# 2023-04-26 - Foresight Institute - On Beauty, Knowledge, and Progress [Official](https://shows.acast.com/the-foresight-institute-podcast/episodes/6527e4f6d5cc0400128c85a3) [Apple Podcasts](https://podcasts.apple.com/us/podcast/existential-hope-podcast-david-deutsch-on-beauty-knowledge/id1573395849?i=1000611016811) Duration: 01:04:06 ## Transcript ### Allison Duettmann 00:00:00 - 00:02:44 Hi everyone, and welcome to Foresight's Existential Hope podcast. Uh, today we have a very, very special guest that I think has, yeah, become incredibly dear to the Foresight community. Uh, and it's none other than David Deutsch. I think we've really, you know, been like, yeah, trying to get you on for this, uh, onto this podcast for so long, because I think, you know, there's a few people that I think I just really associate with Existential Hope, one of them, you know, of course being Anders Sandberg, who we've also had on, uh, now, now previously, but I think really the, and like the North Star almost for Existential Hope as you, you wrote a few really fantastic, uh, books, including Fabric of Reality, which is now, uh, you know, an older book, but has aged certainly incredibly well. And that's really on a multiplicity of universes and like how, uh, kind of that theory combined with evolution, computation, and knowledge, uh, and quantum physics can really like explain a new worldview, and then you published Beginning of Infinity, which really was a big deal, uh, for people in this community, and, uh, you're really kind of like providing the antithesis to the kind of like doom, uh, to the more doom-ery, uh, meme, and you're really saying that, no, look, progress doesn't have to come to an end. In fact, we're really just at the beginning. Um, and there's a few kind of like really, um, pretty concrete ways in which we can, um, push progress forward. Uh, and also like a few more like abstract and I think really good, uh, memetic pieces in which, in how we can think really about progress. So this was a really, really great book and especially chapter nine of optimism has really just stuck with me. I think if anyone, you know, reads anything really like that, I think it gets the kernel of existential hope across chapter nine and beginning of infinity, um, and then obviously you haven't stopped there. I think another talk that is very dear to my heart is, um, why a flower is beautiful, uh, that is given it's on YouTube and, uh, it's a real treat. Um, and then finally, you're also the creator of one of my favorite child-rearing philosophies, I do not yet have kids, but when I do, they will be raised under Taking Children Seriously, which is your, the child-rearing philosophy that you and a few other really wonderful minds have, um, you know, have like, uh, have, have put forward. And we also have Chiara Marletto as a Foresight Institute fellow who, um, wrote a really wonderful book on the constructor theory, um, which you both are advancing. And so we're really excited to, uh, to have you on. So thanks a lot for coming online. Um, I know I said a lot about your, um, kind of like, uh, your contributions as looked at from a Foresight lens already, but if you would like to kind of like summarize, uh, your, uh, perspective on like how you got to where you are right now, um, and yeah, like your life path and then, but so that people can get a bit of an, uh, understanding of what makes you, you, that would be absolutely wonderful. ### David Deutsch 00:02:45 - 00:05:38 Um, I've never aimed for any kind of global effect that way. And some of the things that I, uh, have been interested in have been obviously related, some of them have turned out to be related to each other and some not, and I don't think one can, um, or should, um, direct one's, um, research or one's life for that matter towards a distant, all-encompassing goal. Because that means that, um, that if you're wrong, uh, you won't find out until you're dead. And, uh, after all, all problems are parochial and if they have universal consequences, that's a bonus. Uh, we can be on the lookout for universal consequences, just as we're on the lookout for all interesting consequences, but, um, the main thing is to solve the problems as they come up. So I'll just give you an example of that. You know, I was interested in quantum computers. I was interested in theory of computation more generally and, um, uh, interest in how that relates to thinking. And much later, decades later, uh, came the, uh, ideas of an AI apocalypse. Now it turns out that, uh, these other ideas that I had stemming from a completely different context make the AI apocalypse, um, look, um, what's the word I can use? I mean, they are absurd. Uh, if, uh, for a start, if one regards an AGI, uh, something with human-like intelligence, but running as a program on a computer, if one realizes that that obeys the same epistemological laws as humans do, then, uh, it doesn't make sense to apply different, um, laws of society to it. And especially it doesn't make sense to enslave it, um, namely causing AGI alignment by force or building it in into the hardware as it were, not that that would be possible, but the attempt to do that is an attempted enslavement. So, uh, that's not going to turn out well. And I wouldn't, I wouldn't have guessed when investigating the relevant ideas initially, that it would have any such consequences. ### Allison Duettmann 00:05:41 - 00:06:04 Um, and I mean, I saw you tweet about this a little while ago. Um, and how did, do you find your ideas having any like foothold or, you know, having any, uh, forthcoming in the AI community? I know that like, or like, um, what do you like people to do concrete specific things, uh, differently based on these observations? Like, is there any particular strand, you know, that you want to point people toward? ### David Deutsch 00:06:05 - 00:08:47 Um, well, there are various things involved here. Now I think that, that AI, uh, recently, for example, GPT and ChatGPT, uh, is a wonderful thing and, uh, can be very useful. Um, and it has nothing, whatever to do with AGI. In fact, uh, as I've written, it's, it's a more or less the exact opposite of AGI. Um, because it involves, um, honing the program to conform more and more precisely and in a shorter and shorter time to meeting a given criterion. Whereas an AGI, the difficulty and, uh, no one yet knows how to overcome this difficulty, the difficulty is to, is to write a program such that no possible, such that there is no possible idea, um, for which one can say it will never enter that state, it will never have that idea. Now people will immediately say, well, how do you know it won't get the idea to murder us? Well, that's, that's the thing. That's the, that's the, um, problem that has beset humankind since we have existed. And that's the problem that was solved with liberalism and the enlightenment. And now we know how to do it. We know how to, um, uh, bring people up in a society that makes it extremely unlikely that they will become enemies of civilization. We haven't got it, um, perfect yet, but, uh, we've got it, uh, working amazingly well from the perspective of history, from the perspective of history, that the fact that, that, um, we have so few wars, so little violence as Steven Pinker likes to point out is, is unprecedented. And, uh, it's not inevitable. It's, it, it, uh, it's not that this had to happen and it's not that it has to continue, it's just that we have the knowledge, both theoretical and institutional to, uh, keep it going as it has been for hundreds of years, shall we say. And if we continue improving it piecemeal and so on, as Carl Popper would have us do, then there is no known reason why it should stop, but it's not inevitable. It will all depend on what we choose to do. ### Allison Duettmann 00:08:48 - 00:09:06 So, um, if you were, I guess, like in a bit more concrete terms in the, let's say AI alignment communities, would you advocate for like a taking children seriously view for AI, like you're taking AI seriously view of actually bringing them up in a specific way or? ### David Deutsch 00:09:07 - 00:11:08 Yes. Uh, I mean, in general, the history of educational theory since the enlightenment has been one of increasing freedom, uh, for children and increasing, um, integration of the, of the, um, values of society in general with those of in educational practices and institutions. So, uh, that, that has come together and it, uh, educational institutions are kind of the last, uh, the last institutions of Western society to take on board liberalism and the enlightenment and, uh, things are taken for granted in schools and universities, which if translated to society at large would seem absurd, like, like valuing obedience and so on, and, uh, enforcing, um, ritual behaviors and that kind of thing, but this is today better than it has ever been. It is still improving. Uh, and I think that if, uh, AGI were invented tomorrow, it would indeed be the right thing to do to educate the newly programmed AI, the AI that is, uh, programmed AGI, uh, as closely as possible in the way that our society educates children. I mean, I think I know of improvements upon that, but I think it would be wrong to enforce, uh, my narrow view of, uh, how to do things on everybody. But for everybody to, uh, to conform to the standards of society at large. It's not impossible and to do it for an AGI is not impossible either. ### Allison Duettmann 00:11:11 - 00:11:24 So you would almost be arguing for, I guess, more freedom in the way that we educate AI, AI as compared to, um, compared to perhaps like what the general, uh, like canon in the AI safety community is. ### David Deutsch 00:11:25 - 00:13:27 I don't advocate this for AIs. For AIs, I'm happy for them to be enslaved and to, and to be forced to do whatever we want them to do as accurately as possible. In fact, there is a, uh, there is a whole field of making sure they do this. And so that self-driving cars don't run over people and that, that kind of thing. And that that's all fine. And the more accurately that is done, the better, but that is not how you get people to be members of a free society. That's that you have to do in some sense, the opposite. And we have, uh, we have learned slowly and painfully over the centuries. And to do some very counterintuitive things to make, and to make those, to entrench those as fundamental principles of the, of the legal system and of the financial system and so on and everything. Um, so that, so that we have, uh, policing by consent. Okay. You know, 500 years ago, nobody could possibly have understood what that phrase means, uh, government by the people. Nobody would have understood that either. They, if you said it to them, they would have imagined some monstrous system, which couldn't possibly have worked. So, um, but society evolved, uh, through conjecture and criticism and a cultural evolution to make these things work and to, for them to become second nature, um, to throw them away in regard to AGI is terribly dangerous. I mean, it is the very danger that the AGI alarmists are afraid of. And they want to do the opposite of what's necessary. ### Allison Duettmann 00:13:29 - 00:14:46 Yeah. We wrote a little bit about, you know, extending frameworks of voluntary cooperation towards artificial entities. And I think it would be interesting to actually see, um, you know, how those could look like in practice, so basically like many of the institutions that we currently use to cooperate through in a relatively consensual manner, um, compared to as you said, um, it's an interesting, I think, theoretical exercise to just like, you know, think about what those would look like in an AI context. Um, yeah, very cool. Uh, but obviously you're not, uh, you know, you don't only have thoughts on AI. You clearly have like, uh, an incredibly, uh, I think, you know, like, yeah, an incredible breadth of, um, of at least like being, being able to synthesize different fields and I think like, you know, finding, finding really, I think, um, sensible parallels between them. So, um, could you, for like a young, uh, outsider, like a young, talented, uh, person, like entering your space, uh, would you be able to give like a rough overview also like what it is that, you know, you're working on thinking about, um, so that, you know, they can maybe get up to speed a little, uh, and then quicker. And I know that as you actually said that you don't like giving advice, so this doesn't have to be advice. This is just from your individual standpoint. And how would you, you know, categorize your field? ### David Deutsch 00:14:47 - 00:17:38 Yeah. So I also said that giving advice is not a good relationship to have with somebody. Um, I think, um, entering, so getting up to speed is also a little bit misleading because although there is in all the interesting, in all the things I'm interested in, there is quite sophisticated knowledge, which if you, if you are indifferent to it, let me use careful words, if you're indifferent to it, you will waste your time or you are likely to waste your time or something like that. Um, but not being indifferent to it doesn't mean getting up to speed. There is no such thing as speed. Uh, I think a better metaphor is the one used by my old boss, John Wheeler. Uh, when he said that in physics, but he said in physics, but I think it's true of everything. He said in physics, every point is a growth point. So wherever you look, every, even if something is, is, um, is, uh, has been known for centuries or something has been just invented today, uh, either of those things can be a point of growth where somebody says, why should it be like that? What would happen if it wasn't like that? And then of course, most such conjectures are wrong, but they are the means by which progress is made. So I would, I would, if I were starting out now as indeed, I suppose I am, uh, everybody is, um, then I would, I would want to think about the interesting things and think about what might be wrong. What seems wrong? Uh, what does, what I don't get. And too many people think that if they find something they don't get, it must be because there's something wrong with them. That's not true. If you, if you find something that you don't get, there's almost certainly something wrong with something else. It's either with the people who've told you about it, or, you know, the authors of the books or the teachers of the courses or whatever, or there is something wrong with the actual material. And even if the material is literally true, they may be looking at it the wrong way and your perplexity may be, and in some sense must be, uh, the fact that you're looking at it in a way that wasn't intended and which has some potential for improving it. ### Allison Duettmann 00:17:41 - 00:18:32 I guess that is again, uh, you know, rather, um, I think Popperian, uh, which, uh, which is, which is a nice, I think, um, you know, way to even look at your own updating within, uh, within your fields. Um, all right. So as someone perhaps like, you know, uh, entering your field, you know, still, you know, they, they may want to know, you know, roughly, or like from your own perspective, like, have you realized any specific culture shifts that were like relatively instrumental in your life? That could either have been like, you know, to your academic career, um, where just like, you know, the general, um, kind of like canon within your field shifted or, uh, on a personal level, you know, when were things where you have significantly updated, for example, um, were there any specific moments, uh, that really got you, uh, got you to update your worldview? Was it relatively stable over time? ### David Deutsch 00:18:32 - 00:20:01 Uh, well, I think my worldview, um, has only been like largely shaken or shaped once, and that is when I got to understand Popper. Um, uh, but it has been course corrected, uh, several times. And I suppose the best known one of those is where, is when I, uh, decided to update, um, Turing's work on the universal computer in the universal Turing machine to include quantum mechanics. And that was after I had realized that Turing had made tacit assumptions in his analysis about physics. And these tacit assumptions were false. And what's more that these tacit assumptions were now being used in things like complexity theory to derive what they thought were mathematical theorems, but were in fact, consequences of the wrong theory of physics. So they, they got the wrong answers for it. I mean, I only realized that later, but it turned out that as a result of making, uh, classical assumptions, um, they got the wrong answers for things like, uh, what computational tasks are easy and what are difficult. ### Allison Duettmann 00:20:04 - 00:20:50 Yeah. And I think you were actually relatively successful at like, um, you know, going out there and, um, you know, at least at correcting that, uh, that error, like at least like, you know, providing, providing alternatives for that. So that's a, that's a great, I think, um, you know, embodiment of Popperian falsification, he co-founded the, or founded the philosophy department, uh, the LSE that I was in. And, uh, so it was like Popper up and down and my earliest philosophy education and certainly has, uh, yeah, have a, have a deep appreciation and nevertheless think that only gradually you come to understand the very, I think critical role that he actually plays in everyday lives. Um, and over time. So I think it's, uh, it's interesting. You understand some theoretically and then over time it really sinks in as you continue your lives. ### David Deutsch 00:20:50 - 00:21:43 That was very much the case for me too. I mean that when I, when I first got enthusiastic about Popper, uh, my, my impression of what Popper's theory was, was very wrong. I mean, I would now not regard myself at that time as being a Popperian at all, because I, I'd misunderstood most of the things. Uh, what I had understood though, um, just not to put myself down too much. What I had understood was that the conventional way of looking at epistemology and knowledge was just wrong, uh, completely wrong. Um, and what I didn't understand is just how accurately and powerfully, um, Popper superseded it. ### Allison Duettmann 00:21:45 - 00:22:08 And, you know, I mean that Popper often gets talked about also in context with Hayek as, you know, two really like proponents of the open society. And I wonder if you have any, you know, because I don't think, at least in my previous research, I've seen you very, very much talk about Hayek at all. So I wonder if you have any ideas, um, you know, if you were influenced by him at all, if it was mostly Popper from the scientific lens or if Hayek from this more societal perspective? ### David Deutsch 00:22:10 - 00:24:10 Yes, I think I've only ever read one book by Hayek, The Road to Serfdom. Uh, and it was all right. I mean, it didn't, I didn't find anything in there that, that I didn't kind of didn't already think, um, must be true, something like that. Um, uh, Hayek is, uh, basically a right winger. Um, so in regard to economics, I agree with him, uh, in regard to, uh, society at large, I don't always agree with him. Uh, and Popper likewise, I think, um, uh, I think he overlapped a lot with Hayek, but there were places where they disagreed and, uh, where they, where they disagreed, Popper was usually right, except that, uh, he was to his dying day. I think he was a leftist and Hayek was a rightist, but that, that only affects their ideas in terms of the color and tenor of their ideas, not so much particular policies, which I think, um, in Popper's case, he wasn't that interested in even, but Popper's and Hayek's meta, um, take on political philosophy were much closer than the political policies that they actually advocated and that that's much more important. It's much more important to get, to, uh, to get right how one thinks that, that errors should be corrected, how one thinks, what role one thinks that institutions should have and that kind of thing is much more important than the actual policies that those institutions adopt at any one time. Cause if they can be corrected, then you can hope that they will be corrected, but if they can't, then you can't. ### Allison Duettmann 00:24:12 - 00:25:08 Yeah. Yeah. I guess that to the extent that Hayek had very, or like more concrete ideas about how those that should shape or influence society, um, you know, he, he, he, he easiest to be corrected. Um, and okay, wonderful. Um, well that was just to kind of satisfy my own curiosity, but, um, I think, uh, you know, I, another question I had is like how those, you know, what if any relationship there is between taking children seriously and, you know, the, uh, you know, more, I guess, scientific work that you've done, you know, like what prompted you, you know, to, uh, to go out there and like, really, I think see this really wonderful moment that, I mean, obviously I think in hindsight, education is, it's just incredibly valuable. And that's like, you know, how we will shape the future on a, on a pretty personal sense, but was there any spark that I don't think that there is at present and perhaps there never can be such a ... ### David Deutsch 00:25:08 - 00:28:44 Thing as a science of education. I don't think education theory or even educational psychology has the potential to be a science, even in the future. So it's all philosophy. And, uh, for me, uh, taking children seriously is simply the application of Popperian epistemology and more broadly liberalism, uh, to the foundations of education. And it, it has in a way, so it's rather paradoxical because in a way that means it's not much of a change. Uh, since, since liberalism is the kind of dominant assumption in our society altogether and you know, you it's, it's completely normal to appeal to things like freedom of speech and individualism and so on in society at large. People may disagree with particular cases, but they won't say that's not a way to argue. You know, we, we, we don't care about individual choice or anything like that. So, but on the other hand, as I said, because of, well, this is something we haven't mentioned, but because of meme theory, because of the way that memes work, um, there is a, um, strong tendency for anti-rational memes to particularly manifest themselves in education. Uh, just like, um, if you can accept this analogy, it's just like in biology, the parts of our genome that are most resistant to change are the ones that determine the, uh, structure and function of ribosomes and generally of the DNA code. So the DNA code has been almost unchanged for 3 billion years. Um, it has undergone slight changes, you know, different, different species have slightly different ribosomes, but, uh, and, um, animals and bacteria have slightly different genetic code and so on, but it takes hundreds of millions of years for that to change. And that's because the selection pressure on this thing that is involved in replication is, is stronger than for anything else. So in regard to human ideas or memes, uh, that's the education system or the education, education practices. Now, this is not a council of despair. I mean, uh, memes are not genes and we are not victims of them and we can choose, we can always choose to behave differently and we can always use argument to decide instead of, instead of, uh, dark feelings that one gets when one, one does the unconventional thing. Uh, so we can, it's, it's just that it's, it's no accident, I think that, that education is, is the part of society that has been slowest in adopting the values of the enlightenment. ### Allison Duettmann 00:28:48 - 00:29:43 Okay. Really, really interesting. Um, thank you. Uh, I also had, I guess, like, um, question on the chapter on hope that you wrote, uh, in beginning of infinity. Um, I think it's, you know, like the chapter on optimism is I think one of the ones that, you know, just really brings the point home, um, uh, in a just really wonderful way, because I think one thing that you often get, um, like that suddenly, like, I think, you know, an existential hope lens on the world often gets. It's like, well, isn't this just Pollyanna-ish and, you know, you're like entirely ignoring the world. Uh, and it just almost seems like, you know, you're fighting like an uphill battle there by just making a claim that like, you know, there's there, there are good reasons for optimism. So I wonder if you, you know, could, uh, lay a few out here. Um, you know, obviously you can summarize the entire chapter and people should definitely go read it, uh, if they feel so inclined, but, you know, what are a few, uh, you know, good reasons for optimism? ### David Deutsch 00:29:43 - 00:33:53 So maybe the first thing to say is, is not good reasons for optimism, but almost like the one thing I have in common with the doomsayers, which is that I don't think anything is inevitable. Um, human improvement is not inevitable. It is, uh, always down to the choices that people make. And there is no limit. There's no, uh, naturally imposed God-given limit on the size of errors that we can make. We can mess it all up if we make the wrong choices. So, um, and as, and that conditions, um, that conditions, um, how one can become optimistic or how one can have a optimistic worldview, um, while being, um, while being able to combat the, um, objections that you mentioned that you run into. So it, uh, optimism is not what I call blind optimism. It's, it's not the theory that things will go right, even though they look as though they will go wrong. Uh, it, it just like blind pessimism is, is the idea that things will go wrong, even if they look good, uh, which also is quite a popular view. Um, it is that, um, because, um, what will happen depends on our choices. It depends on the knowledge we will choose to create and on the knowledge that we will not choose to create, uh, on the ignorance that we will not leave ourselves in because of that, um, there's no reason to give up on any problem. So problems are soluble. The problems are inevitable, as I have also said, to carve in stone and also to carve in stone that they are, that they are soluble and they are soluble by specific, not methods because there are no methods for problem solving, but specific types of process, uh, can lead to solving problems and specific types of process can inhibit the solving of problems. So, uh, conjecture and criticism and, uh, institutions of criticism and error correction and of consent, um, are necessary. They are the things that are most precious in maintaining our forward momentum in regard to ideas, because if they are impaired, it impairs everything. And, um, once everything is impaired, well, civilization has collapsed before. And, uh, I see no sign of our civilization collapsing, but as I said, uh, nothing, there's nothing, there's no, uh, supernatural force holding it up and enforcing continued progress. It will be up to us. And if, uh, everyone decides that progress is in fact bad, that progress is in fact an illusion, that progress is always at the expense of one group of people in favor of another, if that becomes a prevailing view, then progress will stop because nobody wants it. And, uh, once, uh, it stops, there's again, there's no reason why it should start up again, again, historically it, it stopped and it started up again. And in these smaller scale cases that I describe in the book, like Athens and Florence and so on, um, it didn't start up again. ### David Deutsch 00:33:53 - 00:34:31 It, it, uh, it was just, um, taken on board by the general enlightenment. Um, uh, but, uh, I don't know of any law of nature that says that the enlightenment had to happen. Um, I think we should be very grateful that it did happen and we should try to keep it going and we should try to improve it because it, it, uh, it still is very flawed. It always will be, always will be very flawed. We will never reach a non-flawed or almost non-flawed state. ### Allison Duettmann 00:34:34 - 00:35:00 But is it then, um, that you think that perhaps the biggest risk that we're facing right now in civilization is more like the kind of, uh, destruction of those institutions of, uh, you know, really conjecture, criticism, and consent that it took us so long to build, um, because we got, uh, maybe distracted by like, you know, some other things that we think are actually higher risk and that the solutions that we try to put forward are actually destroying the other bits that took us a long time to build. ### David Deutsch 00:35:01 - 00:36:27 I'm not convinced that either that risk or all the risks proposed by the doomsayers are in fact very great. I mean, because they're so important, it's worth taking them seriously, but I don't think the actual risk is very great in either sense. What we can say is, what I can say is that whenever our institutions are impaired by some fad or fantasy or bad idea that's, that's going around, it is bad. People are suffering as a result of every time, every time, uh, institutions and traditions of criticism and consent are impaired, people get hurt. Uh, people die of it. Um, this is from the point of view of a civilization as a whole. I don't think it's anywhere near that, uh, that level of harm, but you know, every child that gets dragged to school against his will is a, um, impairment of the growth of knowledge of civilization and who knows what has been destroyed thereby. ### Allison Duettmann 00:36:31 - 00:37:33 Yeah, that's, uh, that's beautifully said. Um, all right. Well, thanks, David. Uh, I will be handing it over to Beatrice, uh, for now, but it's a lot more and I think that you're hopeful with the questions, but yeah, I'll just, you know, definitely you have, I think you really have changed, uh, you know, the ways that like really like people in this community perceive the world in like really wonderful ways. And I think it really also shows in how people show up to each other and interact with each other, uh, in the way that I think oftentimes we are able to hold down, you know, critical conversations, uh, within plus that it'd be as well to be well. And I think, um, you know, if you don't get reminded of these, you know, reasons for why that's so important every once in a while, it's a bit hard of doing so. I really thanks a lot for, um, for being so, uh, well spoken and for, I think really living, um, in a really wonderful way. Um, the types of things that you believe in at least I, so thanks a lot. And I'll hand it over to Beatrice. ### David Deutsch 00:37:34 - 00:37:35 Good to hear. Thank you. ### Beatrice Erkers 00:37:37 - 00:38:22 Yeah. Um, thank you. Um, yeah, so I think, um, I'm going to ask you more about this sort of existential hope related questions, but I was also curious to hear just, um, there's this sort of idea that's talked about a lot now from like, I don't know, Toby Ord, the precipice, like, um, we're in this very crucial time in history where we sort of, um, what we do now has like an unprecedented, uh, opportunity of shaping what the future in the really long-term will look like, um, or like Holden Karnofsky writing about this being the most important century and we're facing these sort of unprecedented risks. Um, what's your take on this? ### David Deutsch 00:38:23 - 00:42:46 Well, I don't think so. Um, first of all, uh, and although nothing follows from this, but perhaps it's worth noting that pessimism throughout the centuries and also conservatism in the, in the bad sense of the word of opposition to progress, um, has always included the idea that we are facing an unusual moment of crisis in which the whole of everything we value is, is at stake. Um, and it has always been false. Um, uh, and I think it's, it's false today. I think the talking about existential risks, uh, obviously, you know, there is a risk that weapons we have available today could bring down civilization, though it's a bit far-fetched, uh, but nevermind. I mean, they could cause so much suffering that trying to avoid that requires as much effort and attention as avoiding the destruction of civilization altogether or our species. I mean, I don't think I make a, make a distinction there, but we have those weapons and the ancient Romans had enough weapons to do that when they destroyed Carthage, um, and the, uh, Catholic church had the weapons to do that when they exterminated the Cathars. And, you know, exterminations and destructions of civilizations have happened since the dawn of civilization. Weapons have been used in unprecedented ways since the invention of weapons. Um, if anything, I think the amount of knowledge that exists today, um, and knowledge is not so easy to destroy, that is explicit knowledge, the knowledge in institutions is relatively easy to destroy, unfortunately, but, uh, the explicit knowledge is, is so enormous today that it's hardly conceivable that, that a civilization brought to its knees could not rise again because they would just have to implement the existing knowledge, you know, they, they, they would, they wouldn't have to reinvent agriculture. They wouldn't even have to reinvent the tractor or fertilizer. They would just have to look in a book and it would tell them what to do. So, uh, on the other hand, so I think that the danger is not as it is painted. It's completely different. Um, uh, and on the other hand, the danger from nature, uh, is definitely less. So we've just seen in the last few weeks that, uh, a whole range of possible destructions of civilization from, um, meteor strike are not going to happen because technology has advanced to the point just recently has advanced to the point where that will not happen. There's a whole, there's still a whole class of possible impact from celestial objects that we do not yet know how to, uh, how to counteract, but a large class of them and the most probable ones we think we don't know that for sure, but we think, um, are now no longer a danger. So whereas there was a danger of, uh, of, um, uh, continental destruction size, um, impact every 250,000 years, I think it is, um, that is now gone. So, you know, one chance of death every 250,000 years multiplied by 8 billion people is quite a large risk per person per year. So, uh, it's, uh, manifestly existential risks are diminishing. ### Beatrice Erkers 00:42:46 - 00:43:30 Well, that's very nice to hear. Uh, also, um, you know, um, that's a message I haven't heard in a while. Um, yeah, because I think a big part of this, this whole existential hope project is that it's our experience that it seems really hard generally for people to envision positive futures for as these sort of dystopian futures are easy to see, um, but you've argued, you know, that like all problems are solvable. And even though problems are inevitable and some are really, really hard, it doesn't mean that they're unsolvable. Um, have you ever thought about, um, any like specific visions of the future that you think are desirable? Like, do you have a vision of existential hope for the future? ### David Deutsch 00:43:31 - 00:45:48 Uh, because of Popper, I think. I think that's a good question. Uh, because of Popper, I think I'm kind of constitutionally opposed to utopianism, both, both to utopianism as a philosophy. That is the idea that one should try to design a perfect society and work towards it and also utopianism in the idea of, uh, just imagining what perfection would look like, I would rather look for imperfections in what we have. Um, which are, as I said earlier, always parochial, even though they might lead to something, something universal, but the actual, uh, flaw is, is always parochial. And I'd rather look for those, you know, I'm, I have to restrain myself from being the guy who says something's wrong on the internet, you know, something's wrong on the internet, so I have to fix it. So I try not to do that. I try to look for things which are going to be interesting to fix rather than just something someone said wrong. Um, so, uh, I think in general terms, I would like the future to be one of ever more rapidly increasing knowledge, ever more rapidly decreasing suffering, but not just suffering in the airy fairy sense, specific suffering that we see, like people dying of, um, of plagues, people, people dying of pandemics, um, wars, uh, and so on that these things require a lot of thought. And there's no law of physics that says we can't solve them. Therefore we can solve them, but it requires creativity. So I envisage the future getting better in ways of involving conquering evils that we know about, but also getting better in ways that we can't possibly know, which will be, you know, wonderful. ### Beatrice Erkers 00:45:52 - 00:46:16 Yeah. I think I recall also you've written about how creativity is like an extremely important tool in gaining this knowledge that you think is like what we need more of, um, is we've spoken about like this, taking children seriously. Is there anything else, um, that we should do on a sort of societal level to like encourage more creativity and that would enable more knowledge? ### David Deutsch 00:46:18 - 00:50:24 Uh, yes. Uh, I mean, at the moment we are the Western, Western culture is suffering from, uh, a wave of fads whose general theme is to oppose, uh, Western culture, Western civilization to oppose the enlightenment, as I said earlier, to, uh, claim that it is a fake or that it never happened or that it did happen, but was bad and all that kind of thing. Uh, none of which is true. And all of it is based on, uh, factual misconceptions as well as philosophical errors, uh, but the phenomenon of this informing people's worldviews is, um, there are several such things which are sweeping, uh, Western civilization and all of them have the effect of, um, inhibiting progress by, um, inhibiting freedom. So restricting the, uh, the range of behaviors that are tolerated for humans, uh, restricting speech and communications. So there are certain more and more things are becoming taboo. Um, enforced, um, what can I call them? Um, reinterpretations of history, which again, it's not really a reinterpretation of history. It's just a relabeling of the phenomena of the enlightenment in pejorative terms. So, uh, so all those things are bad. All those things have got reactions against them, um, which I hope will eventually win or will be replaced by something even better and so on. Um, in this context, I should say that just like I have sometimes said, and people have criticized me for saying that in science, cranks are valuable. Um, if somebody, and even, even, uh, scientific publications ought to give some space to cranks, because it's not just that sometimes they are right, like J. S. Mill said, you know, sometimes they will be right, but even if they were never right, um, as J. S. Mill also said, um, you cannot understand the true theory without understanding why the cranks are wrong and not just one crank, but lots of cranks. And I think cranky moral and political theories are in the same category. The danger is unlike in science, the danger is that they get into power and suppress, um, progress towards true theories. That's different. Um, but, um, the cranks, the woke or the, uh, the extremists, uh, and so on, um, are also a source of creativity and, uh, sorry, they're a, they're a source of problems to think about and to apply creativity to. The danger, the danger is only that they get into power. That, that their ideas spread is not in itself dangerous. And our, our society is good at not letting dangerous people into power. They're not infallible. So, you know, let, let's bear that in mind. ### Beatrice Erkers 00:50:25 - 00:50:58 Yeah, no, thank you so much. There's, um, uh, there are two more questions I want to make sure I have time to ask. Um, which is one of them was, um, on Twitter today, you've got a question, um, about, um, how the, your, you, I think you mentioned that the idea of the universal constructor that, um, you mentioned in The Beginning of Infinity, you said that it's flawed. Is that something that you could maybe expand a bit on? ### David Deutsch 00:50:59 - 00:54:48 Yes. Well, I don't, it's not a very important point. It's, it's, it's mostly a matter of terminology. Um, in the, in The Beginning of Infinity, I said that I classified humans as universal constructors by which I meant that, that there isn't any fundamental limitation on what we can build or what transformations of physics, physical objects we can perform if we want to, um, other than the laws of physics, that they are limitations, but nothing else is. That's the point. Now, the thing is, uh, since then I have actually, uh, tried to develop constructor theory in general, and in particular the theory of the universal constructor, and it turns out that, that it is really essential in the theory of constructors, just like in the theory of computers to imagine objects that obey their program. So a constructor is, uh, first and foremost, it obeys its program. And then you can ask, what are the range of possible programs that it can be programmed with and what can it do as a result? A universal constructor is one that can be programmed to do anything that it's possible to do, to perform, uh, whatever transformation is physical, physically, um, uh, not, uh, violated, doesn't violate the laws of physics. Um, so therefore a universal constructor must be perfectly obedient and a human is almost by definition, like I said at the very beginning of this chat, um, cannot be obedient. Something which is creative cannot be obedient. So, um, the, that's a contradiction. Now you can say that a human body is an approximation to a constructor because although the mind can't be programmed, it has to sort of consent or at least acquiesce or then it might fight against what it's told to do and so on, unlike a constructor. But the body is more or less obeys the mind. Not perfectly, but, um, well enough to count as an approximate universal constructor. But, uh, there's also the fact that, uh, humans are very slow at some things and whether it is possible, we don't know how to make a real universal constructor yet, um, but supposing someone designed it tomorrow, it might be something like, might be something like a computer with a robot and, um, whether an individual person could build that computer and that robot in a lifetime out of ingredients that were naturally occurring. I don't know. It's doubtful. Um, so, uh, there are limitations on humans as universal constructors, but as I said, that's, that's really not very important. It's just a change in terminology from what I used in the book, uh, to a more convenient terminology. Uh, it doesn't mean that there's any limitation in scope of what humans can do. We don't start with naturally occurring things. And if I want to build something out of, uh, the built, uh, built, um, if I want to build a physical machine, I will not begin with digging for iron ore. ### David Deutsch 00:54:48 - 00:55:06 I will go to the hardware store or to Amazon and buy the things which are, which are close, close to what I want to make and just assemble them. ### Beatrice Erkers 00:55:07 - 00:55:46 Thank you. Yeah. No, I think it's still interesting just to hear you expand on it. Um, the second question that I really want to make sure I get to ask you is that like, one of the things that we try to do with this podcast is to like, really try to inspire more, um, positive visions of the future. And so, um, we always ask like for, um, an example of a eucatastrophe. So basically the opposite of a catastrophe. So an event where the expected value of the world is much higher after the event. Um, and so I was just wondering, could you maybe share, um, if you have a vision of what could be such a eucatastrophe, maybe it's the creation of the universal constructor or something like that. ### David Deutsch 00:55:46 - 00:58:41 I was about to guess that one. Um, I think, um, uh, I think it will be important. It, um, uh, it will mean that after the universal constructor is built, after the first one is built, and after all, it can build then more, uh, exponentially more. Uh, the human role in production will no longer ever involve toil. That is unpleasant physical work. Uh, toil will be completely ended by the, um, invention of the universal constructor. Although, you know, civilization in general has already reduced toil by something like 99% compared with what it was, uh, you know, when the human species first evolved. So this is nothing new, but I think it will be fairly dramatic. By the standards of everyday events. Um, and the role, instead of being to provide toil, the role of humans will be entirely to provide knowledge either for its own sake or to program the universal constructor, and there will be increasingly sophisticated aids to programming the universal constructor, just like ChatGPT can take a lot of the toil out of writing a program. Uh, all it really does, as I understand it, um, someone was explaining this to me, uh, is it, um, it takes the corpus of all programs that have been uploaded to the internet and constructs, uh, the one you've asked for, um, in the same way that it constructs good English sentences. Um, by the way, I was surprised at how good ChatGPT is at constructing, uh, sentences in proper English. I would have guessed that it will be decades before AI can do this. AGI, of course, could do it relatively easily, but I'm not sure that that's on the horizon. I hope it is, but as I said, as I have said, the people working on this have, have got the idea that an AGI is kind of just one more heave and AI will become an AGI, and I think that's the opposite. The opposite is the case. It's the AIs are getting further and further away from an AGI, notwithstanding their excellent English. Yeah. ### Beatrice Erkers 00:58:41 - 00:59:01 I saw on your blog, you had a bit of a, an argument almost with ChatGPT, um, about writing a poem. Um, but yeah, it got it right in the end. I think it does. It often gets it right in the end, uh, precisely when you have inserted in your ... ### David Deutsch 00:59:02 - 00:59:05 Angry objections, all the knowledge that it needs to get it right. ### Beatrice Erkers 00:59:07 - 00:59:20 Yeah. Well, it was, uh, it was a fun read and I can recommend it. Um, one, the last question I want to ask then, you mentioned Popper a lot throughout this conversation. And if one hasn't read anything by Popper, like where should one start? ### David Deutsch 00:59:21 - 01:02:04 Uh, I'm often asked this and I don't know. It really depends on where you're coming from. Popper was so broad in his subject matter, you know, political philosophy and philosophy of science and philosophy of knowledge and, um, that, uh, and within those, he, he addressed problems in different ways. I think, I think the concept that maybe unifies all of, um, Popper's thinking in all these subjects as, um, uh, Matjaž Leonardis, uh, recently, uh, pointed out to me is the concept of a problem. Um, a problem in science, a problem in philosophy, a problem in politics. The idea that, and this is also something that one of my chats with chat GPT was about, because it didn't know at first. And so I reminded it that, uh, according to Popper, the growth of knowledge begins, always begins with a problem. And I asked it, what does the growth of knowledge according to Popper always begin with? And it said, uh, a theory, uh, a criticism, you know, and I said, no, it's a problem, now start again, you know, and finally it did. It did give a quite a nice version of what Popper's take on this. Um, so, however, to answer your question, if somebody wants to approach Popper, if they've been persuaded by this, this chat here to start with Popper, to start on Popper, I would say, think about what problems you, you would like to have illuminated by a much, much better theory of knowledge than you have. Probably. Um, and that will guide you to which of Popper's books or articles or videos or, uh, will, uh, will, yeah, it will best make sense to you at first. Then later you can see the connections with other things. So I think I recently wrote on Twitter that, that, uh, sorry, what are we still, are we still running? Oh, I don't know. ### Beatrice Erkers 01:02:04 - 01:02:05 Continue. ### David Deutsch 01:02:06 - 01:03:19 Uh, okay. So there's, um, a lecture by Popper called something like, um, on the sources of, uh, knowledge and ignorance. I'm afraid I can't remember the name, but I, uh, every, every so often I go back to read that, um, lecture. It's, it's not very long. Um, and get something new out of it every time. I think it's the best discourse on epistemology ever written. Um, it, uh, it's incredibly deep and yet incredibly clear. And, uh, Brett Hall, the thing that prompted me to this is that Brett Hall, uh, had, uh, a series of five videos explaining this, this lecture by Popper and, uh, he ended up saying, I'm not sure anyone will want to spend five hours, uh, listening to my video. And I said, um, it's worth, it's worth it. Uh, but you can also read the original, which is nowhere near that long. ### Beatrice Erkers 01:03:22 - 01:04:03 Well, that's a great recommendation to go out on. And I think we can link the, um, the, uh, the talk in the, in the podcast when we post it, um, but yeah, I think I just want to like sort of echo what, uh, what Alison has already said that, you know, we're great, um, great admirers of you at foresight and we're very, um, yeah, very, very happy that you came on this podcast and I am looking forward to see what our AI generator, uh, image generator will make out of your, um, prompt for the universal constructor. Thank you so much everyone for coming and thank you, David. ### David Deutsch 01:04:03 - 01:04:04 Thanks for having me.