremely smart and well informed people I know are really soon. 10 years, 20 years.
- [53:37]
B
Yeah.
- [53:37]
A
But Maybe you think it's very far away, hundreds of years. I think it's very unlikely that it's never, you know, I would go less than 10%, that it's never. And what I think the development of human level artificial intelligence, or greater than human level, where that means artificial beings that can act and speak and reason in the same way that human beings do and like take action in the world. They have a vision of the world, they have goals. That gives a kind of end point for history because the creation, that first generation of artificial beings are immortal, essentially the first creation of the first immortal beings and depending on exactly what happens, could well be the last generation for that reason. Because if I am an immortal, like for most beings, I think if you're an immortal and you're in power, you're not going to want to give that up. And if you're immortal, then you're not necessarily going to want to give that up ever. And so what I see is that that point in time kind of sets an end point. That's a point at which things really could ossify and the underlying causes of kind of model change, such as competition or just people dying over time or just like lack of understanding, the kind of causes of model change over time disappear. And we could have kind of model ossification. But if there's this deadline, then that means that things we do now, even if there's some amount of change, still have an impact kind of on that. So let's say that the point at which you get this kind of human level artificial intelligence, even say that was kind of 100 years time or 200 years time. Well, influence you have kind of politically or morally now would impact the values that are dominant or predominant, at least impact to some extent in kind of 200 years time at that crucial lock in moment. And that's the way in which the question of how much moral progress do we make before then becomes kind of really crucial. Whereas I agree with you that if we were certain that AGI would never come where AGI is general artificial intelligence, human level artificial intelligence, then I just wouldn't think of this as a key priority because I would imagine moral values just wash out over the long run. Because we're just thinking, yeah, tens of thousands, hundreds of thousands, millions of years, at some point I've pushed for some change, but just so much other change happens in the future that the action that I did no longer has an influence.
- [55:53]
B
Yeah. So I struggle to, I really struggle to visualize AGI scenarios and therefore I struggle to have any opinions or intuitions or even fears actually of them at all. You know, and I've gone down the rabbit hole of people debating about when and if we're going to get artificial general intelligence. And somehow I still actually can't visualize what such a world would look like. Is it like how such a world, whether the good scenario or the nightmare or dream scenario, how it would look different for the typical person waking up in bed to, you know, going to sleep at night. And you know, I think if I, if I really could get that visualization, I may have more of an intuition about what's all going on. But I'm curious, you know, when you envision these scenarios, can you really, are you really visualizing what it looks like in the worst case and best case scenario from like a moment to moment first person perspective, like, how is AGI changing my life? Obviously you can't possibly know what it would be like, but what's, can you like sketch a scenario to make it more concrete?
- [57:01]
A
Sure. So I think the thing that is correct about what you're saying is like we're imagining this technology that's extremely transformative. A lot will change. We should have bucket loads of uncertainty about how it's going to play out, an enormous amount of humility in terms of our ability to predict that. But yeah, we can talk about, you know, there are three main scenarios that I tend to think about when it comes to AGI. So one, like the good scenario, which I think is most likely, which is AI, starts to automate process of technological innovation and then, but we're able to like manage it fairly well, then what it looks like is just science speeded up a hundredfold over the course of our lifetimes. Right.
- [57:41]
B
But still within our control.
- [57:43]
A
Still within our control. But now that these companies, like perhaps everyone has a share in the capital stock, so suddenly you've just got these like zeros getting added to your bank account. And that's true for everyone in the world. And not only that, but suddenly you can just purchase things that are far, like things that you would have thought were hundreds of years away. So we're suddenly able to cure like all disease or pain or suffering. You're able to have the most wonderful experiences, VR, that's just like perfect reality and so on. That's the kind of like ideal case. And that's still obviously very sketchy. Then there's the case that people worry about of AI takeover, where the thought is you create human level artificial intelligence, it soon afterwards becomes super than Human level artificial intelligence. Soon you've got agents, beings that are just radically smarter than we are. So in just the same way that chimpanzees would look at humans, or maybe even ants would look at humans, we would be to these artificial intelligences. And in the Wellian case, they just see us as competition. We're just like a risk to them and deliberately disempower us. How would that look for the typical person waking up? Honestly, it might look just like a completely normal life. You're doing your thing, you've not even heard of anything. Because again, in these most dangerous scenarios, things happen very quickly. You're just kind of living your life. One day you get kind of sick and then you die. And that's just true for everyone in the world. And that's because this AI that's misaligned with human goals has set off a virus, many viruses that it created in order to wipe out the competition. That's one example of how things could go badly wrong. Then a third way is via human misuse as well, which I'm also extremely worried about, where a single country or even perhaps a company again develops human level artificial intelligence. Soon after moves to superhuman level. It's extremely powerful. It is able to create technological advancement that's just far, far faster reason about things much better than we might otherwise than humans can. And that single company or country then just threatens the rest of the world. It has technology that gives it military power that far exceeds all other countries combined or the world combined, and thereby takes power. And what again would that look like from the perspective of an individual? Probably not much for most. You know, in the run up, perhaps you've got innovations happening on the way it feels like, oh, wow, things are really picking up, like in terms of technological progress. But then at this moment where, oh, you actually really do have things going much, much faster, then probably you're just out of the loop, you're disempowered, and then, you know, some company or some group of people or maybe even an individual is then able to just conquer the rest of the world and suddenly you're living under their whim. So those are the kind of three scenarios, like I say it's the optimistic ones I think is most likely. But technology, I think this idea of being a dichotomy of people who are like, oh, the AI boosters and the AI doomsayers is just really wrong. I think we should think about AI like we think about fission. So before we had technology to split the atom, you could think like oh, wow. This could be really, really good. You could make nuclear submarines, nuclear power stations. This is going to be amazing. Or it could be really dangerous. We have nuclear, could make nuclear weapons. And if you had the foresight to really see that in advance, and perhaps if we hadn't been living during a war, maybe we would have been able to reap the benefits of nuclear power without the destructive capability. And I think that's how we should think about AI too.
- [61:03]
B
That's interesting. Yeah. And in reality, in retrospect, we've clearly had a mix of both. We've had nuclear power plants that are powered, substantially powered countries like France, which is really good because it's clean energy. And then we've had disasters. Yeah, that was actually very helpful for me. I hope it was for listeners too. So if my listeners want to become part of the Tenet team to prevent the future from hating us, what can they do? What resources are there for them to turn their caring more about the long term future into action?
- [61:36]
A
Right. Well, yeah, I've really decided in terms of getting up to speed with a worldview. I've really started to pack everything in as best as I can into what we are in the future this book and try to make it accessible and readable too. So reading that is a good way of getting to understand all relevant considerations in terms of what to do. I think there's kind of two paths, I think, and this is, I think, true for just anyone who wants to do good. One is if you're like, mainly I don't want to think about this. This is all kind of intimidating. I just want to continue my life with my personal projects and goals, but I still want to contribute. I think the best way to do that is via donations. And I set up an organization called Giving what We can that encourages people to give at least 10% of their income to the causes that they believe do the most good. And you can give to causes that positively impact the long term future. We have the Long Term Future Fund@eafunds.org so yeah, you can sign up at Giving what we can make a pledge to give 10% of your income. And that is if there's a single action that you take to make the world better that is making an enormous contribution. And you figure out where to give and then you can give and then you can focus on the rest of your life. Uh, if you want to go further, then I think the other most important decision you can make than what you use your money on is your career or how you work and obviously, this is a much harder question in that it's going to be a very personal question. What is the best fit for you? Like, what's the best way to contribute? But we have another organization I Co founded, 80,000 Hours, provides an enormous amount of advice where it has online advice, has a podcast, the 80,000 Hours podcast, and it also provides one on one career advice for people who want, like a more personal steer as well. And you can find that on 80,000 hours.org okay.
- [63:13]
B
Wilmacaskill, it's been a pleasure talking to you. Thanks for coming on my show.
- [63:16]
A
Thanks so much. Yeah, I've really enjoyed the conversation.
- [63:23]
B
If you appreciate the work I do, you can support me by subscribing directly to my website, ColemanHughes.org and sharing this episode with friends and family. As always, thank you for your support.