Episode 65: Causality, Time, and Free Will

• Links to this episode: Spotify / Apple Podcasts

Unofficial, reviewed transcript. It may still contain mistakes; check the podcast when wording matters.

Speakers are labeled by first name where identified. Unidentified-speaker labels may refer to different people. Bracketed question marks indicate uncertain wording.

Transcript

[00:00:11]  Peter: Hello, welcome to the Theory of Anything podcast. We've got Bruce and Cameo here. How are you guys doing?

[00:00:18]  Unidentified speaker: Good.

[00:00:19]  Unidentified speaker: Excellent.

[00:00:19]  Bruce: Excited to be here.

[00:00:21]  Unidentified speaker: Yes.

[00:00:22]  Peter: Today, we're going to be talking about something, Bruce's suggestion, of course, that I currently am at the Dunning-Kruger curve of knowledge acquisition. I know it's kind of a silly thing, but where you think you know about something, and then you learn a little more, and you realize you don't know anything. That's where I am, right at the bottom, looking up. But I'm in a fortunate position, because I've got Bruce and Cameo to help me along here, help me up the curve a little bit. You know, that's what I love about this group that I started. And I've got a lot of people who have joined and are far more knowledgeable about this stuff than I am, and are willing to help me along. There was just most… and most people, I just really appreciate how nice people are and generous about sharing their knowledge. There was there was one guy who kind of rage quit after accusing me of what was the acronym for explain it to me like I'm a five-year-old.

[00:01:29]  Peter: But anyway, but as far as I'm concerned, he actually had some pretty major misconceptions about Deutsch himself. So if people have that attitude, it's best if they just leave. But that's pretty rare actually in this community from what I've seen. Most people want to help others learn. And I never feel bad about not understanding David Deutsch, rereading this chapter on causality and time. I don't feel like I'm reading a pop physics book. It's more of a something more like a scientific paper, masquerading as a pop physics book, honestly. And you know, maybe this sounds a little heretical, but I actually don't usually recommend people read David Deutsch, just who are just new to the multiverse and new to wanting to understand about these more out there theories. I mean, I just know for myself, I got a little bit more out of as an introduction, Sean Carroll's book and Max Tegmark's book and some of the other pop physics books I've read.

[00:02:42]  Peter: You know, not that this is a criticism of Fabric of Reality. It's just to me, it's just on another level, really. And I'm trying to get there, trying to understand. And I suspect it will take me the rest of my life.

[00:03:00]  Bruce: Brett Hall once jokingly suggested that we create a new unit that represents how much information you're taking in per sentence and that we name it the Deutsch.

[00:03:11]  Peter: Oh, yeah.

[00:03:12]  Unidentified speaker: Yeah.

[00:03:13]  Peter: That's that.

[00:03:14]  Unidentified speaker: That rings true.

[00:03:17]  Peter: I don't feel bad for for not understanding. But anyway, I know, I know Bruce has a lot of thoughts on this area of time, causality and the multiverse and how these these are sort of intertwined ideas. I thought perhaps we could start by by asking by me asking you a question or two about this chapter. And then and then we can move it, move in, move from there. So, you know, I admit that I've got a pretty rudimentary understanding of the multiverse. How I'm kind of thinking of it is just the whenever there's an interaction, interaction between a quantum event and something larger, which is, of course, basically all the time, the world sort of branches off into different different worlds. OK, so you can think of the world as on a quantum level is as as time moves on, just branching into, you know, most of these worlds are kind of not that much different.

[00:04:18]  Peter: Of course, that's how I imagine it in my mind. But I kind of know enough that to know that that's actually not quite right, that time is… is… it works… This is more me applying my ideas of or my perception of the flow of time to the multiverse and other intuitions, I guess. It's the reality is a little more complicated. So perhaps you could you could explain it to me like I'm five years old, Bruce.

[00:04:49]  Bruce: Okay, so I actually had this idea because for a podcast because of your Facebook page, someone was talking about this quoted from this chapter and it had been a while since I had revisited it and I was thinking about how when I first read this chapter, I had problems with it and you can see it like I still have my book from way back when I first read it and I've got all my notes in the margins and I'm saying paradoxical, huh? Things like that during this chapter. And as I mentioned in some of the episodes, I spent years trying to work out what Deutsch was saying afterwards and I tried to find the best criticisms of what Deutsch was writing. And what I found for the most part was that Deutsch was spot on on almost everything he said. I would go to Penrose, who by the way, if you think Deutsch's pop physics book sounds like a paper, you should read Penrose's where he includes the actual mathematical equations, you know, two or three per page.

[00:05:50]  Bruce: And then he jokes about how he loses half his audience with each equation. But I went and I tried to find the best criticisms, the best alternate opinions, and for the most part I couldn't find any good criticisms of Deutsch's theories. That's not true for this chapter. This chapter is one of the few. I've mentioned some criticisms of the Omega Point, which is obviously one of the chapters of Fabric of Reality. So I guess I have offered some criticisms of that part of the theory, although I really like that theory, just in terms of this is what we're looking for, even if this theory itself might not be right. But this chapter on time and causality and counterfactuals and things like that, I found all sorts of really good criticisms of it. And so I have, I've always had kind of mixed feelings. I do think he's saying some really interesting things here about time, but I think he's leading it towards a series of conclusions that I don't think are tenable.

[00:06:50]  Bruce: So I actually wanted to kind of talk about that for the episode, some of the criticisms that I found of the things you're saying. By the way, this is the one chapter that he recanted quite a bit. So I know he at least somewhat agrees with me that some of what he said, he no longer agrees with. And I actually have a quote from one of his papers where he admits this. So I think this chapter did contain a number of mistakes, and I think that even he fully accepts that at this point, and he's trying to work out a different version of these ideas.

[00:07:23]  Peter: It would be helpful, Bruce, there's a paragraph in here where he sums up, the very end where he sums up what he's saying in this chapter.

[00:07:33]  Unidentified speaker: Sure, why don’t you read that?

[00:07:35]  Peter: [Reading Deutsch:] Let me sum up the elements of the quantum concept of time. Time is not a sequence of moments, nor does it flow. Yet our intuitions about the properties of time are broadly true. Certain events are indeed causes and effects of one another. Relative to an observer, the future is indeed open and the past fixed, and the possibilities do indeed become actualities. The reason why our traditional theories of time are nonsense is that they try to express these true intuitions within the framework of a false classical physics. In quantum physics, they make sense because time was a quantum concept all along. We exist in multiple versions in universes called moments. Each version of us is not directly aware of the others, but has evidence of their existence because physical laws link the contents of different universes. It is tempting to suppose that the moment of which we are aware is the only real one, or is at least a little more real than the others.

[00:08:41]  Peter: But that is just solipsism. All moments are physically real. The whole of multiverse is physically real. Nothing else is. [End of quotation.]

[00:08:51]  Bruce: Yes, great summary. By the way, there's a lot there that I would still say is correct. But it's woven in a way that I think leads to several misleading conclusions. I should also note that this is precisely why Sadia did four episodes on this podcast with me about the problems of physics and particularly the mysteries of time. And she sees time as this unsolved problem. Now, obviously, Deutsch probably wouldn't agree with her because as you just read from that summary, he feels that time isn't a wholly unexplained thing. We know that all there is is the multiverse and a lot of the things that she raises as issues, I think he sees as non-issues. However, he does seem to at least somewhat agree with her. Let me see if I can find the quote. Here's the quote. [Reading Deutsch:] However, this is not quite how the multiverse works. A workable quantum theory of time, which would also be the quantum theory of gravity, which is the theory we don't have, has been a tantalizing and unattained goal of theoretical physics for some time now.

[00:09:58]  Bruce: But we know enough about it to know that though the laws of quantum physics are perfectly deterministic at the multiverse level, they do not partition the multiverse in the manner of Figure 11.6 into separate spacetimes or into super snapshots of each other, which entirely determines the others. [Quotation wording unclear; end of quotation.] On the one hand, he's admitting that the existing theory does not fully have a theory of time, which is, I think, Sadia's point. On the other hand, he's saying we've actually worked out quite a bit, right? And so there is a lot that I think he explains. He spends quite a bit of time on spacetime. And that was actually what led to the argument on your board, is one of the people commenting was taking issue with the fact that they were talking as if the quantum... So some people were talking as if the quantum multiverse was a spacetime, was still a block universe, as they're called, where others were taking exception saying that no, that's not the case, that only Einstein's theory was spacetime and was a block universe.

[00:11:00]  Bruce: And then somebody else was arguing, no, you're all wrong. Newton's theory was also a block universe, and so is quantum theory. So they're all block universes, they're all spacetime. And I commented that they're all arguing over a word, which is, I think, a mistake people make almost constantly, right? And you know how Wittgenstein? Wittgenstein?

[00:11:26]  Unidentified speaker: Wittgenstein?

[00:11:27]  Bruce: Wittgenstein, yeah. He had this idea that all philosophy was just disagreements over words and that none of it was meaningful. And Popper took really strong exception to that because he believes there's real philosophical problems that can be solved. Here's the thing though, they're both right. Wittgenstein is wrong that all philosophical problems are just pseudo problems.

[00:11:53]  Peter: Popper did say that we should never argue about words, though. But he just doesn't take it quite as far. There are actual… He doesn't say that all philosophical problems...

[00:12:03]  Bruce: And I'm fairly certain that the vast majority of philosophical problems are just pseudo problems. Precisely because people are arguing over words. And philosophy is so rooted in trying to define things, which is the essentialist error. Now, there are reasons to define things. I'm not saying you shouldn't, right? I largely agree with Popper, but I think that I could probably lay out a theory of words that's based on Popper that expands a little bit beyond him about when it's good to define things and when it's bad to define things. But this argument was in fact a pseudo problem. And the reason why I'll explain it is because the word space-time came up and was brought about, or at least popularized, by Einstein's theory, where space-time was this block universe where time was just a dimension. And Newton's theory, while you can technically in retrospect think of it as space-time, and that would be an accurate way to think of it, time just simply played no actual physical role.

[00:13:01]  Bruce: And you just used time the way you would normally intuitively think of time. It didn't have any physical consequences, whereas that's not true for Einstein's theory. So the concept of space-time did not really exist until Einstein and is intimately connected to his theory. So when somebody says space-time, they might mean Einstein's theory, or they might mean a concept that comes out of Einstein's theory that could in theory be retroactively applied to Newton's theory. And by analogy, could be sort of analogously applied to quantum theory. And so you could call all of them space-time, having space-time, or it might just refer to Einstein's theory. And we just figured out based on context of how the word was used.

[00:13:50]  Cameo: This is also, we've talked about this in the dogma episodes. I think this is why people get really obsessed with the semantics, because they internalize the portion of that that they're aligning with. And then anytime somebody's arguing with any elements of that, they're taking it personally, but everybody's kind of talking about the same thing.

[00:14:14]  Bruce: It becomes frustrating because you have this giant argument over nothing. So Deutsch, throughout the book, when he uses the word space-time, I'll give you a quote where he does. He means specifically Einstein's theory, right? Or anything before quantum physics. So when he uses the term space-time, he's explicitly excluding quantum physics. Now it's true that quantum physics, even the Many Worlds version, has something a lot like space-time that you could by analogy call space-time. And Deutsch isn't denying that, right? I mean, he's describing how the quantum version of space-time is different than the traditional classical physics version of space-time. Once you realize that's what he's saying, it's not that hard to figure out how he's using the terms and that he's saying something that, you know, makes sense, right? At least you can understand what he's saying.

[00:15:10]  Bruce: He's not talking nonsense.

[00:15:12]  Peter: You know, just looking at this terminology section, okay, flow of time, nonsense. Space-time seems like a meaningful concept, but then he says space-time physics is at best an approximation.

[00:15:29]  Bruce: Yeah, unless by space-time physics, you mean quantum physics.

[00:15:33]  Unidentified speaker: Okay.

[00:15:34]  Unidentified speaker: Okay.

[00:15:35]  Bruce: So this is why there was a bunch of confusion over this and they started quoting from this chapter and things like that. So I wanted to kind of talk about, let's maybe summarize Deutsch's view and what he's trying to pull from this, because he's trying to hit a lot of things, right? So he takes a look at what the Einstein version of space-time would look like. And the key thing that you have to keep in mind with the block universe is that it already all exists, right? It's not like there's moments unfolding. That's how we perceive it because we're living inside of it. But somewhere, in a certain sense, the physics is all there. If you could see to the fourth dimension, you would see that all time already exists and it's fixed. So, Deutsch then uses this to try to show that this causes a problem for the concept of causality. And this is where I think Deutsch is starting to get off the rails, unfortunately.

[00:16:38]  Bruce: It's not that there isn't a legitimate problem here to solve, probably a philosophical problem, but the way trying to solve it through a multiverse, I don't see how it could ever work. And this is what he's trying to do. So let me actually start with the fact that he sees the flow of time as an illusion. So he says, [reading Deutsch:] To exist at all at a particular moment means to exist there forever. Our consciousness exists at all our waking moments. Admittedly, different snapshots of the observer perceive different moments as now, but that does not mean that the observer's consciousness or any other moving or changing entity moves through time as the present moment is supposed to. The various snapshots of the observer do not take it in turn to be in the present. They do not take it in turn to be conscious of their present. They are all conscious and subjectively they are all in the present.

[00:17:31]  Bruce: Objectively, there is no present. We do not experience time flowing or passing. What we experience are differences between our present perceptions and our present memories of past perceptions. We interpret those differences correctly as evidence that the universe changes with time. We also interpret them incorrectly as evidence that our consciousness or the present or something moves through time. [End of quotation.] Okay, first let me try to steel man what he's saying because there is something to what he's saying here. And then let me explain why I have some concerns with the way the way he's saying it. Okay, so he's taking exception to the idea that there could be a flow of time. This is the thing. I’m actually not sure I agree with him on because I think the phrase the flow of time is an emergent concept that properly describes what our experience is like.

[00:18:22]  Bruce: But he's saying it's just an illusion. Okay. And his basis for saying it's an illusion is because of this concept of the block universe that actually there's not a single present with you in it. And then you're you're moving forward. All of those points in time already exist. And as we're going to see, as he the chapter continues, they exist in separate universes because time is just a different universe. Right. And they're all already there. And for the you in that universe, the present is now even the one in the future, even the one in the past, et cetera. Therefore, there is no actual physical concept of present. And that is therefore an illusion, he's arguing. Does that make sense so far?

[00:19:04]  Peter: That makes sense. So just to summarize, when someone says that time is fundamental, including, you know, I think some physicists hold this view, as far as I know, that what they are basically saying is that there is something very meaningful about right now, whereas Deutsch or others, at least Deutsch here, who seems to be saying that the flow of time is nonsense, that's more of a perspective that says right now is not any more meaningful than 10 minutes ago or 10 million years ago.

[00:19:44]  Bruce: So, remember we were just talking about how different words [can mean] different things in different contexts? I am not sure that when you hear physicists say time is fundamental, that any of them are saying the same thing at all.

[00:19:58]  Unidentified speaker: Interesting.

[00:19:58]  Bruce: Sadia keeps bringing that up. That's how come you know about this, because she puts it on the website every single time a different scientist says time is fundamental. And I think to her, it's very similar to what you just said, that there should be something special about the moment, and that the normal way we would look at it. The block universe is the currently accepted theory. So, Deutsch isn't saying something that isn't wholly accepted by physics at this point, right? But the things that she quotes from the different scientists, when I actually go and read them, I don't think any of them are saying what she's saying. And so, for instance, she quotes Lee Cronin. And as far as I can tell, when he says time is fundamental, what he means is that you can look at physics in terms of his assembly theory and that time is fundamental to assembly theory.

[00:20:47]  Bruce: I don't think it's got anything at all to do with the point she's trying to make, right? And I could give you other examples. I don't have them off the top of my head. But I think each of the times she's quoted a physicist, none of them were actually agreeing at all. They were not agreeing with her. They were not agreeing with each other, if that makes any sense.

[00:21:07]  Unidentified speaker: It does, yeah.

[00:21:08]  Bruce: So I'm a little unclear what that means. It probably means something different for each scientist, that they've got some idea that we don't take care of time well in physics, right? That in Newton's theory, time was just, you just used it like you normally did. It had no actual physical component. In Einstein's theory, it became a new dimension, so it was no different than space. And then it thus was part of physics. But we know Einstein's theory is wrong in some way. We know that there's something not quite right about quantum physics, and it doesn't really fundamentally explain time either. So I think that's what Sadia means, is that there should be a physics that describes exactly physically what time is. And in fact, Deutsch, from the quote I just made, says yes, that's what the quantum theory of gravity is going to be. It's going to describe what time is.

[00:22:00]  Bruce: So Deutsch isn't even necessarily disagreeing with Sadia, right? He's admitting that we need a new theory, and that it will wrap time up. And then there's the question of how is time going to be treated? Will it be treated as something basic that everything else emerges from, or is it going to be treated as an emergent concept? Okay, well, we don't know. We don't have that theory today, so we don't know if time is going to end up in the theory of gravity as an emergent concept or as a fundamental concept. Sadia is arguing for it needs to be a fundamental concept. And you can go back and listen to her podcast that she did on this podcast, and she does a pretty good job of explaining her thinking on this. Let me just say, though, that Sadia's theories are non-testable theories, right? This is really not theories at all in the scientific sense.

[00:22:52]  Bruce: These are more like descriptions of what she hopes it will turn out to be, and based on her current intuitions. And as Popperians, we should kind of look at theories in that way. We're always interested in what's the best theory currently available that has the highest empirical content and has survived the most tests, and that is quantum physics, as of today. Yes, there are known problems, and it's just impossible to use your intuitions to know what the final theory is going to look like. Any intuitions you have, like Sadia, as to what that final theory is going to look like, are pure conjecture. There's almost never a particularly good reason to pick one over the other initially until you start to figure out how do I turn this into a testable conjecture.

[00:23:43]  Peter: I hope this doesn't take us down a tangent here, but it does seem to me, do you think that a Popperian framework can ever be kind of used as a way to shut down people from just asking like provocative questions?

[00:23:59]  Bruce: You know, I could see how you could do that. And I know Sadia has complained about that, that when she's talking to Popperians, that they'll say, oh, that's not a best theory. And they won't even look at the conjecture because it's, quote, at odds with the best theory. Well, if you took that too far, you could see how that could very quickly become, you could never replace a current best theory even though it's wrong, because anything you try to replace it with is at odds with the best theory.

[00:24:26]  Cameo: I think that statement is dangerous anytime. If you don't have the patience to even look at a new theory because you believe it's at odds with your best theory, you're not an empiricist.

[00:24:42]  Unidentified speaker: Right.

[00:24:42]  Cameo: You should always be open to the next, that the next theory could absolutely obliterate your current best theory, even if it's completely at odds.

[00:24:52]  Unidentified speaker: That's like your job.

[00:24:54]  Bruce: So let me actually, I have an analogy here that, talking with Sadia on your page, I came up with that I actually think is maybe helpful here. Science is like the halting problem. So you're familiar with the halting problem from Turing machines, where it's impossible to write a program that can detect if another program is going to halt or not. In science, you have a current best theory, and it is irrational [possibly “rational”; word uncertain] to accept that current best theory, if it has no competitors. And it's rational to assume that if that theory has survived the most tests, and it has high empirical content, and in the case of quantum physics, we don't even have a single counterexample of an experiment that hasn't matched quantum physics. So, given that, we know quantum physics is the highest verisimilitude theory we currently have in physics.

[00:25:50]  Bruce: There are no competitors to it. String theory is the best known competitor, and nobody, not even the deepest string theorist, would tell you it's a true competitor to quantum physics as of today.

[00:26:02]  Unidentified speaker: Because string theory has no empirical content yet.

[00:26:06]  Bruce: So, based on that, there's a certain sense in which it's very, very rational to draw conclusions from the highest verisimilitude theory. Why would it ever make sense to draw conclusions from a lower verisimilitude theory? Stop and think about that for a second. Could you ever be rational to draw conclusions from a lower verisimilitude theory if you have a higher one available? Think about that just for a second.

[00:26:32]  Peter: Well, I mean, maybe as a compelling thought experiment.

[00:26:35]  Bruce: Maybe as a compelling thought experiment. But you would never want to go calculate and engineer an airplane based on some compelling version of Aristotle's physics when you know you got Newton's physics or Einstein's physics or quantum physics available to base your predictions on.

[00:26:54]  Unidentified speaker: Okay.

[00:26:54]  Peter: Very fair point.

[00:26:55]  Unidentified speaker: That's a very good point, Bruce.

[00:26:59]  Bruce: So, this is the definition of rationality from a Popperian standpoint. That it just makes rational sense to go with the highest verisimilitude theory, the one that hasn't been refuted yet, the one that you know must be closest to the truth based on everything you know at this point.

[00:27:18]  Unidentified speaker: Okay.

[00:27:19]  Bruce: Now, having said that, you know that the thing that all... there's no guarantee your theory is correct. In fact, in this case, we know quantum physics is false. So one of the main things that Sadia and I get into arguments over is that she'll bring up and she'll say, look, I've refuted quantum physics because of the problem of time. And I'll say, you have, great. I already knew that because of the problem of gravity. That was actually a much bigger problem. And everybody knows quantum physics is false. You're only telling me something that I already know, that absolutely everybody already knows, right? My interest in quantum physics and drawing conclusions from it doesn't come from the fact that I think it's true. It comes from the fact that it is the truest theory currently available. Now, one of those implications is Church-Turing-Deutsch thesis.

[00:28:06]  Bruce: Now, I've said over and over, it could be wrong. We could find out that when we get to quantum gravity, that there's a new type of computer we can build. Nobody's denying that fact. But if I want to sit down and say, is it possible to build an AGI today? The only rational thing for me to do is to look at the current best theory, quantum physics, look at the fact that Church-Turing-Deutsch flows directly from that theory. It's an implication of it. And then say, based on the current theory, yes, I should be able to build an AGI on a regular Turing machine. And that is Deutsch's point. Sadia wants to say instead, look, we know it's false, so you don't know that. To which I say, you're right. It's false, and we don't know that. But I already knew that. That never had anything to do with my conclusion to begin with. I was simply drawing conclusions based on what the current best available theory was.

[00:29:01]  Bruce: And I'm admitting it could be false. And that's really kind of what Sadia and I keep arguing over, and there's really no argument there, right? I've never felt like she understands what I'm saying. That we're not arguing, right? It's, I'm agreeing with you, it could be false, and also it happens to be our best available theory. And according to it, we should be able to build an intelligence on a computer. AGI should be possible. So that is the nutshell of rationality. Now, how do you fit that into conjectures? Well, you have problems that exist, right? We know quantum physics doesn't explain gravity. We know that it has a kind of tentative view of time and it's not very complete, right? There's many things that we know are problems. Now, quantum physics, interestingly, doesn't have what we would normally think of as a problem from a Popperian standpoint.

[00:29:53]  Unidentified speaker: It has no counter examples.

[00:29:56]  Bruce: Most theories throughout history of time have had counter examples. What Popper calls a refutation, although I've made it clear, it really only refutes a combination of the theory plus the background knowledge. So it's not generally perceived as refuting the theory proper. Okay, so when they find that, I forget which planet, Jupiter doesn't follow the orbit that it's supposed to, according to Newton's theory, the first thing they assume is that something about their theory of the number of planets is wrong, not that Newton's theory is wrong. And then they did the same thing when Mercury was found, not to follow the path that it was supposed to according to Newton. They assumed, oh, we're just missing some thing of some body of gravity out there that we don't know about, our theory about what the solar system looked like is wrong. Now, in that case, it turned out to actually be Newton's theory that was wrong.

[00:30:49]  Bruce: But the existence of a counter example does not ever in science cause scientists to say, oh, we're going to abandon this theory, it's refuted. Okay, so if you're thinking of refutation in that way, you've misunderstood the nature of refutation, which is why I don't like the term [refutation]. Because when I say a theory is refuted, the first thing that comes to everybody's mind is this is now a worthless theory. And we need something else, we must abandon it, right? And that just isn't how science works, like even a little bit, okay, because of this concept of verisimilitude. Based on all this, you do want to make conjectures as to how you might go about solving the known problems. Now, to do that, you aren't going to have that conjecture pop out of your mind fully formed as a competitor to the current theory. So think about Einstein. He had that moment, the greatest, happiest moment of his life, where he realized that the force of gravity could instead be thought of as bodies moving together, you know.

[00:31:51]  Bruce: And like if you're in an elevator, you don't actually know if you're weightless or if you're falling. And he had kind of had this thought, and it led to a series of thoughts in his mind, intuitions, as to what physics might be like. He then spent eight years working out what that must be. By the way, when Deutsch talks about perspiration, let me just point out that human creativity has considerable perspiration in it. So eight years is a long time, right, for Einstein to creatively come up with general relativity. So he works very hard on this, and out at the end of eight years pops out a true competitor to Newton's theory. And at that point, the rest of the scientific field has to start thinking about it. And when we do our episode on corroboration, I'll talk about kind of what happened from there. And I don't want to get too far down that road at this point.

[00:32:50]  Bruce: But it's completely acceptable that Einstein must have spent eight years dwelling on versions of his theory that weren't yet competitors to Newton's theory, because they didn't have empirical content yet or had very little empirical content. And he must have at some level just had an intuition that if I keep going down this road, I'm going to end up eventually with a true competitor to Newton's theory. And so he just kept working it out. He kept solving problems. He spent eight years trying to figure out how to fit it all together. And then he was right. He ended up with a theory that was better than Newton's theory at the end.

[00:33:30]  Peter: When Einstein had this revelation about, you know, like, I mean, I don't know exactly how it played out, but what was going on in the elevator where you don't feel like you're moving, what if you had had, I mean, it doesn't seem, it seems like a pretty common thing. If you had asked someone, another physicist at the time, why that was, what do you think they would have said?

[00:33:52]  Bruce: Well, I don't, I don't think they would deny that Einstein was right about that particular revelation. Everybody knows it. The idea that you could rethink gravity in terms of that probably would have seemed completely heretical.

[00:34:03]  Peter: Seemed, seemed weird, but they would have, they would have at least immediately recognized it as a pretty valid problem.

[00:34:10]  Bruce: So here's the thing, because reality is, right, as Popperians were realists, there really is a reality out there. It is, whatever it is, it is the way it is.

[00:34:21]  Unidentified speaker: Okay.

[00:34:21]  Bruce: And it's, it's self-consistent. If you happen to be on a right path, your theory can be developed into an empirical theory, superior to the current empirical theory. And so that is the goal of science, is that we're always trying to make a new empirical theory better than the last empirical theory.

[00:34:41]  Unidentified speaker: Okay.

[00:34:42]  Bruce: So I'm emphasizing the word empirical because that means testable, right? That you can actually, it can clash with actual experience. Now that must not be the case during the eight years. Einstein's theory in its more nascent form is not a better theory than Newton's theory. And yet he is just convinced that it's going to be someday. Now, if he happens to be right, he will be able to develop it into a better theory. If he's wrong, then he won't be able to. Now, I would note that this is exactly what happened to Einstein. He had a theory of what improved theory of everything should look like. And he spent years going down the path following his intuitions as to what that theory of everything should look like. That would be the theory of quantum gravity. And he completely failed to make any progress at all on the problem. Because his intuitions were wrong in that case, whereas the first time his intuitions were right.

[00:35:47]  Bruce: Now, I want to emphasize, intuitions are not sources of truth, not even a little bit. They are sources of conjecture. And so it's completely legitimate for, to use Sadia as the example, for her to have an intuition, time should be fundamental. Or have an intuition, CTD should be wrong. Or to have an intuition, novelty should be worked into the laws of physics. But none of those are sources of truth. Those are just her making a conjecture. And that's it. There's nothing else going on. It's now up to her to figure out, can I turn this into an empirical theory? And if her intuitions are wrong, she'll be unable to. Okay, does that make sense? That's the halting problem. So you never get to actually know. And then you never know. You never know if the reason why you couldn't turn it into an empirical theory was because you were wrong, or if the reason why you couldn't turn it into an empirical theory is because you just didn't get creative enough.

[00:36:47]  Bruce: You never get to know that for sure. What you do get to do, though, is you get to see what other people come up with as theories. And if those are empirical, if they pass tests, if they explain problems that the old theory had, if they explain the success of the old theory, that's something that's really important in science. As Cameo just mentioned, that the new theory may be totally at odds with the old theory. From a certain point of view, that's true. I think people tend to see Einstein's theory as, quote, totally at odds with Newton's theory. But Einstein's theory completely explains the success of Newton's theory. In that sense, it subsumes Newton's theory. It shows why Newton's theory was such a good theory, right? And that's what we're looking for in science, is we want the new theory to be able to demonstrate why the old theory was so successful, why it had so much truth to it.

[00:37:35]  Bruce: And by the same token, though, I think you could argue that the person going with the best theory is also faced with the halting problem. So I, as opposed to Sadia, I think you can build an AGI. So I want to study what algorithms could be used to build intelligence. Now, if Sadia's intuitions are right, then obviously I'm going to be wasting my time because we need new physics before, and I admit this in the podcast when I'm talking with her about this, we need new physics, and I'm going to just be completely wasting my time trying to discover what the intelligence algorithm is because we really need to discover new physics first before it's even possible to come up with the intelligence algorithm. On the other hand, the reason why I've decided to do this is because according to our best theory, quantum physics, CTD is true, and CTD, Church, Turing, Deutsch thesis says that we can do this on a computer.

[00:38:34]  Bruce: So it's rational for me to believe that there should be an algorithm that can do this, even though I know, as a fallibilist, that I'm basing this entirely just on the best theory, which is rational for me to do, but I know I could be wrong. Does that all make sense? And there's still that halting problem. When I fail to find the intelligence algorithm, that could be because I'm relying on a false theory, or it could be because I just haven't gotten creative enough and I don't know which it is.

[00:39:04]  Peter: Everything you're saying makes sense, and I'm with you, and I'm just waiting for this to loop back into the time here.

[00:39:13]  Bruce: Okay, so let's loop this back into time. So I just read that quote from Deutsch. I'm going to take some exception to it now. Not too much yet, okay? But let me take one of the sentences that he writes. To exist at all at a particular moment means to exist there forever. Isn't that a false statement according to his own argument? I mean, even the fact he uses the word forever, right? That implies some sort of outward time that this block universe exists in, and that if we could see that other time, we would see that it's there forever, right? The issue here is that there's just no way to speak except using a sort of analogy of time. And this is why I think his argument that the flow of time… [unfinished clause] doesn’t make a total sense to me, okay? And let me read this again. We do not experience time flowing or passing. What we experience are differences between our present perceptions and our present memories and our past perceptions.

[00:40:15]  Bruce: Well, to me, that sounds like a flow, right? You've got a flow of new memories. So I would say, yes, we absolutely, there absolutely is a flow of time. It's an emergent concept that is a way that we talk about and the way we perceive, and it's accurate, right? It's not false. It's not an illusion. It's a proper description of what it feels like to live inside of a block universe from moment to moment, right?

[00:40:43]  Peter: Yeah, that's an extremely compelling way of thinking of it. It's almost like the reality of abstractions idea. I mean, it's real in the sense that it's just a different level of abstraction. Is that what...?

[00:40:58]  Bruce: Yes, that is what I'm saying. So on the one hand, he's still making a fair point about this concept of the block universe. And from all my looking this up, he's accurate, right? Einstein's theory is a block universe, and honestly, so is quantum physics. So both our best theories say we live inside of a block universe, that the future and the past in some sense already exist out there. If we could somehow imagine stepping out of this, looking at it from another dimension, and there being a magic extra time out there, we would see that it was there forever, which is how I make sense of his statement, right? But all of that, of course, there is no way to step outside of it, and time only exists in the sense of being inside that block universe. And that time that exists inside that block universe, it flows. And the flowing is a real thing, okay?

[00:41:53]  Bruce: And so this is the part that I can accept about what he's saying, and also how I would look at it differently than him. Now, why does this matter? So let me move to another quote here. What Deutsch goes on to do is he tries to link this to determinism. So both Einstein's theory is deterministic, and so is the many-worlds version of quantum physics. Although the determinism of quantum physics is at the level of the multiverse, not at the level of someone living inside of a universe, okay? From the point of view of someone living inside the universe, the world is not deterministic. It's partially probabilistic. It's partially stochastic. But if you could see the whole multiverse, you would see that it's actually still deterministic. Now, this determinism bothers people. And it bothers Deutsch, clearly, or at least it did back when he was writing this chapter.

[00:42:43]  Bruce: So he's trying to use the multiverse, the existence of the multiverse, he's trying to look at the fact that in a block universe, that it's hard to say what causality is, because that whole universe exists. So did something cause something else? And he argues that if the block universe were true, that there is no causation. But then he says, but that's okay, because we don't live in Einstein's space-time, we live in quantum physics. So he then tries to tie that to things like free will. He tries to explain that free will is only possible in the multiverse. This is the thing that he then later partially recanted. Okay, so here's a quote. [Reading Deutsch:] The determinism of physical laws about events in space-time is like the predictability of a correctly interlocking jigsaw puzzle. The laws of physics determine what happens at one moment from what happens at another, just as the rules of the jigsaw puzzle determine the positions of some pieces from those others. [Comparison wording unclear.]

[00:43:45]  Bruce: But just as with the jigsaw puzzle, whether the events in different moments cause one another or not depends on how the moments got there. We cannot tell by looking at a jigsaw puzzle whether it got there by being laid down one piece at a time, but with space-time, we know that it does not make sense for one moment to be laid down after another. For that would be the flow of time. Therefore, we know that even though some events can be predicted from others, no events in space-time caused another. Let me stress again that this is all according to pre-quantum physics, in which everything that happens, happens in space-time. [Bruce’s aside:] Notice how he's using the word space-time to mean specifically pre-quantum physics, even though you could use that term to include quantum physics. [Resuming quotation:] What we are seeing is that space-time is incompatible with the existence of cause and effect.

[00:44:34]  Bruce: It is not that people are mistaken when they say that certain physical events are causes and effects of another. It is just that the intuition is incompatible with the laws of space-time physics. But that is all right because space-time physics is false. [End of quotation.] Now, here's the... And then from there, he goes on and he works out a theory around how we can make sense of cause and effect in terms of quantum physics instead. I don't think any of his arguments make sense. I'm going to be honest. Quantum physics, if the problem of causation exists in Einstein's theory, which I'm going to argue it doesn't, then every bit as much exists in quantum physics. And there is no way around that. And there is no argument you can make that can change that fact. And this is where I really kind of struggled with this aspect of this chapter. And I can even give you examples of how thought experiments, of how I know I'm right, right?

[00:45:33]  Bruce: Because these are just really straightforward thought experience, experiments that will kind of make the point clear. So let me now read another quote from him. He says, [reading Deutsch:] in space-time physics, which is effectively all pre-quantum quantum physics, starting with Newton, the future is not open. It is there with definite fixed contents, just like the past and present. If a particular moment in space-time were open in any sense, it would necessarily remain open when it became the present and the past, for moments cannot change. Subjectively, the future of a given observer can be said to be open from the observer's viewpoint because one cannot measure or observe one's own future. But openness is that subjective sense, in that subjective sense, does not allow choices. [Bruce’s aside:] I'm going to emphasize that term. [Resuming quotation:] If you want [word uncertain] a ticket for last week's lottery, but have not yet found out whether you have won, the outcome is still open from your point of view, even though objectively it is fixed.

[00:46:31]  Bruce: But subjectively or objectively, you cannot change it. No cause that has, causes that have not already affected it can do so any longer. [Bruce’s aside:] Let me say that sentence again, because it's important. [Repeating quotation:] No causes that have not already affected it, that not already affected it can do so any longer. The common sense theory of free will says that last week, while you still had the choice whether to buy a ticket or not, the future was still objectively open. And you really could have chosen any of two or more options. But that is incompatible with space time. So according to space time physics, the openness of the future is an illusion. And therefore causation and free will can only be no more than illusions as well. To be an effect of some cause means to be affected by that cause, to be changed by it. Thus, when space time physics denies the reality of the flow of time, it logically cannot accommodate common sense notions of cause and effect either.

[00:47:26]  Bruce: For in the block universe, nothing is changeable. One part of space time can no more change… [comparison wording incomplete] another part of the fixed three-dimensional object can change another. [End of quotation.] Okay, Deutsch then, his attempt to get around this is to say, and then he talked about counterfactuals, I should probably mention that, that we talk about counterfactuals. We say, you know, if Cameo had made this choice, her career would have been different. And we know exactly what we mean, and we understand these to be, you know, true or false statements. And yet it's hard to say what the, what the statement of a counterfactual even means, because it either did or didn't happen.

[00:48:08]  Unidentified speaker: Right.

[00:48:09]  Bruce: So he tries to get around this. He tries to say we should understand counterfactuals in terms of the existence of other multiverse, other universes in the multiverse. So we say, we could say, we could imagine that if Cameo had made this different choice, that in those other universes where she made a different choice, something else happened. And, you know, let's say that Cameo is now a millionaire. So she makes this choice. And in those other universes, she's not a millionaire. And in the one where she made the choice, she's now a millionaire. So he's trying to work out that we can make sense of counterfactuals by using the multiverse and that based on that, we can also make sense of causation because we can think of causation in terms of that choice made a difference into how the two branches of the multiverse took place. And this even on the surface, it seems kind of like a compelling argument.

[00:49:01]  Unidentified speaker: All right.

[00:49:03]  Bruce: You're with me so far on the argument. And then from there, he kind of works out that this is what free will probably is. So you're with me so far.

[00:49:11]  Unidentified speaker: I'm hanging on.

[00:49:12]  Peter: Can I ask you a question? It might be a dumb one, but it's just what's on my mind. Okay. So I'm thinking about my childhood in Seattle. And then let's think about your childhood in, I assume, Utah. We were in the same universe, correct? The same world. I mean, the puzzle pieces aligned to put us in the same reality. I mean, that's just common sense, right?

[00:49:47]  Unidentified speaker: Yes.

[00:49:47]  Peter: But then the future, I mean, these puzzle pieces are unaligned, at least from our perception. It just seems that the future is a lot different than the past from that perspective.

[00:50:01]  Unidentified speaker: Yes.

[00:50:02]  Peter: I mean, the puzzle pieces, there was an order to the puzzle pieces in the past that is not as apparent in the future.

[00:50:12]  Bruce: So yes. And from our point of view, as someone living inside of a universe, I think that's a completely emergent, an accurate, emergent way to think of things. And that's why I have zero objection to the concept of flow of time, and zero objection to the idea that the future is open and that it doesn't exist yet. Sure, I know that if I were an observer in a fake time, looking at it from a dimension, but I'm not and no such person exists, right?

[00:50:41]  Peter: It's unlikely to happen anytime soon.

[00:50:43]  Unidentified speaker: Yeah.

[00:50:44]  Unidentified speaker: Right.

[00:50:44]  Bruce: So it's supposed to be impossible, right? At least according to the current laws of physics, for someone to exist is for them to exist inside of space time.

[00:50:54]  Peter: So the flow of time is not actually nonsense.

[00:50:58]  Bruce: It's not.

[00:50:58]  Peter: You would not put it. He's pretty clear about that. He calls it nonsense, or at least.

[00:51:02]  Unidentified speaker: Yes.

[00:51:03]  Peter: But you disagree with that.

[00:51:06]  Bruce: I disagree. And I've explained that the way he says, you have this illusion because of this, I simply call that the flow of time. So I've relabeled what he's saying is really going on, the flow of time, and I'm done.

[00:51:20]  Unidentified speaker: Right.

[00:51:21]  Bruce: I'm going to deconstruct Deutsch's argument now. So let's start with the idea of counterfactuals. Okay, that we can understand, we can only make sense of counterfactuals in terms of, in terms of the multiverse. Now it is true that the multiverse may help you kind of make sense of counterfactuals to some degree. But we use counterfactuals that have nothing to do with the multiverse. And at some point after reading The Fabric of Reality, writing The Fabric of Reality, Deutsch realized this. And in the current introduction to The Fabric of Reality, he admits that he's changed his mind on this subject. And in his Constructor Theory paper, he mentions this. He talks about, he says this is, that is part of a more general problem of counterfactuals. Some counterfactual statements are simply undetermined. As for instance, in Quine's 1960 puzzle that if Caesar were in charge in the Korean War, the statement, he would have used nuclear weapons and he would have used catapults, are both arguably true.

[00:52:20]  Bruce: In such cases, adding enough extra detail like, and had been informed by modern weapons, removes the ambiguity. Then he goes on to say, I have previously suggested, and he's talking about in the Fabric of Reality in this chapter, Chapter 11, that this problem of counterfactual is solved in the Everett interpretation of quantum theory, where [propositions] like it would have happened can be understood as it happened in another universe. However, this only works for events that do happen somewhere in the multiverse, such as the statistical fluctuations bring into, such as a statistical fluctuation bring into existence a person with specified Caesar-like attributes and causing everyone to regard him as a military commander. It does not work for laws of physics. For example, the principle of testability is doubly counterfactual. So because of this, he started to realize that there are meaningful counterfactual statements that can't be explained in terms of another universe.

[00:53:19]  Bruce: Now, I don't know if he was joking or not. He suggests that you could work out short of working out the idea that Caesar somehow came into existence due to a quantum fluctuation. The fact is that Caesar doesn't exist in these other universes, right? And so when we say Caesar would have done this had he been in charge of the Korean War, that may just not exist in another universe. And yet it's still an entirely viable counterfactual statement. I would suggest this means that you don't need the multiverse to make sense of counterfactual statements and that the multiverse actually doesn't add anything other than maybe some comfort to some counterfactual statements. That whatever counterfactual statements are, that you don't need the multiverse to make sense of them or for them to be a meaningful thing. So that would be my first argument here.

[00:54:09]  Bruce: And I think this part Deutsch seems to agree with me on now.

[00:54:13]  Unidentified speaker: OK.

[00:54:16]  Bruce: Now, based on that, though, there is there's something interesting about this. Counterfactuals is fundamental to constructor theory. And he points out, for example, and Chiara Marletto points out in her book, for example, that you can't really understand the concept of a computer without understanding the concept of counterfactuals. And she's right.

[00:54:38]  Unidentified speaker: OK.

[00:54:40]  Bruce: I think the concept of counterfactuals will always be meaningful regardless of what physics turn out to be true. I just don't think the two have to have any connection at all. And yes, that does make a bit of a mystery about what do we mean when we say this would have happened had Caesar been in charge, even if that never happened anywhere in the multiverse. You know, let's accept it as a legitimate mystery, but I don't think it poses any problem to understanding counterfactuals and how they relate to, say, what a computer is or how constructor theory works. So I don't think you actually have to in any way invoke the multiverse to be able to be able to make sense of counterfactuals. So let me start with that as my starting point. Now, let's now take the concept of free will. So if you have no free will, and this is what Deutsch is arguing, because we exist in space-time, I would argue that that's just as true if we exist in the multiverse.

[00:55:40]  Bruce: Now, I was just listening to Sam Harris. If you take a look at what he said about, famously Sam Harris believes that free will is an illusion. And famously people pit Deutsch as being at odds with Sam Harris on this, because Deutsch quote believes in free will. But if you actually pay attention, there's no actual disagreement going on here. Because Sam Harris is saying that if you look at it in terms of space-time, then free will is an illusion. And we just read that Deutsch agrees with him. Had space-time been real, that it would be an illusion. But Sam Harris goes one further. He says, OK, let's back up and assume it's the multiverse. Some sort of quantum event randomly causes different universes to branch out. That still doesn't affect your free will, right? In fact, it brings nothing to the table in terms of the kind of stereotypical way of thinking of free will.

[00:56:36]  Bruce: So the libertarian view of free will is still just an illusion in the multiverse. I don't see how Deutsch could argue with that. That is absolutely true. Because it still exists as a block universe. It still exists. It's just happening across different universes. Let's say, though, you want to take a strong stance that, no, there's some difference here. I don't know what it would be. Because honestly, my will has… is… the fact that these universes are branching is entirely controlled by the laws of physics. It's not something that's under my control at all. So if there's some universe where I make a choice and it goes one way in one universe and I make a different choice and it goes some other way in some other universe, it was still the laws of physics that determined that. And I still had no choice in it, right? Because it's still a block universe.

[00:57:19]  Bruce: Do you see my argument here as to why I think nothing really changes with quantum physics before I move on?

[00:57:28]  Peter: Well, yeah, it seems to me that Deutsch and Harris are really sort of a question of emphasis in a way. I mean, Sam Harris in his arguments against free will, I mean, he says, well, yeah, we live in a deterministic universe. But the, and he emphasizes that every thought comes from another thought and all this, whereas Deutsch might also agree that, yes, we live in a deterministic universe, happens to be a multiverse, but it's, like you say, it's still equally deterministic, if not more so, really. But, you know, he just sort of, but still, the growth of knowledge is impossible to predict, and humans as knowledge creators, we don't know where our knowledge will grow, and if you call that free will, it seems like a pretty good definition to me, so.

[00:58:24]  Bruce: Okay, so let's, let's, let's take a look at that. Okay, because you actually were exactly where I was going to go with this. So thank you. So the issue here is the definition of free will, or rather what we mean by free will. Now, it shouldn't be surprising that free will is a concept, as a word, is a concept vague [wording unclear], and that we mean different things by it under different circumstances. That shouldn't surprise us even slightly. Now, what Sam Harris is really saying, so this is where it makes sense to define things. Not because we're trying to define a word, but because we're trying to talk about what someone means. Okay, so when Sam Harris says free will is an illusion, he very specifically means everything is determined and that you have no control over it. And if we reroll time and it plays out again, it will always play out exactly the way it did before.

[00:59:20]  Bruce: Okay, which is true whether we're talking Einstein spacetime or a quantum multiverse. And that concept of free will, it does not exist. And Deutsch is not arguing the point. Okay, because Deutsch knows that's true. That concept of free will does not exist. I don't know if I should say it's an illusion. And I kind of take some issue with Sam Harris calling it an illusion, because when I think of illusions, I think of something that does exist, that this thing seems like it is, but it isn't.

[00:59:55]  Peter: Well, he even goes a step further. And the part of his thing that I really actually don't agree with is that he thinks that you can meditate and practice mindfulness and actually perceive that we don't have free will. I mean, I haven't really meditated very much, but that just seems like kind of nonsense to me, honestly.

[01:00:20]  Bruce: In the way that Sam Harris is defining free will, free will does not exist, and that follows directly from not only both of our best theories of physics, but it follows directly from any theory of physics you can conceive.

[01:00:37]  Peter: That's a good distillation, yeah.

[01:00:40]  Bruce: And so if that's what you actually mean by free will, then yes, it does not exist. I don't even think we should call it an illusion, because that implies that there was something that it could have been, but it was just a completely logically incoherent concept to begin with. We can dismiss it entirely if that's what we mean by free will. Now, Deutsch, then actually in the podcast, Sam Harris is arguing with Daniel Dennett, who believes that there is free will. He doesn't disagree with Sam Harris that that type of free will doesn't exist. He agrees it doesn't. He agrees it's incoherent. Deutsch agrees that kind of free will is incoherent, okay? And Deutsch and Dennett agree, though, that when we talk about free will, it's not a complete meaningless statement. We mean something, and it means something that's real. So Daniel Dennett gives the example of, did you sign this contract of your own free will?

[01:01:35]  Bruce: Okay, whatever we mean by free will in that concept really probably has nothing at all to do with the kind of free will that Sam Harris says doesn't exist. We mean something different by free will in that context. And what we mean in this context probably means, were you being constrained by somebody else's will? And if you answer no, yes, I did this of my own free will, that means this was my choice. Now, the idea of your choice is not an unmeaningful concept. Just as the flow of time is a meaningful concept from the point of view of a person living in space-time, the idea of a choice is equally meaningful from the point of view of a person living in space-time. Yes, we can make it seem really strange by imagining ourselves stepping outside of space-time and there being some fake time out there and looking at it from a different dimension, because it already exists at this point.

[01:02:31]  Bruce: But the concept of you making a choice is an emergent concept, just like the flow of time, that is a perfectly meaningful concept from the point of view of an observer. Because what it really means is that nobody else was constraining you, and that's a perfectly meaningful concept.

[01:02:48]  Peter: And some people are more free than others. I mean, someone living in North Korea has less freedom than someone in an open society. So yeah, freedom means something. And if you want to call it free will, I think it's pretty valid.

[01:03:09]  Bruce: Okay, so what we've now said is free will does not exist under certain understandings of free will, and free will does exist under certain understandings of free will. And really, this should be uncontroversial, okay? When Sam and Dennett argue, what they really argue over is whether or not what Dennett is arguing, the concept that Dennett is arguing in favor of, Sam Harris is not claiming it doesn't exist. He's claiming we shouldn't call it free will. And that's it. They're actually just arguing over a word. And once you realize that, the rest of the argument seems silly.

[01:03:47]  Peter: I think you've distilled the argument very well.

[01:03:50]  Unidentified speaker: Okay.

[01:03:51]  Bruce: Now, let me add one more on. Deutsch is arguing that because we are knowledge creators and knowledge creation is unpredictable, that there's a certain other sense in which we have free will, different than the Dennett sense. Though he's not saying Dennett sense is wrong. Where we can create new choices and options for ourselves through creating new knowledge. And this is unpredictable. For some reason, people really tie predictability and free will together. And I'm a little unclear why. And I'm going to show you why in just a second. So he's saying this is what we could think of as free will. Now, I want to suggest that if you define free will that way, that free will is a real thing. This is what we call compatibilist free will. And it's perfectly fine that Sam says, I don't want to call that free will. Because in Sam's mind, the word free will is just too strongly.

[01:04:54]  Bruce: When he says the word free will, too many people associate it not with the Deutsch version of free will, which he agrees is true, but more with the libertarian sense of free will. That's an incoherent concept. So Sam Harris is really just saying, I'm not going to call that free will, because I think it's misleading, which is actually a fair point.

[01:05:16]  Unidentified speaker: All right.

[01:05:17]  Bruce: And Deutsch is saying, well, when we talk about free will, I think that when we're not talking about the incoherent version of it, I think that really we're talking about the fact that we're knowledge creators and we can make choices for ourselves. And I think that's a fair point too.

[01:05:33]  Unidentified speaker: Okay.

[01:05:34]  Bruce: And if you want to define free will that way, then I think yes, free will exists. Now, here's where things get interesting. Deutsch still wants to tie his concept of free will to the concept of counterfactuals. And he still wants to tie it, if not directly to the multiverse like he used to in Chapter 11 of Fabric of Reality. I found quotes from him where he still says, I think we're going to have to understand how these relate before we can fully understand how to build an AGI, for example. Because he thinks that this is something that's going to turn out to be relevant to understand how the multiverse relates to free will and how it relates to counterfactuals before we're going to build an AGI. He's not claiming you need a multiverse necessarily anymore. But he's claiming that the concepts will relate. I'm going to now give you a thought experiment that will prove that this is just simply not true.

[01:06:31]  Bruce: And I brought this thought experiment up to various fans of David Deutsch, and I can't get them to take it seriously for some reason. But I think this is just a lock-solid thought experiment. So according to Deutsch, and according to Deutsch's Church-Turing-Deutsch thesis, we should be able to build an AGI on a computer. That's true, right? Deutsch is not contesting that. Furthermore, he doesn't think there's any reason to believe that the brain uses quantum effects. So he believes that it can even be a regular classical Turing machine. Now, he doesn't know that for sure. And my experiment won't matter which way we go on this. But even Deutsch accepts that we could in theory build an AGI on a regular Turing machine.

[01:07:15]  Unidentified speaker: Okay.

[01:07:18]  Bruce: So, so far, I'm just sticking with something that's just a pure implication from Deutsch's theories. So let's say that we suddenly, that in the future, we have that knowledge, and we build an AGI on a computer. And let me even just go on a limb here. Like, there's a question of, do you have to build an AGI, do you have to have an outside world? I suspect you do. I suspect that if you took a baby brain and you completely disconnected it from all sense, it would simply never develop into a person to begin with. I don't know that for sure. Some people might say, oh, it might still learn to think and think about itself or something like that. I don't really know. For the sake of argument, just so that we're satisfying everybody's viewpoint, let's imagine that this AGI or maybe it's a set of AGI's exist inside of a virtual world. So they have no contact with the outside world and they exist inside this virtual world.

[01:08:14]  Bruce: But the virtual world is rich enough that if you need an outside world to be able to develop intelligence, that it has what it needs to develop into a set of intelligences. Now we know this should be possible too because of CTD. CTD says we can simulate whatever we need to, to whatever level of detail we need. So whatever level of detail we need of an outside world, if any at all, for an intelligence to develop is included in this simulation. This is the point I'm trying to make. Now this simulation is running and these intelligences develop inside. And their knowledge prediction is, from their point of view, entirely unpredictable. And they develop, in fact, it's even unpredictable from our point of view. If I'm outside and I'm looking at this, maybe one of the intelligences, one of the people living inside the virtual world, develop some new mathematical theorem or something that nobody in our world has ever created.

[01:09:12]  Bruce: And now I see it, I've got some way to view inside in the sake of this thought experiment. And I see this equation, I publish it, and I go, look, my intelligence inside my world came up with this really useful mathematical equation, right, or theorem. And it's completely unpredictable what knowledge that these intelligences are going to create, exactly like Deutsch says. So we're still exactly following Deutsch's theories. Okay, this is pure four-strandism still. Okay, now imagine I run the exact same program a second time, but on a computer that is slower. Okay, so the exact same world is going to unfold. The exact same discoveries are going to take place because computers are entirely deterministic. And the exact same discoveries will take place and the world will unfold in exactly the same way, but it's going to happen at a slower pace.

[01:10:06]  Bruce: Do I now have the ability to predict for the second computer on the slower computer what knowledge they're going to create? Yes or no?

[01:10:17]  Cameo: I'm going to say, I don't know.

[01:10:20]  Peter: I'm going to say no, but I'm still digesting it.

[01:10:25]  Bruce: So people try to say no. People try to say, no, you can't predict it. But let me make this clear. Turing machines are entirely, and computers are entirely deterministic. This is the exact same computer with the exact same inputs. They're running on a deterministic machine.

[01:10:42]  Peter: Fair enough.

[01:10:43]  Bruce: Okay, can you predict what's going to happen on the slower of the two computers?

[01:10:49]  Peter: Well, you would need another Turing machine to predict what's going to happen on the first Turing machine.

[01:10:59]  Bruce: In this experiment, you have [two].

[01:11:01]  Peter: Okay, we have [two].

[01:11:03]  Bruce: Well, you do, right? You've got the faster world and you've got the slower world. But other than that, you’re critical [wording unclear]. Okay, so you're looking at the faster world. You see them invent this new mathematical theorem. It was not predictable. You now know that in two days, the slower computer is going to invent the same mathematical theorem. Okay, is there any way around that being true that you can think of?

[01:11:28]  Unidentified speaker: No,

[01:11:28]  Unidentified speaker: no.

[01:11:29]  Bruce: Okay, I brought this up and I had somebody say, what if there's an error? Okay, well, first of all, you've now introduced the idea that knowledge creation and free will require errors, which is a very strange thing for you to introduce. But we can just tweak the experiment. We can say these are really good computers and most computers are really good, even today. And they rarely, if ever, have hardware errors that would affect the outcome of a program. And even if over a long enough period, if they'll have an hardware error, this thought experiment is going to stop before the first error takes place, which is a very reasonable thing for me to say about the thought experiment. Based on this, and I can keep doing this, anything you come up with that you try to make them differentiate, I'm going to, because it's reasonable for me to do so, I'm going to remove that possibility from the thought experiment.

[01:12:20]  Bruce: I'm going to make sure that it is impossible that these two worlds don’t diverge [possibly “diverge”; wording uncertain]. This is absolutely what follows from Four Strands Theory. There is no way around it. It is a lock solid, impossible to get around implication of knowledge creation, of free will. And guess what? It does not matter that it's in a multiverse, because computers are programmed to not have branches in a multiverse. They are doing the exact same thing in every version of the multiverse from the moment they start running, because they have the error correction necessary to not differentiate. There are no random events in the computer that would take place that would cause them to start branching. If they did, we'd call that an error. I've already said this all happens before the first error takes place. So it does not matter that they're in a multiverse.

[01:13:13]  Bruce: The multiverse plays zero role. Now here's where things get interesting. From the point of view of a person living inside this virtual world, they have free will, and it's the Deutsch kind of free will. They can create knowledge, they can make new choices for themselves, and nobody can predict. Nobody inside that world, even on the slower computer, nobody inside that world can predict the creation of knowledge. I can because I live outside the computer, and I have a version of it that's faster, that's running. But as far as the people inside that world are concerned, even if they're on the slower of the two computers, knowledge creation is zero predictable. And they effectively have free will exactly in the sense that Deutsch claims free will exists. And notice that this can happen, that the Deutsch version of free will can exist even in a circumstance where there is no multiverse.

[01:14:12]  Bruce: And that follows directly from Four Strands Theory, from Deutsch's theories. There is not a single thing I have changed in the slightest. These are lock solid implications of his own theories.

[01:14:24]  Peter: And he would, just to be clear, at least the Deutsch of today would agree.

[01:14:28]  Unidentified speaker: Yes.

[01:14:29]  Bruce: And he couldn't disagree. There would be no way for him to disagree, right? Because these are implications of his theories. Now, based on this, I can draw a conclusion. This is why you'll hear me say the multiverse plays no role in free will. Now, the only way around that would be if something like Sadia’s theories were true, where we find out that Four Strands theory is wrong. Okay, that's what it would take, right? For me to be wrong in this conclusion, it's a conclusion based on the assumption that the theories are right, which is always how we draw conclusions, right? If Sadia’s theories that all of physics is wrong, and we're going to have brand new ideas that take us in a totally different direction, then of course, everything I just said was out the window, okay? But part of our thought experiment is assuming Deutsch is correct. There's more here, though, that Deutsch has just argued, we just read that he argued that in the block universe, there are no cause and effects.

[01:15:26]  Bruce: That is not true. That cannot be true. Because there are cause and effects, that concept of cause and effects inside this virtual world is still a meaningful concept, even though it does not exist inside of a multiverse. So it must be that he's wrong, that cause and effect do not exist and are just an illusion if there's a block universe. Now, I would argue that we need to understand cause and effect also as an emergent concept, just like the flow of time and just like free will, and that they are emergent concepts that we use to explain things from the point of view of an observer inside of a universe. And that from that point of view, they are a completely meaningful thing. And that the only reason why they seem like an illusion is because you are imagining a scenario that never happens, could never happen physically, where somebody has stepped outside that universe and can see that it all now exists in advance and he's watching with a fake time, it happening.

[01:16:26]  Bruce: But you know what, even then, I don't see how that's different than us watching these people living inside this virtual computer, right? We can't predict them, right? The only way we can predict them is by running forward in time and then going back and running again, and then we can predict. I've suggested that we call this retrodiction, retrodiction is predicting backwards, that what we really do is we never predict in the first place, we never actually predict the creation of knowledge, of novel knowledge. What we actually do is we predict the second creation of knowledge. And I want to point something else out too. From the point of view of the people living inside this world, we may from our viewpoint say, well, we've got a version of, let's name one of the AIs Bob. We have a version of Bob running on one computer, and we have a version of Bob running on another computer.

[01:17:21]  Bruce: Those two computers aren't running at the same speed. So one version of Bob is from our point of view, now on year three, and one version of Bob in the other computer is now on year two. But eventually Bob on year two is going to catch up to exactly where Bob is at year three. It's just that his computer is a bit slower. From Bob's point of view, there's only one Bob. This is exactly what the multiverse is like. You could think of it as there's always an infinity of universes that are all identical, and then they branch, and they're not identical once they branch. But at any given moment, there's always an infinity that are also identical. To ask the question, which is the real you, is no more meaningful than to ask, which is the real Bob inside this multiverse? For all intents and purposes, Bob exists in a multiverse of two identical universes that are exactly identical, so we might as well just think of them as a single universe.

[01:18:20]  Bruce: Does that make sense so far?

[01:18:22]  Unidentified speaker: I’m hanging on.

[01:18:26]  Peter: So the knowledge creation is a lot like free will, in that it is deterministic. We can't predict it, or at least can't easily predict it, because if we could predict it, we would already have that

[01:18:44]  Unidentified speaker: Yes.

[01:18:45]  Peter: knowledge. So from that perspective, it's kind of a wrong concept.

[01:18:53]  Bruce: So it depends on what you mean by predictability. If what you mean by predictability is predictable the first time, knowledge creation is 100% non-predictable. Stephen Wolfram has a theory around this as to why that is. It actually is fairly simple. It's that the universe runs at its speed. So for me to simulate our universe would require the entire, first of all, would require an entire universe, a computer size, a computer the size of the universe to do, and probably larger than the universe to be able to do. And then it would still run at the laws of physics. So you cannot actually build something that's going to predict the regular world out there running because your computer exists within the world. So imagine, let's use the same thought experiment. We're on the faster computer, and Bob invents a computer that he's going to simulate his virtual world on.

[01:19:49]  Bruce: First of all, this may be physically impossible for Bob, because by definition his world must have less memory available than the one that he's running on. But let's get around that. Let's imagine that there's virtual memory that exists, and we're going to intentionally increase the memory as Bob does this so that he's able to make a simulation inside of his simulation. So Bob has figured out what the starting point for his world was, and maybe he's figured out he lives inside of a virtual world, and he decides to simulate his own world. Could Bob run his computer forward in time and find out what's going to happen in his world? Would that be physically possible to happen before it happens in his world? Unless we intentionally gave his simulation more clicks of our computer, it would be impossible because he is living inside of a computer that has its computing power.

[01:20:49]  Bruce: So his simulated computer must, by definition, have less computing power than was available to his world. Does that make sense?

[01:20:56]  Unidentified speaker: Yeah.

[01:20:57]  Peter: And then there's also who's going to make sure this other computer works properly and is predicting this knowledge properly. You would almost need another level of computers and people working on that. I guess the take-home point from this thought experience is that it sort of does make sense to say that knowledge creation is unpredictable.

[01:21:19]  Bruce: Yes, it does. And we get that for free. There is just no way around that. We don't need anything special. We don't need Sabine What's-Her-Name's paper where she tries to imagine some set of deterministic numbers that can't be predicted. It's completely unnecessary. Stephen Wolfram has correctly understood this. Knowledge creation is unpredictable. Period. End of story. If by predictable, you mean the first time it happens. Yes, we can imagine predicting it in the case of two computers, somebody on the outside and one of the two computers being slower. But for the people in that computer, knowledge creation is still unpredictable.

[01:22:01]  Peter: I mean, you would almost need like an infinite regress of Laplace's demons.

[01:22:06]  Unidentified speaker: Yes.

[01:22:07]  Bruce: But let's try to get around it without the infinite regress.

[01:22:10]  Unidentified speaker: Okay.

[01:22:11]  Bruce: So let's say that I've got these two computers running, and I decide I'm going to now allow them to communicate. Now, remember, Bob and World 1 is at year 3, and Bob and World 2 is at year 2. And I suddenly announced to both of them, hey, I'm your creator, and I'm giving you access to communicate with each other. This is exactly the same as David Deutsch's description of time travel. What actually, think through what actually happens now.

[01:22:42]  Unidentified speaker: Okay.

[01:22:43]  Bruce: So Bob in World 3, sorry, Bob in World 1, he's at year 3. And let's say Bob in his world that he got a divorce six months ago. So he goes and he tells Bob in World 2, who's still married, I got a divorce and here is why. Will Bob in World 2 be forced to get a divorce, or will he have free will not to? Stop and think about this, because now we're really getting into what free will actually is.

[01:23:14]  Unidentified speaker: Right.

[01:23:15]  Cameo: Or does he even have the opportunity to understand why the other him got a divorce and realized maybe he doesn't want a divorce and he wants to do different things?

[01:23:26]  Bruce: Let's say he does that. Okay, so based on Bob's programming, he receives this news that he's going to make this mistake and it's going to end in a divorce. So he decides to not make the mistake. Bob in World 2 has free will. The fact that, and why is that? Why is it, even though these are deterministic worlds, why is it that World 2 now diverges? Because it will. It will diverge. He now gets to make a different choice, and World 2 will no longer be the same as World 3 from the moment I open up communication between them. Why would that be? And how do we explain this? There's nothing actually magic about it. It still follows directly from theory. It may sound a little magical, but it's not in the slightest magical.

[01:24:11]  Peter: It's a very compelling thought experiment. I need to spend some time thinking about it.

[01:24:16]  Bruce: I'll give you the answer. The inputs are now different. Deterministic function will give a different output if the inputs are different, right? So the moment you open up communication, that's exactly identical to saying that World 2 no longer has the same original inputs as World 1. So we would expect them to start to diverge, even though they run on a deterministic computer. In fact, that will happen even if I don't put them into communication. Let's say that I just simply start communicating with both of them. My communication with both of them will be slightly different. So the moment I start to communicate with them, we would expect both of those worlds to start to diverge a little, even though they're running on a deterministic computer, because the inputs are now different for each of them. You could make the argument they wouldn't be different if I kept them exactly the same, but that's physically impossible for me to do.

[01:25:14]  Bruce: I'm going to do things at slightly different times. There's going to be slight differences that are going to exist, and those will lead to an unpredictable change in the way World 2 unfolds. The moment I start not having… The only reason why they were following exactly the same path before was because I kept them entirely sealed off from the outside world and that they couldn't communicate with each other. The moment they can start to communicate with the outside world, their inputs change, and now we expect them to start to diverge. And now, collectively, they will be unpredictable from that point forward. Even to me, even to me on the outside. This is actually... So there's several things we've now covered. First of all, the idea that in space time, there's no cause and effect, that can't be true, because these are completely deterministic worlds that have no multiverse.

[01:26:08]  Bruce: And yet they have cause and effects. The fact that we need that for free will cannot be true. Unless you want to argue that for some magic reason, even though these people feel like they have free will, that they actually don't, in which case you're right back to Sam Harris again. And that's going to be just as true for us, if we were to have somebody who could step outside of our universe and our space time. Okay? So free will just simply has nothing to do with the laws of physics, because it's something free will in the Deutsch sense follows from having the right kind of algorithm, the creative algorithm, being an AGI or being a general intelligence. And that will be true. And you can run it on a computer, and a computer is divorced from physics, because it's a different layer of abstraction. So physics plays no role in free will, period, end of story.

[01:27:00]  Bruce: Free will is a computational concept that does not matter what the laws of physics actually are. Well, when I say that, I mean, barring them being something totally strange like Sadia wants them to be, I just mean whether it's Einstein's space-time or quantum physics, it will not make the slightest bit of difference. And there's no reason to believe quantum gravity is going to make the slightest difference either. All right, because these will… Turing machine will still work just because we discovered new laws of physics, right? So if you can write an AGI on a Turing machine, under Sadia's physics, you can't write an AGI on a Turing machine until you understand the correct laws of physics. Under this version, no matter what the laws of physics turn out to be, we're already saying as part of the thought experiment that an AGI is possible. So it will not matter what laws of physics we find from that point forward, it will still be the case that free will was a computational concept, that was an emergent concept.

[01:27:53]  Bruce: Now remember, emergent concepts aren't less real, right? This is one of the things Deutsch is really trying to explain, is that reductionism is wrong, that there is no fundamentality towards our world being more real than the world of this virtual world, right? The fact that free will is an emergent concept, that doesn't make it a less important concept. The fact, and whether we should call it free will or not is still an open discussion, but I'm going with the Deutsch version of it, okay? There is no need to study the multiverse. There is no need to understand what the laws of physics are. There's no need to work out a multiverse theory of causality. Those might be interesting things to do in their own right, but they will not even slightly, not even a little bit, not even an iota, inform you on AGI, free will, our free will, our concept of causality, etc.

[01:28:44]  Bruce: I should probably note that physics has no concept of causality. This is something that physicists are well aware of. If you're a physicist listening to this program, you're probably mad that I didn't say this prior to this point. The laws of physics have no concept of causality. And the causality is forward and backwards, right? You can, in theory, reverse the laws of physics and the forces work the same way, both backwards and forwards. That's not true for a computer, by the way. Computers are not reversible, unless they're an intentionally reversible computer, which is very, very, very hard to engineer. Computers can't be reversed. So once you get to a certain state in a program, you can't say, well, back up this many steps and do that again. The information to do that is lost at that point. Okay, so that is maybe one difference between us, where we live in a reversible, essentially we live inside a reversible universe and the people in the virtual world don't.

[01:29:39]  Bruce: But that shouldn't make any difference either, because according to this thought experiment, it must be the case that free will, causality, all those things are completely divorced from that. This means causality is an emergent concept. It's a useful emergent concept. It's very important to the concept of explanations, right? That we work out explanations based on causality. But the concept of causality is not related to the concept of physics. And we shouldn't look for causality in physics. We should understand them as being emergent concepts.

[01:30:15]  Peter: Okay, just to summarize, we live in a deterministic universe. The multiverse, many-world interpretations seem to, if anything, support this idea. Knowledge, creation is inherently unpredictable, which I think your thought experiment captured very well. What Deutsch is wrong about, in this chapter at least, or was wrong about maybe, is that the flow of time does make, actually make some sense. It's not nonsense. Whether many-worlds is not as, maybe nothing to do with that. [Wording unclear.] Is space-time is real or is not?

[01:31:03]  Bruce: Well, it depends on what you mean by space-time. If you mean it by space theory [wording unclear], then obviously it has to be false, because we know space theory to be not true, and we expect many-worlds to be a more true account of it. But if you use space-time to mean many-worlds quantum physics, which you could, it's got its own version of space-time, then it’s free thing… [ending unclear]

[01:31:24]  Unidentified speaker: Okay.

[01:31:25]  Peter: Then free will is, again, depends on how you define it, but I think you've made the case there's some pretty reasonable ways to define free will that you can meaningfully talk about humans having it. I think I agree with that, but none of this necessarily is dependent on the multiverse. So that's kind of like where Deutsch himself seems to feel he got it wrong and fabric of reality is trying to tie it into the many-worlds.

[01:32:01]  Unidentified speaker: Yes.

[01:32:02]  Bruce: I think that’s it [ending uncertain]. I think Deutsch is trying to figure out if there's a relationship, and what I'm trying to say is we actually already know there's no relationship. You just have to understand that that is already an implication of his own theories. So that I don't think he's caught yet, or if he has, he hasn't published anything about it. But I do think this thought experiment makes it quite simple. Maybe we can talk about, since I have brought it up, what are things that people have suggested to me as ways to get around it? And let me maybe try to [steelman] each of their arguments and explain why they're wrong. So one of them was, it's a mystery, we don't know why, but those two worlds, even though they run on a [deterministic] computer, would diverge. I've had people argue that. My answer to that would be, you have now entirely left the theory behind.

[01:32:55]  Bruce: You're violating it entirely, you're basically saying, I'm no longer accepting the four strands theories. Okay, so that's, if that were to be true, we would have a mystery to solve that would mean that everything we thought we knew about physics and about computation is wrong. So that one I can discard on those grounds. I had somebody argue to me that an error might take place and that maybe errors are necessary for divergence. You've now introduced the idea that it's impossible to have free will without errors. I don't even see how this makes sense. I could just simulate an error on a Turing machine because everything is simulatable. So if I need an error to have free will, then I would simulate an error and then I would redo the thought experiment complete with a deterministic simulation of errors and nothing would change in my thought experiment.

[01:33:49]  Bruce: Plus, I can always just change the thought experiment to say, well, errors like that don't happen very often. They don't happen probably within years. So let's only run the thought experiment until the first error happens. And everything I said is still true. So it does not change the thought experiment in the slightest and it doesn't change any of the conclusions in the slightest. I had one person argue to me that I was assuming too much because real people in the real world were fungible because their minds were fungible. And I mentioned this in one of the previous podcasts. And that I wasn't taking that into consideration. Now, first of all, that is a thorough misunderstanding of the concept of fungibility. The person who said this very clearly did not understand quantum physics or understand what fungibility role plays in it at all. Because it does not bring a thing to the table here.

[01:34:46]  Bruce: As I mentioned, they were taking their hopes and dreams and they were stuffing it into a superstition that they were calling fungibility and really not much else was going on.

[01:34:55]  Peter: So they were saying the minds existing in the two different computers might as well be the same?

[01:35:06]  Bruce: No, no, no.

[01:35:08]  Peter: Or were fungible?

[01:35:09]  Bruce: What's that? What they were saying was nonsense. So it's unclear what the point was.

[01:35:14]  Unidentified speaker: Oh, okay.

[01:35:15]  Bruce: So they were saying, you don't know this. And then they were trying to reference what they thought was a mystery, the fungibility of minds. So let's try to take the thought experiment seriously, even though I'm pretty sure it was nonsense. Let's say that you had to have fungibility using quantum physics to have a mind. Well, one of the implications there is that you can't run an AGI on a normal Turing machine, that it's going to require a quantum computer. That's an unbreakable implication of that statement if we take it seriously. So I would just at that point redo the thought experiment, because I'm always willing to change the thought experiment to match things like this. And I'd say, okay, these are now quantum computers, and quantum computers are still deterministic. And so now we can write these minds and these AGI's, they're on quantum computers, because now you're claiming we need a quantum computer.

[01:36:09]  Bruce: For some reason, you're not specifying why, it's just a mystery that we need a quantum computer to be able to run a mind, to create an AGI. So I've got these two computers, one slower than the other. Everything else stays the same. It does not change a thing to try to reference quantum physics or fungibility or anything else. They weren't even worried about the implications. They were just trying to find some way to say, oh, you've missed something. And they weren't trying to think it through, is the honest truth. Furthermore, this shouldn't make any difference anyhow, because in theory, we can simulate a quantum computer on a regular computer. So now, in terms of tractability, this would never work in real life, right? You would need an awfully large computer to be able to do this. But this is just a thought experiment. So we're going to imagine an infinite-sized universe, and we're going to make an infinite-sized computer.

[01:37:06]  Bruce: And how large it needs to be, it can be finite, but it needs to be large enough that I can simulate a quantum computer that's large enough to run these virtual worlds. And now I'm going to take these two large classical Turing machines, and they're going to be simulating, you know, two large enough quantum Turing machines, and they're going to create these people. Now, again, this all follows from Deutsch's theories. They’re [unclear: “Lockhart”] implications. So if we need fungibility in quantum physics for some reason, I can still do this on a Turing machine. I just need a large enough Turing machine. And then you could say, well, it's really slow. Okay. And then it runs long enough that, you know, that eventually, from the point of view of the person living inside the computer, it doesn't matter how slow or fast the computer runs, they experience their life as they experience their life, right, from their moment to moment.

[01:37:58]  Bruce: And so nothing changes. There is not a thing that changes by referencing quantum physics, by demanding that you need quantum physics to run minds, by referencing fungibility. Nothing changes in the thought experiment by doing this. I actually brought this up with Chiara Marletto, and her answer was, I don't remember, so I might get it a little bit wrong, that what we don't know for sure, ultimately the laws of physics will determine if it's possible to do an experiment like this. Okay, that's exactly identical to saying universality isn't actually true, which is an abandonment of part of the basics of Four Strands Theory. Now again, that would be a true statement, right? Anytime you want to say, well, something in the existing theory is false, you get to claim the thought experiment isn't going to work. Okay, but the whole point of thought experiment was to see if you could, what happened if you took the Four Strands Theory seriously, and you didn't violate any of the theories.

[01:38:55]  Bruce: Okay, so that doesn't work either. There is no way around this thought experiment, as long as you're taking Four Strands Theory seriously. There just is no way around it. This is a lock solid, undeniable implication of Four Strands Theory, in its current form, at least. And so based on that, now again, maybe Four Strands Theory is wrong. That could be true, in which case every conclusion I've given you is false. But, or it could be false, I should say. It still might be true, depending on what's different about this new version of this theory we're imagining. The problem is, and this gets back to the discussion about me and Sadia, we can only really work on what we currently have theories of. Yes, I can vaguely reference, there might be some other theory that I don't have today, and that I can't tell you what it is because I don't have it today.

[01:39:48]  Bruce: And under this theory that I can't describe to you, it's going to turn out to be that you can only build an AGI inside a virtual world in such a way that they're completely unpredictable and they can't both work, which would imply you can't build them on a Turing machine, by the way, or really even build them on a quantum computer. It would imply that you need something else entirely because a Turing machine can simulate a quantum computer. So yes, you can reference that to me, which is what Sadia is doing. And yes, I won't deny that's a possibility. But that was never my point to begin with. My point was, what are the implications of the current theory? This is one of them. It is that there should be no connection between free will and physics.

[01:40:31]  Peter: That's a compelling way of putting it. No connection. You've made your case well.

[01:40:44]  Bruce: So that's really what I wanted to cover today. I usually try to find some things that I agree with the person I'm disagreeing with over. So I feel like we're completely leaving out most of what Deutsch actually says. And that's fine because I was trying to concentrate on what I thought we could, was the interesting divergent [wording unclear]. However, he does a pretty interesting job of describing what time would look like in quantum theory. And there's actually a lot there. And if you understand him as saying the flow of time is not real from a certain point of view, not from Bruce's way of defining the flow of time, I still think what he's saying is true. And I do feel like there's maybe a legitimate mystery to understand why we feel the way we do about causality when it has to be largely divorced from physics. Physics is only tangentially connected to it, in that you need physics to create these sorts of universes.

[01:41:52]  Bruce: And from a certain point of view, yes, there's no causality in a block universe. But another point of view there is, because causality is really something about somebody living inside that block time and what the world is like to them. And it's an emergent concept that comes from it. By the way, since causality is closely linked to explanations, this may be a hint about AGI, that explanations are themselves a type of computation. That makes sense. AGI have to run as a computation. And that they're a computational concept that we can work out. There is such a thing as the computational description of what an explanation is, which means we should have a theory of explanations. There should be a theory of explanations. I know Deutsch has resisted that a little. I don't think he strongly resists it. But he said, well, since explanations can, you know, anything can work as an explanation.

[01:42:48]  Bruce: You have to always be open to what can work as an explanation. You can't say an explanation is this and it's not that, because somebody might surprise you. An example, this is the halting problem. That we had Hilbert who tried to say, hey, we're going to someday have a mathematics that shows that everything, we can prove everything is true. And then we had Gödel come along and do his proof, which is the same as the halting problem. That's why I keep saying the halting problem, but I should be saying Gödel's theorem. And he showed through his theorem that that was impossible. Well, the way he showed that was by using math in a totally new way that was outside the way math normally got used, and yet it still worked as an explanation. So this leaves a bit of a mystery as to what an explanation is, and yet explanations have to be a computational concept.

[01:43:44]  Bruce: So there should be a theory of explanation. And I actually think that's, if you want to study AGI, I think that's probably the starting point, is to try to figure out what is an explanation. I've mentioned a few times that there is a type of machine learning called explanation-based learning. And that, first of all, that's really interesting, since explanations are the basis for [possible restart: “computation, for”] universal explainers, according to the current theory anyhow. It makes sense that we would, that explanation-based learning should be a very productive branch, but it hasn't been. It's really just been dwarfed in its successes compared to just the regular probabilistic neural nets, which is what creates, you know, GPT chat and all the huge breakthroughs had nothing to do with explanation-based learning, which is, by the way, one of the ways I know they must be on the wrong path towards AGI.

[01:44:42]  Bruce: There's probably a lot of interesting progress that could be made in the area of explanation-based learning, if we could refine the concept of an explanation. But nobody really knows that, right? This is a case where if you're a machine learning person and you're studying machine intelligence or artificial intelligence, using Four Strands theory, I just gave you a hint of where to go look, what you should be researching, right? But nobody knows that. That's not something that has cross-pollinated enough yet. So I think that's an example of how Four Strands theories should be able to inform artificial intelligence research, at least by narrowing where you need to go do your search.

[01:45:31]  Peter: It seems to me that Deutsch's definition of explanation is quite far-reaching. I mean, I can see why he would be hesitant to try to put a definition on that, as far as I can tell.

[01:45:46]  Bruce: Well, I have no idea how you do it. If I were to define an explanation for you today, what I would actually do is I would define it exactly the way that Judea Pearl defines causality. I would use a graph and I would work out causalities. Because that's the only thing that I can currently think of when I think of explanations. When I try to say, okay, this explains that, I almost immediately my mind goes to causalities. And there's already a theory of how to model causality. So that's what I would try to use, right? It's maybe all wrong, but that's about as far as I can get it at this point. I've been wanting to get that thought experiment out there because I feel like it enlightens so much about this that people are stumbling over. And it just sort of cuts it all away. And you start to realize that free will has a limit in what it can mean.

[01:46:43]  Bruce: It has a limit in how it can be connected to physics. It immediately reduces the search, if that makes any sense.

[01:46:50]  Peter: It's kind of a riff on Laplace's demon, right? I'm kind of taking in a bit of another direction, right?

[01:46:59]  Bruce: So Laplace's demon is correct, right? In a deterministic universe, in theory, that would be correct. I mean, the reason why Laplace's demon doesn't work is because of Wolfram's theory, that whatever you're going to use to make the prediction has to exist within your universe, and therefore has to be slower than your universe. The reason why I was able to get away with it with the virtual worlds is because they're both simulations and I exist on the outside. The key thing here is that this whole thought experiment only works as long as I never communicate with the people inside the virtual world so that they don't know about me, and so that they can't communicate with each other and everything stays the same. It has to be closed. It has to be entirely closed off. The moment there's any extra input from the outside world, things will diverge.

[01:47:50]  Bruce: You might imagine a thought experiment like this. Let's say that you get to year 100 with both these thought experiments, and then the Bob in World 1 discovers he lives inside of a computer, and he even happens to figure out that by doing certain things, he can determine something about the outside world. I don't know how you would do that from inside a computer, but maybe you could imagine that he figured out that this in his simulation is actually tied to the registers in the CPU. He somehow has discovered that he didn't know before that he actually has this ability to know something about the outside world. So he starts to take measurements about the outside world. Well, of course, that's now a new input. So the worlds will start to diverge from this point. They can't be the same anymore. This is actually one of the main things about just our own brains.

[01:48:51]  Bruce: We try to imagine, oh no, I'm a deterministic, my brain is deterministic, that means I have no free will. And there was even a story a science fiction writer wrote where you had this device, and if you tried to, anytime you tried to push the button, it would light up before you pushed the button so that you knew it had figured out what your intention was before you did it. And so everybody realizes they have no free will and they get depressed. First of all, I suspect we could make such a device. It wouldn't be that hard.

[01:49:24]  Peter: Well, I think it's more than a science fiction story. They've actually performed experiments like that, right?

[01:49:30]  Unidentified speaker: Yes.

[01:49:30]  Peter: Neuroscience.

[01:49:32]  Unidentified speaker: Yes.

[01:49:33]  Bruce: I think he was basing it on that. He was trying to make it so everybody gets a chance to experience it. And first of all, it wouldn't be that hard. Let's say you had this device, had a little MRI device in it, and it's scanning your brain, and it detects you're about to touch it so it lights up. That didn't impinge on my free will in any way. And so you say, actually this device would not in any way make me feel like I don't have free will, if you understood it. So now you say, what the device is really doing is it's running your brain algorithm, and it's figuring out what you're going to do before you do it. And it's a little bit faster than you. Well, that's not possible because the inputs of that device and the inputs to me are not the same. Okay, if you made an exact clone of me, and you put that person in the world, they were me up until the clone took place.

[01:50:27]  Bruce: But from that point forward, we're two different people, and we start to diverge because our inputs are different. This all follows from determinism. So there was never an issue here. That device is either, yes, you can make it and it's no big deal, or no, you can't make it. It's a physical impossibility. Those are the two possible outcomes. There are no other outcomes possible under current theory, right? So a lot of these things that people are worried about, they're just nothing, they're just literally nothing to worry about. There are implications though, but Sam Harris makes a big deal out of the fact that the concept of deserts doesn't make sense. He's right about that. So one of our concepts, deserts, the idea that you punish people because they, quote, deserve it. Punishment does make sense. There's arguments to be made that, good arguments, correct arguments, that even if we're all deterministic, that you still have to have penalties.

[01:51:26]  Bruce: This is what Daniel Dennett argues, okay? Because the system doesn't work without them. But the concept of deserts is wrong, according to Four Strands Theory. And that is the part of free will that I think people have a hard time giving up on. And this is actually something that I've noticed, right? When people try to cite the Deutsch version of free will, they'll say it's because I'm a knowledge creator, and I'm unpredictable, and knowledge creation is unpredictable, I can make new choices for myself. None of that will bring back deserts as a coherent concept. And it's easy to use the thought experiment I just gave to show why, right? That's all true for these beings living inside this virtual world. And yet, from a certain point of view, they're just a program that's unfolding, right? There's no libertarian free choice going on, and so therefore they deserve it [as spoken].

[01:52:21]  Bruce: So the concept of desert is what we're starting to lose, because the implications of science is causing us to realize our penal system has this giant thread of deserts in it, and deserts is an incoherent concept, and we're removing it. And it's taking a long time, and it will be centuries before it's totally gone, I suspect. But I think we will eventually end up with a penal system. It will still be a penal system, but it will be built around, built without the concept of people deserving things. I actually think that will be a better world, right? I don't think that's a negative, like, at all. There'll still be consequences for actions. The people are worried there'll be no consequences for actions. That will still be there. Like, that will never go away, right? But I don't know what that would be like. Like, I'd be guessing what the future is going to look like, right?

[01:53:10]  Bruce: But you could imagine some future where, when we put people in jail, that we make their lives as comfortable as possible until we have a way to cure them. You know, or something along those lines.

[01:53:23]  Peter: In some ways, it's kind of similar to how we look at animals. I mean, you kind of, sometimes you got to put down a pit bull or something, but you don't really blame the pit bull. I mean, it could be seen as a sort of a dehumanizing thing, I guess, from that perspective. And you could look at it both ways.

[01:53:43]  Bruce: So there's other interesting things that do come from this. It's the fact that we're slowly starting to accept that people with certain disabilities shouldn't be held responsible, quote, quote, responsible for their actions. Now, of course, by that, what we mean is they don't deserve anything, but we still give them consequences for their actions. It's just that we have a very different attitude towards it. Do you remember when you said, maybe we're all disabled in our own way? That's what I hope for this. [Wording unclear.]

[01:54:14]  Peter: Well, I think it's that being in education, that for various reasons, we've moved pretty far in that direction. And there may be some pluses and minuses there, I'll say that.

[01:54:27]  Bruce: It's something we're going to have to work out over time. I mean, there's a lot of things that I just don't know how this is going to play out. It doesn’t sound negative to me, though. I've seen people get scared over these things. And it seems to me that the fears of the deterministic nightmare that Chiara writes about, or even the fact that Deutsch wrote Chapter 11 of Fabric of Reality, was all based around fears of free will disappearing and trying to salvage it. Specifically, the incoherent version of free will disappearing and trying to salvage the parts of it that they like. And really, I just don't think it's going to work, right? Yes, we can talk about compatibilist free will, but it's exactly and no more of the Deutsch version. It doesn't carry the same moral implications of the non-compatibilist version of free will. And if what you're trying to do is save those, which is what I think people are trying to do, I think you're wasting your time.

[01:55:34]  Bruce: I think you've got an impossible task.

[01:55:38]  Peter: To save the non-compatibilist version of free will.

[01:55:41]  Bruce: I think people who are trying to salvage parts of it would admit that the non-compatibilist version isn't real, but they're still trying to salvage [deserts]. They're still trying to salvage. You have to be seen as responsible for your action in the sense of you get a [desert] if you do something wrong. That's what we're going to lose. There is no way to salvage that part. That is part of the in-coherent version of free will. We need to start concentrating on the coherent version of free will. Maybe we should stop calling it free will. Maybe Sam Harris is right. Maybe we should just call it will. We have wills. This is a suggestion from Douglas Hofstadter. As he said, [quotation wording uncertain:] I call it free will. I call it will.

[01:56:26]  Peter: That makes a heck of a lot of sense.

[01:56:29]  Bruce: So it captures a different... The word free will so strongly is associated with the in-coherent version that, yes, I can say, did you sign this of your own free will and choice? Maybe we should just say, did you sign this of your own will and choice? Because now we're not even indirectly hinting at the in-coherent version, right? I don't know. This is all language. It doesn't matter what we call it. It does matter from a political standpoint. If the word free will is dragging philosophical baggage that we don't need, we should get rid of it. But from a certain point of view, they're just labels, right? It's the concept that matters, not the word.

[01:57:11]  Unidentified speaker: Well, thank you for everything, Bruce.

[01:57:13]  Unidentified speaker: That was real nice.

[01:57:13]  Cameo: Plus, that last end might have been a better end than the other end.

[01:57:18]  Bruce: Cut out the previous end and put this new end in. The Theory of Anything podcast could use your help. We have a small but loyal audience, and we'd like to get the word out about the podcast to others so others can enjoy it as well. To the best of our knowledge, we're the only podcast that covers all four strands of David Deutsch's philosophy, as well as other interesting subjects. If you're enjoying this podcast, please give us a five-star rating on Apple podcasts. This can usually be done right inside your podcast player, or you can Google The Theory of Anything podcast Apple or something like that. Some players have their own rating system, and giving us a five-star rating on any rating system would be helpful. If you enjoy a particular episode, please consider tweeting about us or linking to us on Facebook or other social media to help get the word out.

[01:58:07]  Bruce: If you are interested in financially supporting the podcast, we have two ways to do that. The first is via our podcast host site, Anchor. Just go to anchor.fm slash four dash strands, F-O-U-R dash S-T-R-A-N-D-S. There's a support button available that allows you to do reoccurring donations. If you want to make a one time donation, go to our blog, which is fourstrands.org. There is a donation button there that uses PayPal. Thank you.