00:00:15.360 Yeah, it's great. We're recording now. Perfect.
00:00:22.080 Hello, everyone. Welcome to the Q&A session with Professor David Deutsch,
00:00:26.800 hosted by the Oxford Quantum Information Society. We are a student group at Oxford
00:00:32.960 organizing events about quantum information sciences and quantum computing. Actually, the first
00:00:37.760 such group founded a few years ago. Today, we have like from the society Nissot to Simone,
00:00:44.000 Amin, Jan, and Benji joining us. I think Maria is also here. She was the president two years ago
00:00:49.920 or three years ago and one of the founders of the society. I'm Matt. I'm a physics
00:00:55.200 field student at Oxford and the president of the society. I must mention that this session
00:01:02.560 is personally important for me. During my freshman year at Stanford, I was taking a
00:01:08.160 philosophy of physics class and I was exposed to a code by Professor Deutsch where quantum
00:01:14.000 computers were given as an argument to the multiverse. I learned that factorizing large
00:01:19.760 prime numbers efficiently could be done on quantum computers as shown by a computer short and to
00:01:26.160 code Professor Deutsch is large composite numbers. Yes, thanks. So to code Professor Deutsch
00:01:39.120 from the public of reality, so if the visible universe were the extent of physical reality,
00:01:45.360 physical reality would not even remotely contain the resources required to factorize such
00:01:49.680 a large number. Who did factorize it then? How and where was the computation performed?
00:01:57.520 After this non-intuitive and no description, I got extremely interested in quantum computing,
00:02:03.280 or at least that's what it seemed non-intuitive at the time. And now six years later, I'm
00:02:09.040 I'm doing a PhD on it. We're extremely lucky to have Professor Deutsch joining us today.
00:02:14.240 It would not be an exaggeration to say that Professor Deutsch is one of the pioneers of quantum
00:02:18.560 computing. He defined universal quantum computing and discovered the first form algorithm that
00:02:24.480 showed quantum computers can be efficient and more efficient than classical ones, demonstrating
00:02:29.760 that demonstrating what we now refer to as quantum advantage. And Professor Deutsch is currently
00:02:36.160 a visiting professor here at Oxford. Today we'll have the chance to ask him some questions.
00:02:42.000 We prepared some questions, including the ones you sent us while signing up,
00:02:45.920 but you can in fact, you're encouraged to ask questions as we go along. Just use the
00:02:52.160 hand raised feature on zoom when you have something to ask. And if you don't, like see you for
00:02:57.360 some reason, please feel free to shout out. We want this to be a free open for all this discussion.
00:03:04.800 Actually, we're aiming for this to be as interactive as possible, so it would be nice if you
00:03:08.000 could turn on your cameras. FYI, the meeting is being recorded, so if you might do some edits
00:03:16.320 before publishing, if you have some problems, just email us messages. And then the questions we
00:03:23.200 prepared mostly evolve around Professor Deutsch's book, The Fabric of Reality. We divided the
00:03:29.360 questions into the fourth trends of reality, discussed in the book. Computation, quantum physics,
00:03:38.800 evolution, and I don't know if you can see what epistemology. Okay, I guess we can get started
00:03:45.840 with computation and Professor Deutsch asked me to address him as David, so I'm going to do that.
00:03:53.040 David, you proved universal quantum computing under quantum physics, extending on
00:04:00.240 Alan Turing's universal computing principle. And now quantum computers are being built today.
00:04:06.240 What do they tell us about the multiverse interpretation of quantum physics?
00:04:11.760 I think the computers that are being built are from the point of view of telling us about
00:04:19.360 the multiverse interpretation, they are just a matter of psychological impact.
00:04:28.080 The real implication for the multiverse was already there when people first described single
00:04:38.880 particle, single photon interference. And then especially when people analysed entanglement
00:04:53.200 in those terms. And then especially when the theory of the universal quantum computer came along,
00:05:01.600 and then there was also my thought experiment which involved an artificial general intelligence
00:05:10.320 running on a quantum computer and performing an experiment to actually test the many
00:05:23.120 universal interpretation, or ever at every quantum mechanics as I prefer to call it.
00:05:27.840 And those were all theory. So, well, I suppose the single photon interference wasn't just
00:05:41.040 theory, but everything else was pure theory. And when we have a working universal quantum
00:05:48.080 computer, as a matter of logic, the case for every in quantum theory was already made and was
00:05:57.440 already unanswerable before. But it is just that much more cognitive dissonance for people to cope
00:06:17.520 Okay, I'm going to go on with our questions if no one is raising their hands.
00:06:22.560 And feel free to shout out again, if that's fine with you. We actually had a lot of questions
00:06:29.840 come up about your inspiration for the Deutsche Algerteman, the Deutsche Joe's algorithm.
00:06:36.160 Can you briefly talk about the algorithm itself and maybe your inspiration for it?
00:06:41.200 Yes, well, the first algorithm that I proposed in my first paper on universal quantum computer
00:06:55.520 was in fact not the one that's called the Deutsche algorithm today. It was much had much less power.
00:07:03.360 So, I've really usurped some bodies, somebody else invented the thing called the Deutsche
00:07:12.800 algorithm or probably several people simultaneously. But then the Deutsche Joe's algorithm
00:07:19.280 was invented basically by Joseph. He came to see me and said he wants to talk about quantum
00:07:28.880 algorithms and he showed me how you could execute something that in the end turned out to be
00:07:41.280 what's called the Deutsche Joe's algorithm. And when he was halfway through his explanation,
00:07:46.560 I said, wait a minute, that violates the theorem that I proved in my first paper that quantum
00:07:56.400 quantum parallelism can't increase the average speed of computing anything, but can only
00:08:09.360 increase the speed the fastest speed. And so he just casually said, oh yeah, that was wrong.
00:08:17.680 So, oh, my theorem was wrong, right? Good. And the reason it was wrong is that I had
00:08:28.000 what I called quantum parallelism in that first paper was a particular form of quantum parallelism,
00:08:34.480 which was the first one I thought of, and which I described in that paper. But in fact,
00:08:39.040 there are many other forms, and they are more powerful, and in fact, exponentially more powerful. So,
00:08:48.240 we ended up writing the paper about that, and that was the Deutsche Joe's algorithm.
00:08:55.840 It, one way it is conceptually an improvement on my original algorithms was that it involved
00:09:07.920 directly manipulating the phase. And manipulating the phase is obviously a more direct way of
00:09:16.000 getting to interference, which is the basis of the power of such algorithms. Some people say
00:09:24.240 it's entanglement, but it's the same thing really, because you adjust the phase somewhere,
00:09:31.120 and you also adjust it somewhere else in the Schrodinger picture, very misleading, but there you are,
00:09:35.680 that's the Schrodinger picture for you. And you mentioned entanglement, and we had a question
00:09:43.920 about that as well. How is entanglement phenomena explained without any signal passing between
00:09:49.920 two entangled pair of particles? But it does pass. The thing is, in order to test entanglement,
00:09:58.720 or to demonstrate any entanglement phenomena, you need to do experiments on one locality depending on
00:10:11.200 the results of experiments in another locality. And how did the information get from,
00:10:18.560 you could, so how did the information get from one to the other? Well, the answer is,
00:10:23.520 it gets there in the inside, the classical qubits that one sends from one side to the other
00:10:35.680 to determine what experiment to do. That shows that perfectly classical bits or bits,
00:10:47.760 which to all local experiments look perfectly classical, actually contain
00:10:56.880 quantum information inside them, which is stable to decoherence. Because it's stable to decoherence,
00:11:04.400 you can't measure it, but you can use it to convey the quantum information from one system to
00:11:11.600 another. And then, and only then, can you detect the entanglement phenomenon?
00:11:21.440 And I'm assuming that's a classical signal, or being the limits of speed of light.
00:11:28.960 Yes, although the calling it a classical signal kind of begs the question or makes the
00:11:39.520 explanation more obscure than it needs to be, the term classical signal or classical information
00:11:47.920 is ambiguous. It can either mean a signal expressed in classical bits, each of which can be zero
00:11:56.960 one, and that is not a physical thing. There is no such thing in nature, or it can mean a quantum
00:12:05.920 system, which all systems in reality are, in which the information is decoherent. But it can be shown,
00:12:17.760 and Hayden and I showed in a paper that such classical systems can contain hidden quantum
00:12:31.280 information. And by Hayden, I assume you meant Patrick Hayden? Yes.
00:12:44.720 I guess we moved into the world. I might have a very naive follow-up on that.
00:12:53.040 So when you talk about the classical bits having some hidden quantum information,
00:12:57.600 does that have anything to do with the hidden variable's theory?
00:13:01.760 No, no, it's the opposite, really. We called it locally inaccessible information. That is,
00:13:10.480 it's information that cannot be accessed by a local measurement on the bit that travels from A to B.
00:13:18.320 But it can be accessed by doing an experiment that involves both A and B one after the other.
00:13:27.600 So the information is there. It just can't be accessed by an experiment purely on the traveling bit.
00:13:42.880 I guess we moved into the realm of quantum physics. And I mean, I guess you were arguing that
00:13:52.160 the force trends are linked. So that is natural. Could you give an overview of the experiments
00:14:01.680 that provide the best evidence for parallel universes? In other words, what should an
00:14:07.040 experimentalist know about any worlds? This actually came from an experimentalist in the department,
00:14:12.880 so. Right. Well, it depends. So any experiment that provides evidence,
00:14:21.760 one of the other strands, namely the epistemology strand, where tells us that
00:14:29.040 evidence is meaningless unless there are at least two theories predicting different things
00:14:34.800 about the evidence. Otherwise, one can always reinterpret. If one only has one explanatory theory,
00:14:43.280 one can always reinterpret any result as conforming to that theory, if only by saying it was
00:14:49.600 an experimental error or a fraud or something we don't yet understand or or whatever.
00:14:55.440 So, which experiment is most convincing depends on which rival theory you have in mind as being
00:15:11.040 the one you would jump to if the quantum, if the ever at in quantum theory predictions were refuted.
00:15:18.560 So, the basic experiment, as I said before, is just any interference experiment.
00:15:31.680 And the more complex the variable that's involved in the interference, the more convincing
00:15:44.240 the experiment is. So, that's why people did experiments with interfering neutrons and then with
00:15:50.880 interfering molecules and interfering bucky balls and they keep trying to make viruses interfere
00:15:58.320 and so on. And eventually we'll have, as in my thought experiment, we will have interfering
00:16:04.640 AGI programs and that will rule out that then one can actually perform the
00:16:13.840 vignus friend paradox in real life and show that vigna was wrong and that what actually happens
00:16:21.360 is that the AGI can recall evidence of itself having had more than one experience simultaneously.
00:16:37.920 That I think would if it could be done eventually would be the ultimate answer but really it is
00:16:46.400 as I keep saying and hope nobody's offended by this, but I think it is it has been incoherent
00:16:54.720 to deny to insist that there is only one universe ever since 1957 when Everett wrote his
00:17:09.280 paper describing his so-called interpretation. It's not interpretation, it's a theory and the theory
00:17:22.960 Oh, I should say, so as I said, it depends on what alternative theory you have in mind.
00:17:29.920 If you have in mind the the bowman interpretation,
00:17:32.960 that is an interesting case because as I have often said, the bowman interpretation or the pilot
00:17:44.960 wave interpretation is the many world's interpretation in a state of chronic denial because the
00:17:56.480 wave function in that theory has got unoccupied grooves where the representative particle
00:18:05.920 doesn't travel and yet those grooves affect the particle. So some of those grooves are
00:18:13.840 human-shaped grooves in which there are thought-shaped events and those thoughts affect other
00:18:21.600 thoughts and so it's really just a matter of terminology to deny that those thoughts are the
00:18:29.040 thoughts of real people. If we could do such an experiment explicitly, we would have exactly
00:18:37.040 as much evidence of the people in the other universes as we have of people in our universe
00:18:45.120 and in both cases, that is more evidence than we have of the existence of dinosaurs in the past.
00:18:57.440 In both cases, we cannot experience them directly, but we can experience evidence that
00:19:04.320 it is inexplicable without postulating that they exist. I guess matching bowman mechanics
00:19:18.000 builds up to our next question. Oh, we have a question from Ma. Yeah, hello to an auditor to speak here.
00:19:26.800 I was wondering, so I read about the moral implications of the many world's theorem.
00:19:31.920 I was just wondering if you had any thoughts on that. Yes, I think the only moral implication of it
00:19:39.600 that's different from what morality was before and knew about it is that one ought to advocate it
00:19:51.600 that it is morally wrong to insist that it's false. Sorry, I meant more that
00:19:58.240 so do we have to stop worrying about all these other possibilities? Yes, so I'm saying that
00:20:05.040 we don't have to worry about those because it's interesting. I'm often asked that question,
00:20:12.960 but in two different forms. One form is saying, well, since you know that every time you drive out
00:20:18.720 in your car, you will necessarily run someone over in some universe. So shouldn't you be more reluctant
00:20:24.640 to drive a car? And the other is, well, whenever you run someone over, they always survive in some
00:20:31.840 universe. So maybe you should be a bit less than worry a bit less. So I think the quantum theory
00:20:38.560 of probability and decisions, which comes out of the Everett, Everettian quantum theory,
00:20:46.240 tells us that precisely that well, at least when we're doing decoherent things, so it wouldn't
00:20:55.360 necessarily hold during my thought experiment, but when we're doing decoherent things,
00:21:00.800 we should attend to probabilities exactly as if the outcomes were determined by probability
00:21:10.160 with the born rule. So it should not make us more reluctant or less reluctant to drive a car.
00:21:21.440 What happens when we are fully uploaded into quantum computers? And then I'm thinking
00:21:29.760 sorry, we're thinking coherent quantum coherent thoughts. No one knows yet, I think,
00:21:35.520 it's certainly I don't. So we've got a really interesting feature. Thank you.
00:21:42.480 Do you have a question from Bobak? Yes, thank you. And hi David. I was wondering how you would explain
00:21:53.360 the quantum xeno effects or paradox however we want to call it within the Everettian quantum theory
00:22:00.000 which. Yes, so do you mean the actual xeno effect or do you mean the Elizur vitamin
00:22:11.520 bomb phenomenon? Oh no, so I mean the so the effect being you know or the watch part effect.
00:22:24.400 where one is again it it's it's another experiment that confirms the Everettian quantum theory.
00:22:39.200 What one is doing is one makes a when makes a measurement which is overwhelmingly likely to have
00:22:45.840 the answer A and has a very low probability of having the answer B. Then one does it again,
00:22:52.000 then one does it again and so on. And classically the probability of getting a B would
00:22:59.200 would go up. I actually quantum mechanically as well, but you can you can arrange but because
00:23:11.040 quantum probabilities add up in a different way from classical ones, you can arrange so that
00:23:16.800 so that in the limit B will never happen or happen within with an arbitrarily small probability.
00:23:24.640 So that that is simply again I think that's better understood indeed
00:23:30.720 in Everettian quantum theory because you are simply putting yourself in a larger and larger
00:23:54.480 Hi, I'm I'm doing quantum some of I meant and someday someday I explain the
00:24:01.520 multiverse and quantum configuration speed up to my friends who are doing us and they take a very
00:24:10.560 different viewpoint of it and the question that consumes a lot. Did it decide that would it be
00:24:17.840 not be still resources from other universe to speed up the atlas and they have at least in
00:24:24.160 the universe. So this concerns me as if collaboration between universes to make the computation
00:24:31.200 faster or just a computation between different universes. And this perhaps correlates with your
00:24:40.720 conservation laws between the universe. I noticed that there is a conservation of knowledge
00:24:46.800 here. Oh, I saw this two questions you did but I'm not sure. Yeah, well, so I think this conservation
00:24:56.640 of knowledge and that kind of thing happens where there's not only quantum theory but close
00:25:01.120 time like lines as well. So just setting that aside for the moment, quantum computation
00:25:09.120 never involves stealing anything from other universes. It's always a collaboration,
00:25:14.000 but not only a collaboration, it's a collaboration between different versions of the same people
00:25:21.520 that started the computation. So somebody programs the computation and that programming happens
00:25:31.520 in a certain range of universes. Then as the quantum computation proceeds, that range of
00:25:42.640 of universes which were initially identical with everybody putting in the same program and wanting
00:25:50.160 the same answer wanting the same number factorized or whatever. And then as the computation goes
00:25:58.800 on, this range of identical universes differentiates itself into a vast vast number of different
00:26:08.480 universes all performing a different computation. And then they are recombined but not all with
00:26:16.240 the same phase. So in such a way that they all, well, in some you know in sorry, for some algorithms,
00:26:25.040 it's not all, it's just most. But let's say in a simple case, they all end up with the answer.
00:26:31.200 So there's no stealing going on. They have entirely, but not only are they collaborating,
00:26:42.880 I see you, thank you. Thanks, Ante, has a question.
00:26:59.200 Hi. Recently in your conversation with Robin Hanson, you said that the problem of a Y
00:27:07.040 risk and frequencies can be approximated by probabilities is an unsolved problem. And I was just
00:27:15.600 wondering why, how can the quantum theory doesn't solve that problem? Or why isn't it that
00:27:25.920 the probabilities are just frequencies in the multiverse? Yeah, very important question. It's because
00:27:32.960 when you, when you're betting on on roulette or something like that, which show we can be approximated
00:27:42.560 as a, as a quantum process, what you mean by the, by using probabilities is, is that you'll get a
00:27:50.640 different answer in different universes. But when you bet on let's say one company over another,
00:27:57.200 you're not, you're not betting that, that the company, like, suppose you have,
00:28:06.480 you can bet on two companies and you're, you're wondering which one of them is going to make
00:28:11.520 more profit than the other or give a higher dividend than the other or whatever. It's not that
00:28:16.800 you're expecting the universe is to divide equally, supposing its risk is 50, 50. You're saying
00:28:26.560 that you don't know how they're going to divide. In fact, you might be fully of the opinion
00:28:33.040 that the universe, universe is aren't going to divide very much with respect to which company
00:28:39.040 is better, that one of them is unequivocally better, but you do not know which it is.
00:28:46.160 And that ignorance is somehow translatable into a probability, but that
00:28:54.320 violates the theory of probability, the deriving knowledge, including knowledge of probabilities
00:29:02.000 from ignorance is a cardinal mistake in probability theory. So it's not that, it must be that
00:29:08.960 there is some way of dealing with knowledge, which gets you to a, say, a rational decision.
00:29:16.880 And then you can work backwards and say, well, because that's the rational decision,
00:29:21.520 the implicit probability must have been so and so. By the way, there are other cases where
00:29:32.000 using probability and using classical decision theory at all is a mistake. So this equivalence between
00:29:41.440 decision making in the, where decision making under real ignorance, decision making under
00:29:52.640 probability doesn't hold. And really, the only thing you can say is, I don't know. And you have to
00:30:02.640 resort to other objectives, such as minimizing your maximal loss or maximizing your minimum
00:30:11.360 gain or whatever, or just to keep out of it, if you can. So I don't know, as I say, it's an unsolved
00:30:20.000 problem. Did I understand your question correctly? I mean, you said, why isn't it the same? Well,
00:30:29.680 for probability to make sense, according to quantum theory, it must be because things are distributed
00:30:36.800 in a certain way in the multiverse. It cannot possibly work when it isn't distributed much
00:30:44.800 in the multiverse, but you just don't know. Yes, I think you understand it correctly.
00:30:55.280 Okay. We have another question, if there's no follow up on that, from Jan, who's part of our society.
00:31:11.200 Oh, you're in the room, I guess. Okay. Yeah, so we have a viewing party. I'm not in the same room
00:31:30.880 so that we don't have the quote. Yeah, I'll ask you. Thanks so much for taking time. So on the
00:31:37.520 topic of bermium mechanics, I was just into this. So if we, if you look at like modern versus bermium
00:31:44.240 mechanics, there's usually for bermions, there's these two important different frameworks. There's
00:31:50.240 a configuration space in which the wave function evolves, and there's also a physical space of
00:31:54.720 particles in which particles evolve. And then there's, for example, interpretations were
00:32:00.240 once as the evolution of the wave function is what guys the evolution of particles. And
00:32:05.920 there's not really, well, I guess my, my main question is, for a lot of bermions, this is the way
00:32:12.160 they think about the world that sort of, we have to model things as particles and therefore we
00:32:15.760 have to find out how particles evolve. And I was just wondering how you think about the world around
00:32:20.880 you in sort of the paradigm of, of many worlds in the multiverse. Do you think along the lines
00:32:26.320 of Wallace and sort of Daniel Dennis, a French-wist interpretation, and sort of what you have to say
00:32:34.720 to this sort of bermion interpretation, they have these two conflicting frameworks of,
00:32:41.600 or not conflicting two related frameworks of. Well, I'm not quite sure what you mean by
00:32:47.120 two conflicting frameworks, because I thought bermions, they are compatible, in fact,
00:32:56.960 necessarily complementary or whatever. But anyway, the thing you have to ask yourself,
00:33:04.400 if you are a bermion, is, is the pilot wave real? Is it a real feature of the universe,
00:33:14.160 or is it not? Is, is only the representative particle, or whatever you call it, is only that real?
00:33:21.600 If the representative particle is the only real thing, which I think is the psychological
00:33:26.000 motivation behind bermion mechanics, then you have to concede that it is driven along by
00:33:38.960 something that isn't real. It's driven along by pure mathematics or something. Whereas,
00:33:45.680 if you concede that therefore the wave function must be real, the pilot wave, or whatever you call it,
00:33:52.800 then you must concede that the pilot wave contains people shaped unoccupied grooves,
00:34:01.120 where the particle, where the particle is not, but where people are having conversations,
00:34:09.360 people in the unoccupied grooves, some of them believe bermion mechanics, some of them think
00:34:15.760 is rubbish. The full panoply of things that happen in the world represented by the particle
00:34:26.000 also happen in all the other grooves. And denying that they exist is logically the same
00:34:37.280 as denying that the dinosaurs existed, or denying that Australia exists, or denying that anyone
00:34:44.240 beyond your computer screen exists. Now, in contrast, I don't think there's much of a question
00:34:52.400 to answer in the case of ever Etienne quantum mechanics, because at least in decoherent ever Etienne
00:35:00.640 quantum mechanics, the ontology of the world is the same as it was classically. If we think of
00:35:09.440 coherent observers and that kind of thing, then that breaks down and we need a new conceptual
00:35:16.880 framework and new language and so on. But for everyday life, there's nothing new to understand.
00:35:27.760 Just today, I tweeted a lecture given by Harvey Brown,
00:35:33.840 which I think has said the last word on this issue in my view,
00:35:46.160 where he described some of the philosophical knots one has to pull oneself through
00:35:55.680 if one is to defend the bermion interpretation. As for functionalism and physicalism,
00:36:02.480 I think that's an unrelated issue. Those are issues of the metaphysics of thinking
00:36:11.680 and of what a person is and so on, which are questions orthogonal to, well, they are orthogonal
00:36:21.520 to quantum theory to ever Etienne quantum theory. Now, of course, if you make up a theory in which
00:36:29.920 the mind has a special place, then the two issues collide. But in ever Etienne quantum theory,
00:36:46.640 I think we have a follow-up question from Jamal.
00:36:52.480 Hello. Thank you for the opportunity of being in this question and answer session.
00:36:56.560 My question refers to the book, the science can and cannot that want to be a grad student,
00:37:04.720 so I created it. As when I was reading that book, when I was reading that book,
00:37:11.040 I came across how you were trying to use some instances of set theory and information theory to
00:37:19.120 try to give more breadth to constructed theory. And my question in regards to what the halting
00:37:30.400 problem in Goethels and completeness theorem also would be some of the things that would
00:37:36.640 also be added to constructed theory when you add set theory.
00:37:40.960 No, they wouldn't. At least we don't expect them to be added because
00:37:51.120 And constructed theory regards, although the universe may be infinite,
00:37:56.320 it regards physical systems as always finite. It only makes statements about finite systems.
00:38:04.880 They can be arbitrarily large, but still finite.
00:38:07.760 So, therefore, constructed theory does not make statements about the universe as a whole.
00:38:18.080 The universe is not what we call a substrate. It's not something that a constructor can act on.
00:38:23.520 So, you can think of the universe as being an emergent property of finite object, or many finite
00:38:31.600 objects. But you can't think of it as a first class object within constructed theory.
00:38:43.600 By the way, I should say that constructed theory is a work in progress.
00:38:50.080 I don't think, I mean, some people disagree with me, but I don't think it would be the right
00:38:55.600 thing to do to first make a rigorous formalism for constructed theory and then work out what it
00:39:08.160 means. Of course, it might be that's the way one has to go, but I don't think that's a good
00:39:16.400 idea if I'm possibly avoided. We need to work out what it means first and then think of a
00:39:23.440 mathematical framework that suitably expresses what it means.
00:39:28.960 We have a mathematical framework that's good enough to be going on with.
00:39:35.280 I guess, yeah, we will be talking about like creation of theories. So, maybe we should ask you
00:39:42.320 one final question on quantum physics. And as I said before, they're all interlinked, so
00:39:48.640 we can come back to it, but maybe let's talk about time a bit. In the fabric of reality,
00:39:56.960 you state that the flow of time is not real. It's an illusion. So, why can't it's not an illusion?
00:40:06.880 It's not real and it's not an illusion. It's a mistake. It's a mistake. Okay.
00:40:11.440 But then why can't we access past snapshots of the universe? Whereas it feels like we are
00:40:20.320 flowing towards the future snapshots. Right. What we can access is a question of which physical
00:40:30.480 processes are available under the laws of physics and which aren't. There are plenty of
00:40:37.040 physical processes like obtaining a sample of the center of the sun or seeing the center of
00:40:46.960 the sun or seeing a neutrino or whatever, which the laws of physics happen to not provide
00:40:54.240 the right kind of interactions for us to be able to harness those interactions to perceive
00:41:02.080 those processes. And the same is true of most of the past and the same is true of most of the
00:41:12.400 future, though they're interestingly different subsets of the past and the future that we can
00:41:19.040 know anything about. And I think that's a very powerful insights to be gained there.
00:41:28.160 But so we shouldn't be surprised when there's something we don't know. We shouldn't be surprised
00:41:33.440 when there's something that the laws of physics forbid us from experiencing what we should
00:41:42.560 be surprised if there's something that the laws of physics forbid us from understanding.
00:41:48.160 I think that would be a very supernatural kind of world to live in, and I don't think we live in it.
00:41:58.480 I see. I see. Okay. I guess to tie the multiverse to evolution, we have a question.
00:42:09.120 It's fine tuning necessary for intelligence. To explain, we found ourselves in these laws of
00:42:16.080 physics because we could only exist in these laws of physics. Would there be a way for the laws
00:42:22.080 to be different, but intelligence to exist? And by intelligence, we mean someone or something
00:42:28.160 who understands the laws that govern it. Should we call that life or are laws necessary for your
00:42:35.920 description of life? Oh, that last question seems to be at all. It's with all the rest of the
00:42:41.920 questions. Are laws necessary? Well, laws, I can only understand the term laws in the context of
00:42:50.400 understanding something. We understand something via laws, and if that's what you mean by laws,
00:42:56.480 then yes, they are necessary. But I think you may have meant our particular laws necessary,
00:43:03.600 like via the anthropic principle. I think the anthropic principle is kind of obviously true,
00:43:13.360 but what it tells us, I think, is very, very limited. I actually discussed this quite a bit in
00:43:22.960 my other book, the beginning of infinity, where I give an argument due to Dennis Sharma
00:43:29.840 that shows that if there are many different kinds of laws instantiated in reality,
00:43:44.240 and the only reason we see the ones that we do are that the only reason is that there's no one
00:43:53.760 in the other ones there to ask the question, then the most likely thing is that we are only just
00:44:03.680 in such a universe. I think this argument was also made in a slightly different way
00:44:09.920 by Richard Feynman, but I heard what I think is a devastatingly convincing form from Dennis Sharma.
00:44:18.080 And so if the universe is only just fine-tuned, which by this argument it must be,
00:44:31.760 then we should expect at any nanosecond to be wiped out because we would be surrounded by
00:44:39.520 a hot stuff that will kill us in the next nanosecond. And so what I've just said proves that
00:44:48.640 there isn't so in a certain sense with the alive. Do we have any follow-up on that? Yes, we do. Andrew,
00:45:01.840 Salto. Sorry, you're muted. Can you hear me now? I put the question in the
00:45:14.640 thing. What is wrong with the probability as a measure of the number of worlds in which one is to
00:45:19.040 find? Oh, well, this is an old, you know, truism in probability theory. Probabilities are only
00:45:31.760 equal to frequencies if the things in question are equally likely. So probability is frequency
00:45:43.600 over things that are equally likely. What does equally likely mean? It means they have the same
00:45:47.920 probability. What does probability mean? And so on. So that would lead to an infinite
00:45:53.440 regress or to circularity or whatever. And it's worse than that in the case of like multiverse
00:46:02.000 type reasoning because you can easily make a model of a multiverse that doesn't obey quantum mechanics
00:46:10.080 where the frequencies are obviously not probabilities. So somewhere I wrote an example of this where
00:46:20.720 you have, there are a lot of copies of you all in identical prisons cells and you are also identical,
00:46:30.480 you know, this is a non-continent thought experiment. And then half of you do one thing and half
00:46:39.200 of you do another. And then the second group, half of them does one thing and half of them does
00:46:44.000 another. And then by probability that the first group would have probability half, the second group
00:46:50.640 have probability one quarter, third group have probability one quarter. But you can achieve the
00:46:55.040 same physical effect by dividing into three, in which case they should all have probability one
00:47:02.160 third. So that's a world in which probability doesn't make sense. The fact that probability
00:47:09.120 doesn't make sense in our world is in my view an amazing property of quantum theory. We had to work
00:47:18.240 hard to prove that property. And if you changed quantum theory just slightly and no longer has
00:47:24.240 that property. Thank you very much. Thanks David. I guess we can still talk about evolution
00:47:37.920 to change topics a bit. A meme as described in your book is a theory or joke that survives through
00:47:46.320 iterations if I'm not mistaken. Well it's any idea that can be replicated. Yes. I guess it's a
00:47:56.560 container for knowledge. Yes. Yes. So if it can be replicated to more than random degree,
00:48:05.040 then it must contain, it can be said to contain knowledge. Now memes are a widespread internet
00:48:14.480 phenomenon. By the way, what I just said is and what you said is true of genes as well. It's true
00:48:22.640 of any replicators. You guessed the question. My down the question. Our memes are related to genes.
00:48:30.080 So memes and genes are both replicators. So in that sense the theory of replicators applies to
00:48:39.040 both of them. And there are illuminating properties of both that both share where each of them
00:48:50.800 can tell us something about the other. But memes and genes have a radically different replication
00:49:01.120 mechanism. Again, this is in my second book, the beginning of infinity, which you should all buy.
00:49:12.480 Genes, when they replicate, they are copied blindly. So the only bottleneck to their
00:49:23.120 to their exponential growth until they cover the entire universe, the only bottleneck is their actual
00:49:30.000 effect on their holders. Whether they increase or decrease the chance that that holder will
00:49:38.240 survive will survive in time to replicate the gene. Sorry. But a meme has got two bottlenecks.
00:49:51.920 First, it must pass the test of being understood and replicated into the mind of a new holder.
00:50:05.360 And then when the holder then executes it, it must do something which like the gene.
00:50:12.800 So that second bottleneck is the equivalent of the gene. The first bottleneck has no analog
00:50:19.120 in genes. And that makes the natural history of memes radically different from that of genes.
00:50:31.360 It means that these two kinds of memes that they have different replication strategies,
00:50:38.880 neither of them has the same reputation strategy as a gene. There is no way of
00:50:44.240 there is no way for a meme to replicate itself by blind copying. Humans are incapable of blind
00:50:54.320 copying anyway. But even if you try to engrave your meme on a stone obelisk, that obelisk
00:51:03.600 will wear away unless somebody tends to it. And that is not true of genes. They can survive
00:51:12.480 for billions of years and because they are maintained by a blind error correction process.
00:51:21.840 Whereas human memes have to be maintained by humans actually going out of their way
00:51:28.720 to maintain them in every generation. That's another difference.
00:51:34.000 Genes need not be enacted in every generation. If you don't, I have an example I gave you
00:51:42.400 if you don't break your arm. You will still pass the genes for repairing broken arms onto your
00:51:48.080 children. But if you don't enact your religion, then your children will never hear of it.
00:51:59.360 See, so that first bottleneck is a difference between
00:52:04.240 a live replication and non-living replication. Whereas a meme or your book is a container for
00:52:11.600 knowledge, it replicates as a printing house copy it and I acquired some knowledge to it.
00:52:18.320 But if without me, it would not be able to replicate the knowledge would go away or without the
00:52:23.600 readers. Without the re... And just printing it is blind replication. But really, that's not...
00:52:31.840 If we look at generations, then if humans don't read the book, then almost all printed material
00:52:40.000 will never be read again nor enacted. We have a question from Arjun.
00:52:48.240 Yes, it's a pleasure to be asking a question. So since we're on the concept of memes,
00:52:57.120 I think what you are in the beginning of infinity, which I have right here,
00:53:00.960 is in chapter 15, it's extremely profound when you come by with the idea of memes and
00:53:08.000 the fact and applications of society, and use right a little about and explicit
00:53:14.960 the in explicit component of memes and ideas. We should always be very interesting, but
00:53:20.560 please stand on that a little since... In fact, all ideas have some in explicit content,
00:53:27.520 since even in the knowledge of the meanings of the world, of words have largely in explicit
00:53:32.560 in our minds. But just to unpack that a little, maybe, I'd love to hear.
00:53:38.800 Yeah, well, so first of all, the way that it couldn't be is that all knowledge is explicit.
00:53:51.680 That couldn't happen, because if knowledge, all knowledge were explicit, then the meaning of words,
00:53:57.280 let's say I use a word like castle, and somebody doesn't know what a castle is. They would have
00:54:05.680 to look in a dictionary. And then in that dictionary, a castle will be defined as a type of building,
00:54:12.880 which has some following properties, etc. But in order to, if the person doesn't know what
00:54:19.200 building is or what a type is, then they have to look up a type. And it's obvious by accounting
00:54:26.480 arguments that the dictionary cannot possibly define, even in that straightforward sense,
00:54:34.800 cannot possibly define every word in it, because they would have to erute to every definition,
00:54:46.640 and that root would itself be undefined. So, how is it that in explicit knowledge gets around that?
00:54:55.360 Well, in explicit knowledge doesn't have definitions. And in fact, as Karl Popper says,
00:55:03.840 definitions are highly overrated. They're nearly always useless.
00:55:11.120 We have knowledge encoded in language, but not as definitions. They're encoded
00:55:20.160 as ideas expressed in language. And the way children learn language is by guessing
00:55:31.840 what words and grammatical structures mean. And when they guess wrong, they correct their guess,
00:55:41.360 and when they guess right, they use the word, and then perhaps correct it again.
00:55:48.960 Nobody, no true people end up with the same meanings of all the words. We loosely speaking,
00:55:54.640 we say that everybody who speaks English can understand each other. But that is not true. As soon
00:55:59.840 as you get into any kind of detailed discussion of anything, you find that people have a different
00:56:05.920 idea of what it means like to exist or what it means to be true. Or just basic things like that,
00:56:13.600 people have radically different intuitive ideas about which if they are naive, they think that
00:56:19.520 everyone else has them too and not just bring up twos. So, I don't know if that answers your question.
00:56:27.280 I mean, that's, I said like, what can't happen and what does happen.
00:56:32.320 Okay. We have another question. Also, how are we doing on time? David, do you think we could
00:56:43.200 go on a bit more? Yeah. Yeah, up to you. Okay, maybe like 30 more minutes, I guess, then.
00:56:51.120 In the shop. Maybe less. We have a question from Babak.
00:56:54.880 Hi, thanks. Hi again. So, we have physical theories of what we understand now as physical
00:57:04.560 theories of information starting with thermodynamics and second law of thermodynamics, especially.
00:57:11.760 But it, it seems to me that we are missing something there as a physical theory of knowledge
00:57:18.560 and how that might have a bearing on all of this. Do you have any thoughts on that to share?
00:57:27.600 Yeah, well, first of all, we already do have some quite deep knowledge about knowledge,
00:57:37.440 namely epistemology, which of which the best that I know of is called Popper's epistemology.
00:57:47.040 It obviously is insufficient, since if it were, if it could tell us everything about knowledge
00:57:53.920 and how it works. And so we would be able to use it to make an AGI. And I'm, I take it as
00:58:02.880 axiomatic, if you like, that until we can make an AGI, we, there is a huge gap in our understanding
00:58:12.560 of knowledge. Now, as a separate matter, I am kind of hoping that constructor theory
00:58:20.800 will tell us something about knowledge, since knowledge is an abstract constructor.
00:58:27.200 And that can't be an irrelevant fact. So, I think there will be a,
00:58:34.000 a constructed theory of knowledge eventually, which may or may not improve on
00:58:42.800 Popper's theory of knowledge. But even if it does improve on it,
00:58:48.080 I'm afraid I still think it won't be the full answer. Well, when I say the full answer,
00:58:53.040 I mean, it won't solve the problems that we have now. I mean, nothing is ever a full answer in
00:58:59.280 the sense of solving all future problems. But I think that the various issues about knowledge,
00:59:09.440 like the question of identity and the question of qualia and so on, knowledge, consciousness,
00:59:17.680 freewill and so on. Those, we have partial knowledge of those, especially knowledge of
00:59:25.200 what is a bad theory about them. But we don't have an actual explanatory theory of them.
00:59:37.360 And well, so we don't know that that's another, it's a whole other vast area that we don't know about.
00:59:46.080 I think it's a vast area. Maybe somebody will come up with a single equation that just solves it.
00:59:55.680 Wouldn't that be nice? Yeah. Cool. Cool. We have a question from Charles.
1:00:08.960 Yeah, my question is about the time and the multiverse. I guess I'm going back a bit.
1:00:17.760 I wonder how can the multiverse picture of time accommodate the idea of an open future?
1:00:25.920 If like in classical space time, determinism ends up fixing the multiverse.
1:00:30.240 And by open here, I refer either to people's choice or to knowledge creation, I guess.
1:00:39.600 So the, the open universe is, is not, is not intended to be open in the sense that there's
1:00:53.680 indeterminism. In fact, determinism does not help in any way to make the universe open in the sense
1:01:02.080 we want it to be open. In fact, if anything, it makes it worse. That it makes it worse.
1:01:10.320 We have to, the, the, the problems though arise at a different level from the level of microscopic
1:01:16.400 physics. They arise at the level of, of understanding knowledge again. When, when I say that
1:01:24.560 when I say that I have freely chosen to do something, I mean that I've created in, in the process
1:01:33.600 of deciding it. And I, I don't mean simply a choice like whether I should have strawberry or
1:01:39.120 butter scotch ice cream, where it doesn't really matter much to me. I'm talking about decisions
1:01:43.600 where I had to think about it and where I am a different person as it were after having thought of
1:01:49.440 it. In those cases, I have created something new. I have brought something new into the world,
1:01:59.040 namely the knowledge that I then use to guide my, my choice. So the universe can be open to that
1:02:12.000 or not. That's, that's the sense in which Papa used the word open in the sense the open society
1:02:17.520 and its enemies. No society is totally closed. No society that we know of is totally open,
1:02:25.040 but still there is a vast and important difference between societies that are basically open or
1:02:32.480 which at least strive to be and societies which are basically closed or at least strive to be.
1:02:38.480 And that's the sense in which it is important for a society to be open. And I think exactly the
1:02:46.720 same is true within one person. That's, that's what we mean by being open to ideas or it's what
1:02:56.640 we mean by having had a new idea. You know, I could say I had a new idea yesterday about so
1:03:02.640 and somebody, somebody, you know, a rapid determinist would say, no, that will idea was already there
1:03:10.080 at the big bank. But in a sense, it was and in a sense, it clearly wasn't.
1:03:20.720 So we have to basically in order to think of openness, we have to kind of put back in the drawer
1:03:29.760 all the framework of laws of motion, initial state, determinism. Otherwise, it might be just
1:03:38.640 meaningless and we're just going to loop around. Yes, because we shouldn't be surprised that
1:03:44.400 the answer to the problem is at the same level as the problem. The problem isn't about electrons.
1:03:52.960 Exactly. Yeah. If you have a, you know, if you're worried about how your theory of knowledge
1:04:02.560 is compatible with your theory of electrons, then that's kind of, I don't know what you call
1:04:08.480 a category error or something. That that's, that's, it's, it's like, it's like saying,
1:04:17.040 well, the example at one of the examples I give in, you know, efficiency is like, if somebody
1:04:23.360 said, if you're watching a conjuring trick in, in the theater and the counter it does something
1:04:28.400 that looks impossible. And you say to your friend, how the hell did he do that? And the friend says,
1:04:35.200 laws of physics, he hasn't, he hasn't answered your question. And if he says, oh, the
1:04:40.960 country did it still hasn't answered answered your question. You want, the problem was at a different
1:04:47.200 level from those two levels. It's at one level up. The answer is something like, when it looked
1:04:55.840 as though he was putting the ball in the second cup, he wasn't and he then put it in the third cup.
1:05:01.520 That's the beginning of a real answer. And it's compatible with the first two answers,
1:05:08.320 but the first two answers don't answer the question.
1:05:19.920 Hey, thanks for the question Charles. We have two more questions. So I guess we moved to the,
1:05:26.560 the strand of epistemology now. But let's see where these questions come from. Toby.
1:05:39.120 Hi, Toby. You have a, hello. Can you hear me? Yes. Yes. Hello there David. I have a question
1:05:50.160 about decision theory and the hour of time. And this could be a bad conjecture. I want to know
1:05:56.480 whether observation of time asymmetric increase of entropy could feasibly be a result
1:06:01.920 of agents maximizing their rational expectation values in a time symmetric world,
1:06:07.680 much the way you derive probabilistic bond rule from non probabilistic many worlds.
1:06:12.160 So that agents such as ourselves are playing like a game as if we can make sense of the world
1:06:17.760 in terms of causation change in particular in the future. But where those things have no intrinsic
1:06:23.120 or where by that time symmetry has no intrinsic existence of its own. So I was wondering if there's
1:06:29.680 been any work in this area or or whatever, this is a bad idea to think about. There have been
1:06:37.440 people who have tried to link what's he called the the entropic hour of time with the knowledge
1:06:46.880 hour of time. I'm I'm not very persuaded by that. I think there's a lot more to knowledge in
1:07:01.440 general and to the knowledge hour of time in particular than just physics. Yeah. As I said before,
1:07:08.800 construct a theory may provide a closer link, but it doesn't it hasn't yet.
1:07:19.200 Yeah. I mean, I mean, the sort of conjecture that I had was I think one of my criticisms of it was
1:07:30.400 that it would mean if it was true causation would be it would mount to some kind of I think it
1:07:38.000 would mount to some form of subjectivity because it the I can imagine random agents saying the
1:07:47.760 multiverse. They're determined by laws of physics and stuff like that. But the actual arrangement
1:07:55.360 in time would be a fiction or a story that would be correlated between sets of agents. But that
1:08:03.680 would mean that the kind of causal structure that I tend to think would be objective would have
1:08:08.400 to be subjective. So I thought that that was probably one major problem with the idea.
1:08:16.320 I'm not sure I understand quite what the idea you're putting forward is, but it kept sounding to
1:08:22.800 me that you were mixing levels of explanation again. The question of what can communicate with what
1:08:32.160 is a question of physics. It has an answer in terms of laws of physics and no higher level
1:08:40.400 explanation can add anything to that. Yeah, right. If on the other hand, you're talking about like
1:08:49.680 what things mean, like when you realize that Napoleon isn't what you thought he was,
1:08:58.320 then at a level of physics, that doesn't change Napoleon. But it does, it may change how Napoleon
1:09:11.360 affects the future from now on. Yeah, yeah, thank you. That's a, I think about
1:09:22.160 separating levels of explanation a bit harder. Thank you. Yeah, we had a few more questions,
1:09:32.400 pop up. Owen. Hi, yeah, thank you. So earlier when we're talking about means and genes,
1:09:40.320 you called them containers of knowledge. And I thought that's on a little bit strange to me because
1:09:46.320 I would consider them more containers of information rather than knowledge. Maybe I'm just mistaken
1:09:51.840 there, but no, no, that's a very important distinction. Information, so knowledge is a species of
1:09:58.480 information, but very little information is knowledge because the vast majority of the information
1:10:07.520 that exists in the universe does not have causal power. It doesn't have the power to replicate
1:10:15.280 itself, that is basically, but it also doesn't have the power to systematically cause transformations
1:10:21.520 in other systems. So, you know, you could take, you could take a census of all the grains of sand on
1:10:31.040 all the coastlines in the world. And that would be a vast amount of information. And almost none
1:10:36.800 of that information is involved in the explanation of anything else. So, whereas a much higher
1:10:46.560 proportion of, let's say, the information in the books in the bodily library, still, you know,
1:10:54.000 not a hundred percent, probably not ten percent, but vastly more than in the grains of sand,
1:11:01.440 have the power to cause things to happen. Historians can go in there, convinced of one thing,
1:11:08.640 come out, convince of another thing, and then go and do things in the world that have
1:11:15.440 much greater mass and momentum and generally causal effect than the grains of sand.
1:11:28.800 Yeah, okay, I see. Okay, yeah, and also earlier you were saying time is a mistake,
1:11:34.000 and I'm not sure I understood exactly what you mean by that. When we think that time flows
1:11:42.240 like a river or, you know, whatever metaphors people say, that's obviously not true. You know,
1:11:49.680 the time at this moment today is not going to flow anywhere, it's not going to flow to tomorrow,
1:11:59.680 it's not going to flow to the day after. Today, we'll always stay at whatever today's date is,
1:12:06.480 and that doesn't change, and nothing about the sequence of time changes at all.
1:12:14.400 Again, causation happens from one time to another, but the time itself doesn't do that,
1:12:28.480 Okay, but we still need the time as sort of a substrate to understand what
1:12:32.720 causation even means, like we still need the concept. Well, we need the concept of causation,
1:12:39.760 but I don't think we need the concept of the time flows. I think- Okay, yeah, I see.
1:12:52.800 Yes, hi. I had not sure is a very serious question, but you said earlier that
1:13:03.440 memes replicate by using basically two methods, and one of them is that you need to annex
1:13:11.040 the meme during some's lifetime for them to remember it, and then they can pass it on.
1:13:16.240 What is an AGI reborn with like pre-programmed knowledge of say the Bible or something,
1:13:21.040 or of certain religious traditions, so that you could bypass that without in any way change
1:13:28.640 meme evolution? I don't think it would. By the way, I'm not sure which, so
1:13:36.800 memes have two bottlenecks to pass through, unlike teens.
1:13:41.760 There's also, in meme replication, there are two different kinds.
1:13:47.360 So, I think you went first with those things. Yeah, yeah, I think you said, if you don't
1:13:54.240 annex your religion to your child, then they won't know about it, but if you break,
1:14:00.160 if you'd never break your leg, then they will still have- The best you can do, I think,
1:14:04.880 is to engrave something deeply on a diamond, but then you've got to put the diamond somewhere where
1:14:13.200 you can ensure that somebody won't grind it up and use it for industrial purposes,
1:14:18.240 or eventually so that it won't burn or evaporate or something like that.
1:14:25.040 If you put it in a robot like you said, and the robot let's say is immortal,
1:14:31.840 then that won't work because the robot, let's say you put the Bible or whatever,
1:14:40.240 the robot will have a different interpretation of the Bible next year from what it had this year.
1:14:47.440 It's presumed, I assume, by hypothesis that it's a thinking being.
1:14:52.880 It's opinion of what things mean will change, that's what humans do, that's what any person does.
1:15:01.200 If it's not a person, if it's just like the diamond, then it will eventually be ground up.
1:15:08.240 So it will only be preserved if it's useful for something or if somebody has a theory that it
1:15:17.280 ought to be preserved. And in either case, that's a property of the information that it has in it,
1:15:24.480 which makes that information into knowledge. But if it's just information,
1:15:29.760 or if it's useless information, I mean, then it will eventually fade away.
1:15:38.880 I suppose there's an exponentially small chance for any particular piece of information of
1:15:45.360 useless information that it will eventually survive. I mean, no doubt, somewhere on one of those
1:15:50.480 beaches around the world, there's a configuration of sand that will still be exactly the same
1:15:57.200 after the tide has come in and gone out again.
1:16:03.520 Thanks for your answer. Thanks for the questions, Sam. I guess let's take one final question
1:16:10.480 from Ante and then bring the conversation more to how to describe our more, more comprehensible,
1:16:17.040 immediately accessible world. Ante, yeah, go ahead. Thanks. My question is,
1:16:26.160 in the constructor theory, the universal constructor is there a meaningful distinction between
1:16:37.840 let's say something like DNA that can make a person and then it's automatically a universal
1:16:44.720 constructor, or does the effect have to be in some way or immediate?
1:16:52.400 So I have to apologize to everybody that in my book, the beginning of infinity,
1:17:01.120 I talk in a very loose way about the universal constructor, which turns out to be
1:17:08.720 not very suited to constructor theory. In particular, I spoke of a human as being a universal
1:17:16.960 constructor, and that's not quite right in an important way. Namely, that a universal
1:17:23.520 constructor, like a universal computer, has to be programmable. So it's universal in the sense
1:17:31.200 that for everything that can be constructed, there exists a program that will make the universal
1:17:37.760 constructor constructed, which means that it has to be perfectly obedient, like all
1:17:45.360 constructors, it's an idealization. But to say that something's a good approximation to a
1:17:51.280 universal constructor, you mean that it's very, very, very obedient. But humans, neither
1:17:57.440 humans individually nor human society as a whole, is at all obedient. So now you might
1:18:07.760 stay, well, it's not the human that's a universal constructor. It's just the human body,
1:18:13.920 and forget about the brain. You have to program the brain, instructs the rest of the body. But
1:18:20.160 then the rest of the body has a finite lifetime. And anyway, most constructions involve
1:18:32.960 cooperation between different people, and it involves machinery. And one day, when we have
1:18:40.320 universal constructor robot type things, it will involve those. And then it doesn't really make
1:18:48.320 sense to consider the human body as a universal constructor, when all it is, is a front end for the
1:18:58.240 real universal constructor, which will be a type of robot. Okay, so having made that apology,
1:19:05.280 you have forgotten what your actual question was. Well, if DNA can build a human, then is it
1:19:15.120 automatically universal constructed? Yes, well, so since a human isn't a universal instructor,
1:19:23.120 the answer is no. So if there were a way for DNA to not only build a human, but to install
1:19:31.680 in the human brain, a program for doing an arbitrary thing, and for passing that on to its offspring,
1:19:42.880 so that they would work for generations and generations to build rockets to go to the moons of Mars
1:19:50.320 and convert them into busts of Napoleon. Unless there's a way for DNA to do that, then the answer is no.
1:19:57.200 I don't think there can be a way of doing that, but that's another issue. I mean, that's not a
1:20:04.720 fundamental issue. It could be that DNA can do this without going via the intermediary of a human.
1:20:16.640 DNA could build an animal robot that builds other robots that builds other robots and so on,
1:20:24.320 and then the DNA installed in a certain cell, because something has to make the DNA
1:20:34.480 work, like with the ribosomes and so on. It could be that the DNA can indeed make anything
1:20:41.120 when installed in a suitable cell. Again, I don't think so, but it's not impossible
1:20:56.160 Thanks for the question. I guess we can bring the conversation towards more
1:21:03.520 practical questions. If I may say more accessible in our everyday world with our hands,
1:21:13.520 which like as quantum computers right now, we have there being built that we can touch and
1:21:20.400 play with. What is your opinion on current progress on quantum technologies and the industry in general?
1:21:28.080 Unfortunately, I haven't got my finger on the pulse of current technology.
1:21:34.720 My view of current technology is just that of a layman who sees things on Twitter from time to time.
1:21:40.880 So, I'm naturally skeptical about claims to have achieved universality,
1:21:51.520 and I'm naturally gullible about things to claims to have even universality in the sense of
1:21:58.880 universal quantum computer. And I'm naturally gullible when it comes to claims about having achieved
1:22:05.840 something amazing, like a new functionality. Just like in the case of AI and AGI,
1:22:14.240 I'm skeptical about claims to have made progress towards AGI, and I'm fully expecting amazing
1:22:26.080 Okay, but was there maybe a technology could development that you were surprised to see,
1:22:36.400 positively surprised to see in the last few years, or what are some scientific trends that
1:22:41.920 you were excited about? You are excited about right now.
1:22:46.000 Oh, well, that's a much wider question. Do you mean in regard to quantum technology?
1:22:56.560 Yes. Again, I'm afraid I don't know. Maybe generally, actually, maybe not even,
1:23:03.520 maybe not immediately related, but related.
1:23:07.200 So, I think the big picture is that I think that progress in fundamental progress in physics
1:23:17.200 has slowed down during the last few decades, and it has slowed down not because
1:23:26.000 we've reached all the low-hanging fruit or for any reason like that. There's a sociological
1:23:31.680 reason that has stultified the scientific world in regard to almost all fundamental issues.
1:23:47.440 So, most things haven't made progress. There has been a lot of progress in cosmology,
1:23:54.800 which again, I don't know much about. I think that the inflationary model
1:24:05.680 seems like a fairy story to me, and if it turns out to be true, it will be amazing.
1:24:13.840 And I mean, when I say it turns out to be true, I mean, if the underlying mechanisms
1:24:19.040 are found to have some basis in a new series of elementary particles and that kind of thing,
1:24:28.000 then that will be amazing. If it turns out to be false and it's superseded,
1:24:32.080 that will also turn out to be amazing. So, that's a nice state for a field of science to be in,
1:24:39.280 that whether it succeeds or fails in its current hopes, that will either way it will be amazing.
1:24:56.480 Computer technology, aside from quantum, is of course making huge progress and computer software
1:25:04.240 in the sense of machine learning and AI is also making rapid progress. I've probably forgotten
1:25:12.000 the most, I'm not remembering the most important thing and somebody will come to me after
1:25:17.040 and say, hey, you know, you dised our field, but I don't mean to dis any field, but you've been
1:25:23.040 talking about where to reality a lot as a theoretical construct to explain the fabric of reality
1:25:28.160 in the book. Yeah, well, that technology is improving, but it hasn't improved in any fundamental
1:25:35.920 way for many years. It's just the resolution has got better, you know, the programming has got better,
1:25:45.440 but there's nothing that would amaze anybody from the 1980s that say,
1:25:54.400 they'd be amazed that the actual technology, like they'd say, you know, shut up and let me try this
1:26:03.120 out rather than answer your questions, but there's nothing fundamentally new that they didn't
1:26:09.600 think of. Whereas, for example, if you go a bit further back and look at what was considered
1:26:15.760 to be the most amazing possible future in the 1960s, let's say, with Star Trek, there are a lot
1:26:24.480 of things from that that were actually achieved, and there are a lot of things from that that
1:26:28.960 were not achieved, and there are a lot of things that were achieved that were not thought of then,
1:26:35.680 not even conceived of. And that kind of thing has slowed down, and I think it's because
1:26:50.480 of irrelevant, I mean, fundamentally irrelevant things, like the sociology of science and the
1:26:56.480 the structure of scientific careers and the structure of universities and the public conception of
1:27:07.280 the educational system altogether, you know, thinking of the educational system as a machine for
1:27:13.760 passing on knowledge is not very compatible with an open society, and open science is and should
1:27:23.600 be the epitome of an open society, but it's less so than it has been in the past.
1:27:36.720 I guess we'll take one final question before we wrap up from Andrew.
1:27:43.520 Hey David, I'm coming over here from the US. I am actually an analyst and we invest in those
1:27:49.760 cutting edge technologies on my PC firm, and so I'm curious to learn well with regard to
1:27:57.920 trying to remove abstraction layers from what is cutting out science and then commercializing that.
1:28:06.320 So, and I think you've already talked on this a little bit, but what would be, I don't know,
1:28:12.800 like an inflection point in the science community that can help proliferate like these
1:28:22.240 more cutting edge frontier technologies, and because you said it's kind of slowed down, so what
1:28:28.320 can kind of stop that stagnation? Well, so one thing is to do with a combination of funding and
1:28:36.240 organization. Scientific research ought to be organized in the form of whether the fundamental
1:28:52.400 unit is the research group. So, if the research group could be one person, in which case that
1:28:59.440 person should be free to do whatever that person wants to do, or it could be a group, in which case
1:29:07.680 that person, it should have a leader who decides who they want to work with, but the organizational
1:29:18.880 structure should be flat, so that the leader is the one that talks to the funders, but apart from
1:29:26.160 that, when they're actually in the lab or in front of the whiteboard, everybody's equal.
1:29:33.040 So, that's one thing. Another thing, this may sound like a very parochial issue, but I think they
1:29:40.880 ought to leave graduate students alone, that there's been a sort of mission creep in universities
1:29:47.520 where it used to be, like when I was a graduate student, it was taken for granted that you were
1:30:01.280 an undergraduate, and then either you went, if you wanted to do research, you would go into
1:30:08.480 research and you would start on day one, being in some research group or working on some problem
1:30:13.120 with your own or whatever it was, or you would first have to do an MSC and then start being a
1:30:20.880 proper graduate student, but now graduate students are from the point of view of the research itself,
1:30:33.680 they are not in a creative role, they're in a subordinate role. It's hard enough for people who
1:30:47.760 come out of say a school with its regimented structure to do research, to build up on the
1:30:54.720 on-ramp of an undergraduate degree, to actually do creative research, but this has been made
1:31:01.120 harder and harder by more and more burdens being placed on graduate students, and then it doesn't
1:31:11.440 stop with graduate students, postdocs now have to spend a significant proportion of their time
1:31:19.280 meeting irrelevant requirements from their departments, from the funders, from the government,
1:31:28.560 whoever is in charge of them, so there's somebody who is called a postdoctoral researcher
1:31:36.960 that is, you know, doctor means learn it, they're a learned person, they've become a learned
1:31:44.160 person, they're supposed to be doing research and creating new knowledge, the job description
1:31:51.360 says that they're supposed to create new and original knowledge, and yet they have to spend most of
1:31:56.880 their time meeting existing criteria, that's not creativity, so those are a couple of things,
1:32:07.520 they're no doubt there are many more. Thank you. So as the final question, then David, what would
1:32:17.920 be your advice to those people, what would be your advice to future potential,
1:32:24.160 the covers of algorithms or scientific theories? In other words, what would be your advice for
1:32:30.080 seekers of the truth? Well, I don't give advice because when you give advice, what happens next
1:32:42.880 is then your fault, and I don't want anything to be my fault. So the, I think the only, like,
1:32:53.440 I could say, just be aware that the hierarchy and the structure whose ostensible purpose
1:33:10.560 is creating knowledge actually acts largely to prevent that, and the fact that this doesn't
1:33:23.040 result in immediate collapse of society is due to the fact that people, some people manage to find
1:33:30.960 their way through the minefield, through the labyrinth, through the dark forest, and do the research
1:33:37.680 anyway, despite all the discouragement and the compulsion to do irrelevant things. So if my advice
1:33:48.720 is find your way through the minefield, or find a field that doesn't have minds or something,
1:33:54.960 it's not going to be much help. All right. All right. Thank you, David. A plows. I don't know
1:34:08.880 how we can do it on Zoom, but it has been fun. Thank you all. Thank you so much, David. Thanks for
1:34:16.160 everyone who showed up. After maybe some edits, we'll post this on our website and our YouTube channel
1:34:24.400 by Quantum Information Society. And I've seen that there were many questions in the chat.
1:34:30.880 We had many questions that were unanswered. You couldn't take any questions from the chat.
1:34:35.040 So we'll find a way to maybe somehow record those and make use of them in our future sessions.
1:34:41.120 But thanks for everyone who showed up. I guess I'm going to stop the recording now. Thanks, David.
