AMA | April 2023

0:00:00.7 Sean Carroll: Hello everyone. Welcome to the Mindscape Podcast. I'm your host, Sean Carroll. This is the April 2023 Ask Me Anything edition of Mindscape. And I thought I would start out by telling a somewhat amusing story about last month's AMA episode, the March 2023 episode. You know, I start off each episode with an introduction. If it's a regular podcast, I'll talk about the guest and the topic. If it's an AMA, I'll just ramble on for a little bit, maybe with some information about what I'm doing lately or whatever. So, last month I was getting, you know, still in the mode of having moved to Baltimore, having moved to Johns Hopkins. There wasn't a lot of big news to talk about. So I talked about the fact that we had had a very mild winter in Baltimore, almost no snow. I've never lived in Baltimore over the winter before, but I have lived in Philadelphia.

0:00:52.8 SC: It's not that different. And I use this to make a completely banal observation about human psychology. You know, as we know, we are facing a crisis of global warming caused by human beings. The climate is changing. This is going to be bad in various ways. Now, the fact that we'd had a mild winter in Baltimore this one year, is not strong evidence in favor of global warming. It is some evidence, it's one little tiny data point, but you need a lot more data than that. It could just be a random fluctuation. That's not a big deal. I was not by any means making the claim that the fact that we hadn't had any snow in Baltimore is evidence for anthropogenic climate change. I don't think we need more evidence for that. We just had a new IPCC report come out. It's incredibly, overwhelmingly strong, the evidence that climate change is real. It's being caused by human beings. The effects will be disastrous.

0:01:50.3 SC: The point I was making was one about human psychology, namely, that you can know intellectually that climate change is happening and you can know that it's bad and you can be in favor of all sorts of ameliorating programs and so forth, but when your winter is warm, you think about it in a different way. It sort of hits home even if, again, you know, completely well, intellectually, it could just be a random fluctuation. Like I said, a very mild banal observation. That was the intro to the podcast last month. And as many of you know, I put these podcasts in addition to the usual audio feeds all over the place that you can find the podcast. I put them on YouTube as well. There's no video, it's just a still image with me and the guest talking.

0:02:38.6 SC: But some people like watching on video. I don't mind doing it. I lose money doing it, 'cause I don't get the ad revenue and it's work, 'cause I have to make the video, but I do want to spread the podcast as widely as I can. Usually it's not that many people who prefer to listen on YouTube. So, that's okay. It doesn't really make a difference. I might get 5,000 listeners or whatever. The interesting thing about YouTube though is you can go viral on YouTube, like, you can with a tweet or with something like that on various other social media, because you're watching a screen while you're on YouTube and there's all these suggestions next to you. And if you finish a video, it'll suggest more videos. This doesn't happen interestingly, for regular audio podcasts, because people listen to podcasts, they're driving their car or they're on a walk or whatever with their earphones.

0:03:26.4 SC: They're not watching a screen most of the time, if they're watching a purely audio listening to a purely audio podcast. So there's no way to get that recommendation, right, to get the suggestion to go somewhere else. So, usually for my podcasts, I have a very, very stable audience. The number of people listening is more or less the same. It fluctuates plus or minus 10% depending on who the guest is, what the topic is or whatever. But it's basically the people who subscribe to the Mindscape feed, listen to Mindscape. On YouTube that's very different. So rarely, but occasionally, something can take off if it pushes a button. Last month's AMA episode took off, which is hilarious, because it's a three hour long Ask Me Anything episode about all sorts of crazy topics. But that three or four minute intro somehow got people's attention. So instead of 5,000 listens to that YouTube video, like I would usually get, it's like, 500,000 listeners, which is absurd.

0:04:29.7 SC: And guess what? They didn't like it. [chuckle] So, somehow the YouTube algorithm was pushing my AMA episode into the feeds, into the attention span of people who are very against the idea of climate change and that human beings are in charge of it. The climate denialists, somehow got pushed this AMA episode. And you can tell, not just 'cause the numbers are much bigger, but they leave comments. You know, you can read the comments. You can't read them now, you can't, 'cause I delete the the worst ones, but there were some bad ones, let's just put it that way. I leave the benign ones up there, even if they're wrong. But the completely, crazy pants ones, I will just go ahead and delete. And over, like, I got much worse reviews on that podcast than anything else that I've ever put on YouTube.

0:05:20.4 SC: I don't know why, I don't know why that got pushed to these people or why they clicked on it. Right? There's nothing in the title, nothing in the description of the show notes that even mentions climate change. Somehow it must be... So there's only two possibilities. One is that somehow it got linked somewhere else, but... Yeah, that I can't keep track, but I suspect that the YouTube recommendations are not just generated by the text or the description or the title, but the actual audio, right? That YouTube listens to your audio, transcribes it somehow and uses that to make recommendations. And so, somehow... I'm just guessing this could be completely wrong. Somehow YouTube thought that my podcast was about climate change. And so, I got all sorts of comments about how... You know, you can see what kind of people it is.

0:06:12.4 SC: Oh, I bet you believe that the earth is round too. I bet you got vaccinated. I bet you think black holes are real. [chuckle] So yes, guilty on all of those accusations. Sorry about that. Anyway, I thought that it was really interesting, because not only it's illuminating about the YouTube algorithm, but it's illuminating about, as we all know, we live in a hyper-specialized, localized information ecosystem where you can choose to hear what you want to hear. And it wasn't that long ago, you know, 10 or 20 years ago, there is what I would call a semi-respectable debate about climate change. All the respectable people knew that climate change was real and that it was problematic. But there were plenty of sort of semi-respectable people who were trying to raise problems with it.

0:06:58.2 SC: These days, the semi-respectable people will argue either that it's too late, you know, the climate has already changed too much, we can't do anything about it. Or that, yes, it's changing and it's our fault, but it just costs too much to do something about it. That's where the semi-respectable debate has shifted to. Just like there was a semi-respectable debate 20 or 30 years ago about intelligent design, right? I mean, there's the respectable debate and then the semi-respectable one. Today, There's not that many people, even in the semi-respectable sphere who deny the reality of climate change or that human beings cause it. So, since that's, you know, where I get my information from the union of the respectable and semi-respectable, I'm more or less not paying attention to the completely disrespectful, disreputable, whatever you wanna call it, spheres out there.

0:07:45.0 SC: And I forget that they're there. There's a whole large number of people who are really very, very committed as an important part of their self-image to denying that the climate is changing and that human beings are causing it. So, it's educational for me anyway. I'm just trying to say the words climate change and denial and flat earth enough in this intro, so we can do the experiment again. See if it is once again fed to people in their YouTube algorithms. I really, I really don't know one way or the other. But other than that, we actually have a lot of good questions in this month's AMA and I mean that, like, it was hard for me this month to not answer all the questions. I mean, it's easy for me to not answer all the questions 'cause it would take me six hours.

0:08:30.3 SC: But they were all really, really good questions. You folks are getting good at asking these questions. By you folks, I of course mean, people who support the podcast on Patreon. If you go to patreon.com/seanmcarroll, you can sign up to support Mindscape a dollar an episode, $2 an episode, or you know, pledge $100 an episode. I don't care, I will not stop you from doing that. If you do pledge any amount, you will get to offer questions every month for the AMA episodes, as well as getting ad free versions of the podcast that you can listen to. And also, it's nice and I very, very much appreciate the support. Thanks for all of you folks out there in Patreon land. And I will also offer up apologies. I am going to hope and try that the audio quality for this episode is good, but I'm on the road now. I'm not at home with my usual setup, with my usual nice microphones and non-echoey room and so forth. So I do my best. I try very hard to make the audio quality good, but the mobile podcast unit is not quite as reliable as the home studio. So, we do our best. And with that, climate change, climate change, climate change, let's go.

[music]

0:09:56.2 SC: Our first question is from Kyle Stevens who says, in your last AMA on black holes being the source of dark energy, you stated that papers are often not worth devoting much thought to as they are almost certainly wrong in their conclusions. I assume this paper and others like this are written by credentialed and educated academics. What then leads to the publication of papers such as the black hole, dark energy paper that aren't worth taking seriously? Do the authors themselves not believe in what they're writing? So, I've talked about this kind of thing before, but it's a really important issue and I haven't really succeeded, I think, in articulating how I think about this. You know, very, very briefly, the authors of this paper are certainly believing what they're saying. They're being very, very sincere. They're very excited that they got what they think is a new and important result.

0:10:41.4 SC: They're doing it in good faith and earnestly that is absolutely not the problem. And other people might completely agree with them, right? I mean, obviously this paper, a set of papers along these lines by these group of authors got published. So they got through the refereeing process, got through the editors, etcetera. They have colleagues who they've talked to, I'm sure. So, there's nothing intentionally misleading or anything like that going on here. And I wouldn't even say that it's almost certainly wrong. What I would say is, it could be right, I should back up, because maybe you don't even know, maybe depending on what your background here is, you haven't heard about this, but there was a claim in the news and in the scientific literature that black holes can grow in mass relatively faster than we think that they do as the universe expands.

0:11:35.1 SC: So these people did some general relativity, did some calculations, and they said that when you take a black hole and you look at the solution to the metric of spacetime in general relativity, given by Einstein's equation, by itself, the black hole could just sit there and wouldn't grow. But if you try to look at what happens to the black hole in an expanding universe, so as a background in which you've embedded the black hole, you don't just take empty space with nothing in it, but you take an expanding universe like we live in, they claim there is a cosmological coupling. So, the black hole will grow in mass all by itself, because it is living in a universe that is expanding, not because there's an accretion disc or stuff falling into it. And in fact, they say that the expansion of the mass of the black holes goes as the expansion rate of the universe to some power.

0:12:28.5 SC: The scale factor, the size of the universe to some power. And in fact, [chuckle] So here's what we would ordinarily say, like, if you're a regular garden variety conventional cosmologist, you would say the energy contribution to the energy density of the universe from black holes is from the mass of the individual black holes. And we think that the mass remains roughly constant. And just like any other particle, which has a mass, and its mass stays constant, as the universe expands, the total amount of energy in any region of the universe stays fixed. Total number of particles or black holes stays fixed. The total energy per particle or black hole stays fixed. So the total energy stays fixed. And therefore, if you're in a box in three dimensions, it is growing by some factor, the total energy density will go down like the volume of the box going up.

0:13:22.7 SC: So, the scale factor is the linear size of your box. The scale factor cubed is the volume of your box. So we would say that the energy in black holes, roughly speaking, goes as the scale factor to the minus three power. Okay? The total energy stays the same. The energy density goes down as the volume goes up. That's the conventional story. These folks are saying that they've looked at the equations in great detail and there's very complicated equations you can look at. And when you do that, what you find is that the mass of individual black holes increases like the scale factor cubed, coincidentally enough. [chuckle] And so, the total energy in black holes in that region remains constant. Sorry, the total energy density remains constant. The energy goes up. This is just like dark energy is supposed to do, just like the vacuum energy or the cosmo constant is supposed to do.

0:14:21.1 SC: And therefore, they claim that black holes can be the dark energy. And the short version of my claim is just that, these people have not done enough to convince me that this claim is worth looking into. I could very easily be wrong. You know, I have not looked into it very, very carefully. And so the two things I wanna talk about are, number one, why would I not be convinced? And number two, what can someone do to convince me? Like, that's the actual intellectually interesting question. On the general relativity question, you know, I've done general relativity a lot. I wrote a book on it. I have a certain feeling personally for what the equations of general relativity predict in certain different kinds of circumstances. Nothing in anything I have ever learned or thought about would make me think that a black hole embedded in an expanding universe would have its mass just spontaneously increase, like, the scale factor cubed.

0:15:18.6 SC: I have no idea why that would happen. That is completely contrary to everything that I've ever thought about general relativity. Now, that's not, again, to say that I know it's wrong, because I've never sat down and specifically solved these equations, okay? But the people who are writing this paper are not claiming that what happens in our galaxy affects the black holes in our galaxy. They're claiming that what happens in the whole universe affects it. And the thing about general relativity is, the equations are difficult. The equations are non-linear. There are not a lot of exact solutions that you can look at. So, instead what you do is you look at some exact solutions and you look at some approximations and you build up, you look at numerical simulations also, for in spiraling black holes, for example. And you build up some intuition for what seems reasonable and what seems unreasonable.

0:16:11.4 SC: And the claims in this paper just don't seem reasonable to me in terms of what I expect general relativity to be predicting. So that's my general relativity part of the story. It goes beyond that, because if it were true the black holes were gaining in mass, you know, black holes aren't fixed. They have motions through space, right? And so, in fact they pull together each other. They have a gravitational attractive force. So you'd expect that these black holes would fall into galaxies and into clusters of galaxies, and we would notice that there was a huge amount of mass in the form of black holes. And I don't mean by like, microlensing or individually looking at the black holes themselves, but there's three times as much energy in the form of dark energy as there is in ordinary matter and dark matter. So, I would expect that this whole prediction of this model says that galaxies and clusters of galaxies are actually three times heavier or four times heavier than we would expect.

0:17:11.8 SC: And that's not true. We've done the observations and that's not true. So they have to come up with another story in which the black holes actually repel each other and they leave the galaxies. And that's why we haven't noticed them there. Again, that is entirely in conflict with everything I have in my intuition about how general relativity works. Having said all that, I could be wrong, right? So, how would they convince me to read the paper and think that maybe they were right? 'Cause there's a lot of papers that come out. You know, it's easy to say, well, if the claim is right, then it would be hugely, hugely important, therefore, you should spend time looking at it. But that logic doesn't actually work, because that's true for a lot of claims, a lot of crazy claims. Might be, might have the feature that if they were right, it would be really important.

0:18:02.2 SC: But that doesn't mean I'm gonna spend time thinking about them. There's not enough time. You gotta pick and choose. You have to do some triage here. And it can be difficult. This is, you know, something that is important now for other reasons that have nothing to do with black holes and dark energy, which is why I wanna spend some time on it. Because there's this whole fad now for doing your own research when it comes to claims that you don't agree with or you think that the mainstream media is perhaps misleading you in some way. The establishment is hiding the truth from you, so you're gonna do your own research. And the people who like to do their own research, I mean, for one thing, they're not actually out there performing randomized controlled trials, [chuckle] they're not collecting data, they're not acting like the actual scientists.

0:18:44.6 SC: They're reading other people's work and they're drawing their own conclusions on the basis of them. And my point is that, that process of reading other people's scientific research in an area that you are not an expert in, is a lot harder than you might think, because scientists disagree with each other. Scientists can be wrong. So it's not that, you know, these people who wrote these papers are disbelieving their own result or that everyone disbelieves them. It's that, I just don't think that the probability that they're right is very high. And so, I wanna give you a little bit of the insight as to why I came to that conclusion, which might be relevant for other people doing their own research. So the point is that, you do have some expectations, right? None of us enters the world of trying to figure out science or the way the universe works completely as a blank slate.

0:19:36.8 SC: We have a picture of how the universe works. We have priors, we have propositions to each, we assign credences that we say, oh yeah, this is pretty likely, this is not very likely, and so on. And then, we have evidence that comes in and as good bayesians we update. So, when you have a claim in a scientific paper that goes entirely against everything the reader might have expected to be true, it is your job as the author to do two things. Number one, indicate that you get it. That you know that this sounds very different than what people expected and list the reasons why people might not believe it. Okay? So, you're not giving off the impression that you just solved some equations and are taking the answers at face value. Because like I said, in general relativity, the equations are hard to solve exactly.

0:20:28.5 SC: You have to interpret sometimes, you know, what effects are important, what effects can be neglected, what are the boundary conditions? And it's all very, very important and difficult and what makes some people good at doing general relativity. So, when I read a paper like this, I'm looking for an acknowledgement that here's what you might expect and here's why we are different. And then, the other thing of course is that, that latter part here is why your expectations are not going to come true in this particular case. Here is why your intuitions are off. And I didn't see that in these papers. [chuckle] All they're telling me is, we've solved some equations, we got this result. Isn't that awesome? And I think that, you know, you need more than that. I'll give you two examples, and very, very opposite sides. I don't wanna talk about these people too much, 'cause I think that they are good scientists.

0:21:20.0 SC: They're trying to do their best. I'm just not convinced by their answer. And that's okay. If it weren't for like, the media talking about it and acting like it's very exciting, it would just be the usual process of scientists talking to each other, and that would be great. Maybe I'm wrong and maybe I will eventually be convinced. So, one example is, I wrote a paper with Jackie Lodman saying that energy is not conserved in quantum measurement. And some people say, well, that's just perfectly obvious thing. We knew that all along. But other people say, no, that's completely crazy. How could you believe that? So Jackie and I really tried to say in our paper, here's why you might not believe it, and here's why those reasons don't apply here. I think that's an important part of what you have to do.

0:22:02.7 SC: I'll give you a completely other... And the second example I do with great trepidation, 'cause I don't want to tar people with bad associations, but it sticks in my mind, The Bell Curve. Remember The Bell Curve? You know that book, Charles Murray and James Herrnstein from the '80s, and it came out in the '80s and it was talking about a distribution of IQ etcetera, and what the implications of that are for public policy. And I'm not an expert in this stuff and I was just a college student when it came out, but you know, I'm curious, it was a big deal when it came out. People are still talking about it today. And so, one of the... I didn't read the whole book by any stretch of the imagination, but I did look at it.

0:22:42.2 SC: I opened it and paged through. Again, to do this judgment, is this worth my time? Right? That's the question. And one of the very first questions I asked myself was, you know, look, there have been a lot of claims over the years that I can do science and show that some people, some racial groups are dumber than others, right? And these claims have a very bad history of being bad science badly done, because people wanted the results to be true, because it's a bunch of white people saying that black people are dumb. Usually, historically, that's what it's been. And these historical examples have been used to justify some terribly racist things. So, when a book comes along in the 1980s, which purports to be super duper modern and scientific, but yet reaching conclusions that are very similar to conclusions that have been reached a hundred years ago by pseudoscientific nonsense people, what I'm looking for is, once again, an acknowledgement.

0:23:45.8 SC: You might worry that this is just pseudoscientific racism, because that's been... We've had a long history of pseudoscientific racism and here is why we are not that. So, when you open The Bell Curve, you do not find that discussion. In fact, you find that they cite some of the pseudoscientific racism of the past approvingly. And at that point. I said, meh, this is not really gonna be worth my time. And every individual person, every individual scientist is gonna draw those distinctions differently. Again, it's just not enough time in the world. This is a point that Zeynep Tufekci made in our podcast conversation. The most valuable thing that people have to deal with right now in the world is their attention. What do you spend time thinking about? And there are hundreds of papers, scientific papers coming out just in fields that I care about every day.

0:24:42.5 SC: And I'm not gonna read all of them and leaf through them. Sadly, I probably read fewer than I should because I'm busy doing other things. But you have to really think about what is going to be most likely to pay off. And just because a result would be really, really important, if it were true, doesn't mean it's going to be worth reading, 'cause you don't think it's likely to be true. And this can happen with very, very well-known people. I once went to a talk by a very, very well-known physicist, but someone who, you know, bless his heart, was known for exaggerating a little bit. And he gave a usual sort of bombastic exaggerating talk. And afterward, people were just not that excited. And, you know, we were chatting afterward about other things and he said like, why weren't people excited about my talk?

0:25:27.1 SC: I'm claiming all these amazing things. I didn't know what to say. But you know, the truth was like, dude, you claim these amazing things all the time. They never turn out to be true. So we have a filter a little bit. And that's what's hard to get across to people who are trying to do their own research. You know, when you read a paper, you can't just say, oh look, this paper found this result, therefore it's true. Like, everyone knows that. But you know, how do you judge whether it's true? Well, who are the people? What are they saying? How are they framing it? Do they seem reasonable? Do they seem open to understanding why they might be wrong? Or are they more rigid in their ideology and pushing an agenda? Are they a little bit too enthusiastic about their own results? So they're perhaps insufficiently skeptical.

0:26:09.0 SC: There's a lot that goes into it and it's tricky. But anyway, I do not mean in any way to disparage the individual researchers here. I think that they're absolutely doing their best and they might be right. I'm just explaining why they have not yet moved me personally to spend time digging through those long, complicated equations. They've got to explain to me why all of my intuitions about general relativity are being violated in this case. Okay, that was a longer rant. I actually thought about turning this into a whole solo podcast, but I think that's enough. 10 minutes of yammering about this is more than enough. Okay, let's go on to Venkitus Morkvenas. Hope I got that right, who says, let's assume we live in a simulation, then it might be the dark matter and dark energy are just parameters set by simulation runners, e.g., to confuse us.

0:27:02.5 SC: Would it mean, we'd never be able to resolve their mystery? Or on the contrary, this lack of this resolution would prove the simulation hypothesis? This is an even less likely to me, potential hypothesis to explain dark energy than the black holes are. So again, not trying to give Venkitus a hard time here, but you know, yes, it is absolutely conceivable that we live in the simulation. It is absolutely conceivable that things like dark matter and dark energy are just parameters set by simulation runners. In some sense, I think everything about physics and cosmology would qualify as a parameter of the simulators, right? Someone's gotta set these parameters somehow. We usually think that nature does it, but if we live in a simulation, then some simulator did it at, at some point. The idea that they did it to confuse us, I find pretty wildly unrealistic.

0:28:01.8 SC: Like, why do they care about us that much? Like, it took 14 billion years of evolution, of running time, to simulate 14 billion years of evolution even to get to us. You know, are we the only people here in the universe? You know, it seems like a really bizarre game to play. For them, maybe it's all just a few cycles on their super intelligent iPhones for all we know. So maybe it's not a big deal, but it just seems weird to me. More importantly, the reason I thought this was a good question was, what would it mean in terms of being ever able to solve the mysteries of dark matter and dark energy? I don't know if... I don't have an opinion on whether or not simulators could make the parameters of our physical world such that we would never be able to explain them.

0:28:52.8 SC: That's something I honestly don't know. I do... I am a believer that at some level there are brute facts in physics and cosmology. There are just true things that we measure rather than explaining. So, I don't even think that I would ever come to the point where, oh, it must be simulated. It must be like, set by some God-like beings to make it that way. I could just say, oh, that's the way it happens to be in our universe. I'd be perfectly happy with that. In terms of the lack of resolution proving the simulation hypothesis, that I think is very, very unlikely. For one thing, how do you ever know that you cannot explain something, right? We try, like, science tries to explain things, tries to come to better understandings of things, but there's never, you know, an indication that, oh, okay, you have failed perfectly irrevocably, you cannot possibly explain this thing that you're thinking about.

0:29:47.0 SC: That's just... You know, there's never a sign. Like, maybe you just haven't done it yet. And in fact, this is kind of relevant to various questions in modern physics. Like, there are some hard questions. Is it possible that we'll never answer them? Yeah, yeah, it's possible, but probably we just haven't answered them yet. You know, like, be a little bit more patience. So, I would say there's never going to be a point where I would say, oh yeah, we probably live in the simulation, 'cause I don't know what the dark energy is. I think I know what the dark energy is. I think it's the cosmological constant. I don't know why it has the value it does, but we're still working on that problem. It's still very worth exploring. Okay, I'm gonna group together two questions. Apparently, my subconscious has put all the dark matter and dark energy questions here at the beginning.

0:30:30.3 SC: So, Paul Conti asks, given the amount of time and the number of experiments that have tried and so far failed to yield any concrete evidence of dark matter, do you think it is time to give more consideration to the MOND theories of gravity? MOND is Milgram's theory, modified Newtonian dynamics, which purports to change gravity rather than having a dark matter particle. Paul continues, do you believe it is likely that a dark matter particle may still be discovered? And Walter E. Miller says, if it turns out there are no dark matter particles, but that the observational evidence for dark matter is explained by an adjustment of how gravity works, what are the implications for our cosmological models? The value of Lambda and the interpretation of the cosmic microwave background? So again, here's an example where maybe a little more patience is warranted.

0:31:18.3 SC: So, in terms of dark matter, the first thing is, the evidence that there is something in the universe that we would label and understand as dark matter is overwhelmingly strong. And I say this as this is a testimony against interest. I wish it weren't strong. I have worked myself on the idea of modifying gravity. I think that would be cooler than the idea of just another particle being the dark matter. But in terms of science, when you like, sit down and look at the theories and compare them to the data, the idea that dark matter is an independently new propagating thing that adds to the energy of the universe is supported overwhelmingly by the evidence, mostly, most importantly by the cosmic microwave background. And I've talked about this before, so I'm not gonna go into great detail, but the cosmic microwave background and isotropies, the small fluctuations in temperature from place to place are related to fluctuations in density in both dark matter and ordinary matter, if there is dark matter there.

0:32:20.7 SC: And you can make predictions based on your theory of physics for how the different amounts of fluctuation at different scales across the sky should be related to each other. And there's a story that you can tell about the ordinary matter in the early universe, and it oscillates before recombination, before electrons have joined together with protons to make atoms, the universe had a pressure in the fluid, because everything was charged particles and the charged particles would keep bumping into photons, and that creates pressure. So there were literally acoustic waves, literally sound waves, a region of space where the density of matter was slightly higher than average would want to collapse under its gravitational pull. And then, it heats up, because there's pressure and it pushes back and it bounces. Okay? So, you see these fluctuations that have been evolving for a while.

0:33:12.5 SC: On small scales, they have bounced, they have squeezed and bounced and squeezed and bounced again many times. On larger scales, maybe only once or only half a time before, the CMB, the cosmic microwave background was formed. And the thing is, dark matter doesn't bounce, right? Dark matter just collapses toward the regions where the gravity is stronger. So there's a very clear unambiguous prediction in a universe with dark matter versus a universe without dark matter. In a universe without dark matter, these acoustic waves will damp, they will just get slightly softer over time, because, you know, the over dense regions send out photons and neutrinos into under dense regions and vice versa. So you make a prediction for how the different peaks in the power spectrum of the microwave background evolve as a function of scale. With dark matter, there's a very different prediction. There's the same basic story, but now you have a new ingredient, the dark matter, it doesn't bounce.

0:34:16.5 SC: So when the dark matter is in phase with the ordinary matter that reinforces the density fluctuation and therefore the temperature fluctuation, when it is out of phase, it goes against the density fluctuation and smooths out, sort of compensates a little bit for the overheating in the ordinary matter. And so, there's a difference in the first, third, fifth peak of this power spectrum of the CMB versus the second, fourth, sixth, the even numbered peaks odd versus even number peaks. A difference between the odd and even numbered peaks in the power spectrum of the cosmic microwave background is incredibly strong evidence for dark matter. It's evidence that there is a gravitational pull that is being sourced by something other than ordinary matter. And that is exactly what we see in the data. It is incredibly... It was a prediction made ahead of time that came out to be true.

0:35:15.3 SC: And guess what? People who care about modified gravity, also made a prediction and their prediction ahead of time was that you should not see these differences in fluctuations. And that prediction came out wrong when we compared it to the data. So, why do we pay attention to MOND theories at all? The answer is, what do you mean by a modified gravity theory? Generally, when you start with a modified gravity theory, you have to like, write down some actual theory, like Milgram posited a phenomenological model that would work in galaxies. He wasn't trying to explain the microwave background or anything like that. So later, people like Beckenstein and others have actually come up with good old respectable field theories that give you MOND like behavior in galaxies. To do that, you need to invent new fields and those new fields vibrate and propagate and have energy.

0:36:10.7 SC: And so, the people who invent new theories of modified gravity figured out that they can explain the cosmic microwave background data by inventing new fields that propagate and have energy and move around in ways other than the ordinary matter moves around. So that's how they can claim to fit the cosmic microwave background. But guess what? That's just dark matter. They just have dark matter. They have new fields that propagate independently that we haven't yet seen in experiments. That's just what dark matter is. That's all it is. So, I'm totally happy with the idea that gravity is modified. I would be thrilled if that were true. That would be, you know, a playground for theoretical physicists. I would love it. But there's overwhelmingly strong evidence, there's something out there besides ordinary matter that is causing gravity in the universe. So, we haven't found the dark matter particles yet.

0:37:03.4 SC: That's true. What is the probability that we should have found then by now? That's a good question. People sort of don't ask that question as much as they should. They're a little bit less willing to put their money where their mouth is when it comes to these things. But what you can do, and I've done this, is go back, you know, 20 years before we had a lot of the data we have from current experiments and say, what are all the different possibilities? What is the space of parameters that we had in mind for what the dark matter could be and how does it compare with what we've actually looked for and ruled out yet? Very, very, very roughly The answer is that we've ruled out about half of the parameter space for weekly interacting massive particles, which are the most popular dark matter candidate.

0:37:49.6 SC: So that's, you know, half is substantial, but it's not everything. It's not 99%. So it's not like we looked everywhere. The dark matter could be and haven't found it yet. That's completely false. Furthermore, the dark matter might not be weekly interacting massive particles. There's great evidence that there's dark matter, but we don't know what it is. It might just be axions, for example. Those were always the second most popular candidate. And we have not ruled out any substantial region of parameter space for axion dark matter, yet. It's completely possible that the dark matter particle is something we haven't thought of yet. It just needs to be something that moves slowly, that is massive and it doesn't interact with photons very strongly. Now that's really all you need, that's not very hard to come up with models. There have been thousands of models, predicted, and not all of them are easy to experimentally test.

0:38:37.9 SC: The reason why axions and weakly interacting massive particles are so popular is because they are not only the dark matter, they were actually originally proposed to solve completely different problems. And if you can solve two problems with one particle, then that's considered a triumph. So look, the dark matter might just be axions, it might be a weakly interacting massive particle we haven't discovered yet, or it might be something completely different. These are all viable possibilities, and the fact that we haven't discovered them today and might not have discovered them by next month doesn't mean they don't exist. You have to keep thinking. Having said all that, if someone comes up with a better theory, then that will be great. The better theory will almost certainly involve new propagating amounts of energy, because that's what the CMB tells us to believe in.

0:39:23.6 SC: Rue Phillips says, as a fellow physicist, lover of Jane Austen, I would very much like your perspective on the author's genius. Take Pride and Prejudice, for example. Comparing my feelings of reading the book and my feelings after watching the Faithful A&E adaptation, I must say that so much more is infused in the adaptation that Austen does not capture in words, feeling, romance, etcetera. Having read other books of hers, I believe it is her style. She masters the character like no other, but also leaves something wanting. Aside from being a groundbreaking female novelist of her time period, what do you see as her genius and shortcomings? Love this question. I don't get enough chances to rhapsodize about Jane Austen here on the podcast or the AMAs or anywhere else. So, I don't quite agree with your diagnosis there, Rue, but I get it. I get what you are saying. I would just put it slightly differently. So the thing about Jane Austen is, I don't know, maybe many of the people listening have read Jane Austen or are fans. Many probably have not. She's a novelist who you know what the plot is going to be, okay?

[chuckle]

0:40:28.8 SC: You don't read Jane Austen novels for the plot, except that you know it's going to be a happy ending. The female protagonist is going to get married and live happily ever after. So that you're not there for the twists and turns of a Sherlock Holmes or Ms. Marple kind of story. What you maybe are there for, but I guess my point would be what you might be there for comes in different layers. You can absolutely read Jane Austen's novels at a completely surface level and go along with it. Enjoy the fantasy, because as Rue says, she's really a genius at, I mean, Rue says character, but I would put it a little bit differently. What Jane Austen is a genius at is, we all have our romantic adventures or misadventures or experiences or whatever. In retrospect, when you think about what was going through your mind and what you did, when you were infatuated with someone or dating someone or someone, you go, boy, I was really kind of silly back then. I did silly things. I obsessed over silly things. I did not say sensible things. Anyway, if you're like most people, including, certainly me. And Jane Austen is really quite good at going into exactly all of those little obsessive things you worry about when you're thinking about the object of your affection, etcetera.

0:41:53.5 SC: Like, every word you said, every word they said, what does it mean? Do they like me? Do they like, like me? All of these things and you can enjoy that give and take. The specifics of the journey to the happy ending are quite well done and you can enjoy it for that and that's what makes for good movies. But I don't think that's what makes her a genius. That's not quite enough. Other people can do that. For me, what makes Jane Austen a genius is that there's a quite substantial gulf between the words on the page and what Jane Austen wants you to think. And there is enjoyment in figuring out what it is that she wants you to think, because it's not like, hopelessly esoteric or hidden. If you put a little effort into it, you can figure out what she wants you to think, but you absolutely need to put that effort in. The most obvious example, you know, there's the famous opening line in Pride and Prejudice where she says, it is a truth universally acknowledged that a single man in possession of a good fortune must be in want of a wife. And I had a friend who read the book and objected to this.

0:43:00.9 SC: No, it's not true that a single man in possession of a good fortune must be in want of a wife. That's a very bad thing to start a novel with. To be fair, this was not a native English speaker and they didn't perceive the sarcasm. Clearly, Jane Austen was not trying to be serious. And this is the most obvious thing. I think most native English speakers would not take that expression in the first line of Pride and Prejudice as serious, okay? But it goes way deeper than that. Just to give a couple of examples in Pride and Prejudice, Elizabeth Bennett, who is the protagonist, her father, Mr. Bennett, is portrayed in part, because we're seeing him through Elizabeth's eyes as a good man. Maybe a little grumpy sometimes, etcetera, but smart and kind compared especially to Elizabeth's mother who is portrayed very badly. But If you actually look at how he behaves, in many ways he's not that great. He's careless. He sort of removes himself from the cares of his family. He does not provide for them. After he dies, they're going to lose the house. They have no special means of support, etcetera.

0:44:12.0 SC: In many ways you could easily say that he is not that great. But the novel never says he's not that great. The novel just shows you him not being that great and lets you figure it out. An even better example, for those of you who have read the book, you know that the reason why Elizabeth and her family are going to lose the house when her father dies is that the will he's going to give it to his cousin, Mr. Collins, the Reverend Mr. Collins. So, Mr. Collins visits with the idea of wooing one of the daughters of Mr. Bennett and then the house would stay in the family. He actually proposes to Elizabeth and she refuses him because he's terrible. He's just a nincompoop. And that's absolutely believable, because Elizabeth is a hopeless romantic and she thinks that she's going to hold out for something better. That all makes sense, but then Elizabeth's best friend, Charlotte Lucas, later gets proposed to by Mr. Collins and accepts his proposal.

0:45:14.8 SC: And from Elizabeth Bennett's point of view, this is crazy. Why would you pledge your life to this terrible, terrible person? You're going to be miserable. And Charlotte explains that this is the best she can do. These are the options that she has. And I think that a lot of people reading the book are like, oh, poor Charlotte, she made a mistake. Elizabeth stuck it out and she eventually got to marry Mr. Darcy, who is much more rich and a wonderful person and a loving husband, etcetera. I don't think that's the full story or anything close to it. I'm not the first person to point this out, but Jane Austen is telling you, sure, I'm going to give you a fantasy about living happily ever after, but I want to remind you that life is not a fantasy for most people.

0:45:57.7 SC: In the real world, it makes perfect sense to do what Charlotte Lucas did, given the position of women in that society in that period of time. Jane Austen is both letting you enjoy the happy story of Lizzie Bennett and Mr. Darcy, but also, not letting you forget that life was not like that for everybody. And This happens just over and over again in many, many different ways in Jane Austen's novels, and that's what I really love about them. In terms of shortcomings, she obviously wrote about a very narrow slice of possible things in the world. I don't even think that's a shortcoming. That's what she did. I think that's good. I don't want every novelist to write about the widest possible landscapes, etcetera. I think that Jane Austen, for what she did, was more or less ideal, so, I'm not gonna talk about her shortcomings there.

0:46:49.9 SC: Charles Grant... I gotta give shorter answer to these questions. Oh my goodness. We're gonna be here forever. Charles E. Grant says, decades ago I was a physics undergraduate. My questions about the philosophical foundations of quantum mechanics were waved away, with Bohr and the Copenhagen School sorted all that out. I tried to read some of Bohr's papers on the subject and found them impenetrable. Recently, I took a course on the philosophy of quantum mechanics. We started by reading the EPR paper, Einstein, Podolsky, Rosen, and it seemed very clear. I then tried to read Bohr's response to the EPR paper. It was still impenetrable to me. I consulted my professor and they sighed and said, yes, Bohr is a very difficult read, and folks widely disagree on what he actually meant. Bohr was obviously a great physicist, but how did he come to be so influential in the foundations of quantum mechanics when his philosophical writings were so difficult and themselves subject to multiple interpretations?

0:47:40.7 SC: Yeah, that's a very good question. I think it's more of a historical question than a physics question. Somehow, the character of Niels Bohr did manage to exert a huge influence over generations of physicists. And if you just read what he wrote, it is a little bit difficult to figure out why... I mean, he's welcome to have his opinions and his opinions could be even correct sometimes. He was clearly a genius physicist. His contributions to quantum mechanics are unquestioned, but why were, like you say, his philosophical writings so influential? And I don't know. If you go back to the podcast I did with David Albert, we talked a little bit about this. And David very amusingly said that if he were able to have dinner with one person from history, it would be Niels Bohr. And the reason why is because the smartest people in the world would go visit Niels Bohr and then they would come away spouting nonsense about the foundations of quantum mechanics. And David Albert was very curious as to how Bohr could exert this power.

0:48:43.1 SC: I think that part of it was that Niels Bohr was a genuinely good person and people really, really liked him. John Wheeler, who talked a lot with Albert Einstein, but worked under Niels Bohr, said that he admired Einstein, but he basically worshipped Niels Bohr. Niels Bohr was a saint as far as John Wheeler was concerned. And I think that there was something in what Bohr said that physicists wanted to hear. They didn't want to worry too much about the possible difficulties of quantum mechanics. They wanted to get on with their lives and just do the calculations, shut up and calculate, and make some theories of particle physics and radioactivity and stuff like that. But honestly, I don't know the details, I don't know, I completely agree that Bohr's actual writings are very, very hard to read. I would not suggest reading Bohr or even reading EPR as your introduction to the foundations of quantum mechanics. There are other places you can go that are much more modern and have sort of figured out what the deep issues are.

0:49:44.8 SC: Ken Wolf says, Johanna Hoffman said some interesting things, had some interesting things to say about the future of cities on a recent episode. But she lost me completely as soon as she said that her ideal city would have at least 20 million people. I am of the opinion that there is a sweet spot of city size that maximizes the available facilities and venues while maximizing the ease of navigating the city. And I think that sweet spot lies between somewhere between half a million and one million people. Based on where you have lived do you have any preferences for city size or thoughts on what would be the ideally sized city to live in? I kind of want to say there's no such thing as an ideally sized city to live in. This is definitely one of those areas where I'd rather let a thousand flowers bloom. A lot of different people have different options, which fortunately we have in the real world. We have this weird, weird thing right now where our biggest cities are making it very hard for people to afford to live there.

0:50:41.6 SC: I am what is labeled Nimby in the discourse. Yes, in my backyard as opposed to Nimbi not in my backyard. I think we need to build a lot more housing. I think it is very, very obvious, that's the best way to make housing more affordable, build more of it. And we don't do that necessities right now, and that is a problem. But putting that aside, I think that people are different. You know in fact, not directly related to your question, Ken, but we did talk about this in the episode with Will Wilkinson on Political Polarization and Urbanization, where he talks about how different personality types as classified by the big five personality inventory are drawn to either live in urban or rural/suburban environments. And I think that's good, that's a feature, not a bug. So, some people are gonna like 20 million person sized cities. The number of options you have for dining at 2:00 AM will be larger in a city of 20 million people than in a city of half a million people.

0:51:41.0 SC: Even a city of half a million people just doesn't quite have a capacity to support a lot of the more esoteric, small-scale idiosyncratic needs or wants or desires that its inhabitants might have. I absolutely appreciate what you're saying as someone who lives in a half-million person city right now, the ease of getting around is just amazing. [chuckle] I like to say that in Baltimore like, every point is 15 minutes away from every other point in Baltimore, and there's like no two points in Los Angeles that are 15 minutes away from each other. So the time it would take me to go from one end of Los Angeles to another would take me two different states when I'm in Baltimore, which is great and wonderful.

0:52:21.1 SC: But there is less going on in Baltimore than there is in Los Angeles. And that's a trade off and that's absolutely fine. There's plenty going on in Baltimore, more than enough to fill my dance card, as it were, but there are people who get out a lot more than I do and maybe want more variety than that, maybe they need a bigger city. That's completely fine with them. There're absolutely by the way economies of scale. I talked about this also with Jeffrey West on one of the original Mindscape episodes. There are things that just can happen in important ways the bigger the city gets. So, whatever individuals might want I think that it is good for the planet to have some cities with 20 million people in them, I think it's a perfectly good goal to have.

0:53:03.1 SC: Sid Huff says, The US Government has been dominated by just two parties for nearly all of its history, whereas countries governed by parliamentary democracy, such as the UK, Canada, France, etcetera, typically have a number of parties of very sizes between four and eight. He recently taught a course at Johns Hopkins on Physics and democracy, does physics shed any light on why the US has just two parties, while most other democracies have had multiple parties? I don't think you need physics actually to address this question. I think the political scientists understand this pretty well, although, there might be and I'm actually thinking about this, so I can give you a full-blown answer right now, but there might be a little bit of insight you can shed by asking exactly that question from the physics point of view. But like I said, there's a perfectly conventional political science point of view, which is that...

0:53:48.9 SC: Well, let's put it in physics terms just because that's what you asked, and it's kind of fun. In physics terms, what we're gonna do to describe the collective behavior of some object with many, many little constituents in it is, invent some emergent description. Right? So I have the table in front of me right here, it's made of a lot of little atoms, but I'm not gonna describe it in terms of atoms, I'm gonna describe it at a higher level by talking about wood of a certain size and tensile strength and things like that. So I invent a whole new category of stuff, table versus atoms to describe this collective behavior. And it's more or less unique that description.

0:54:29.8 SC: I mean, I could describe the table in different ways, but the fact that there is a table in front of me and I'm gonna group a whole bunch of atoms together to call it a table, makes sense in a variety of objectively true ways. I can describe the table using a small number of data points, which gives me a handle over how it will behave and things like that. It's not arbitrary why I chop it up into a table. So in some sense, there's an analogy there with politics and with democracy. You have a bunch of people in your society, in your country or what have you, and they come in to make a collective. They make a country or a state or a city or whatever, and that country or city or collective society makes decisions through a democratic process, which is kind of an emergent process, right?

0:55:17.1 SC: The individual people have their individual preferences and wants and so forth, but you have to somehow figure out how to aggregate those into social preferences. Here's the difference, the difference is that rather than the laws of physics giving us the right way to describe this emergent behavior, we have an extra layer in the political discussion, which is we invent a system of government. Even if you say, Well, we have a democracy, okay? How exactly is your democracy going to work? How are we going to vote, is it gonna be plurality voting or ranked choice voting, or approval voting or some other system? Are we going to vote directly for our representatives or will we vote for people who vote for the representatives? How are we gonna split up power between a legislative branch and the executive branch and the judicial branch and other branches and so forth?

0:56:11.2 SC: So all of these choices come into how we go about aggregating the preferences of the individuals and do what we call the preferences of the societies, and those choices... And here's where we actually answer your question, those choices go to a long way to figuring out how political power is fought for within the society. In particular, the existence of political parties. Very, very roughly speaking to over-simplify quite a bit, the point in the United States that we have a President. In a parliamentary system, you don't directly vote for the chief executive, you vote for your local representative, and then, whichever party gets the most representatives can vote for their leader, the Prime Minister.

0:56:57.4 SC: Again, simplifying a lot here, but that's the basic idea. In that case, you can have multiple parties in the parliamentary system, because they can form coalitions. If you have multiple parties, it will very often be true that no one party has a majority, so no one party can just pick the prime minister. So the different parties have to come into a coalition to get together, and therefore, a small party can have a huge amount of power if one party gets 45% of the seats, another party gets 45% of the seats, and a third tiny party gets 10% of the seats. The two big parties are gonna fight to make the little party happy, so that they can come into a coalition and choose the Prime Minister, right?

0:57:40.2 SC: So there's absolute power that is had by these smaller parties. Whereas in the presidential system where the people just vote for their chief executive, even if it's through an electoral college, someone's gonna win, someone's gonna lose. Right? And basically, if you have a tiny party, the chances that you will ever win an election for the president become vanishingly small. And so, it's much harder to wield power as a small party in a presidential system than in a parliamentary system. You don't need physics to explain that, maybe you can think about it in terms of physics, but it's a pretty straightforward result, I think. Now, is that good or bad? Is a harder question that I'm not gonna answer right here. I think that people can argue over the pros and cons. I think that there are both pros and cons. I'll just leave it at that. You can imagine what some of the pros are and cons are yourself.

0:58:33.6 SC: James Swift says, in your excellent episode with Rafael Milliere on how artificial intelligence thinks, he discusses examining structures within the large language models and the question of whether they understand what inputs they've been given. This immediately made me think about Mary, the color scientist, the philosophy thought experiment, and whether this could be considered in some way a real version of that experiment and perhaps what it would mean then to give a running model in visual inputs, which would be in some way equivalent to experiencing red. Am I connecting two things that I shouldn't or is there something interesting there?

0:59:10.3 SC: No, I don't think you're connecting two things that you shouldn't, but I'm not sure there's something interesting there or not. To me, the Mary the color scientist... So if you don't know the Mary color science experiment. The idea is, although I certainly talked about it in the podcast with Philip Goff for example, maybe also with David Chalmers, I don't remember. The idea is that Mary, the color scientist, grows up in an environment where are no colors, all that she ever sees in her life are shades of black and white and gray, but she's a scientist and she learns by reading books or watching YouTube videos that have been grey scaled, she learns everything that there is to know about color, about the color red, okay?

0:59:53.5 SC: So the point of the thought experiment which was first put forward by Frank Jackson philosopher is, Mary knows every physical fact about color, she's a color scientist, she's read all the books, etcetera, etcetera. Then we let her walk out into the world and see something that is red for the first time. She now experiences something, she gets a kind of knowledge that she didn't have before, the knowledge of what it is like to see the color red, therefore... So the premises of the argument are, Mary knows all the physical facts before she leaves the room, before she leaves her environment, she learns something new by leaving her room and experiencing the color red, therefore, the physical facts are not all the facts. There are facts that are not reducible to the physical facts of science, etcetera, and therefore, there's something about consciousness that is not just an emergent physical phenomenon. Okay? That's the Mary, the color scientist argument.

1:00:54.6 SC: By the way, Frank Jackson, I love pointing this out, he himself has repudiated the argument. He says, It doesn't really work. I don't think it works. I think it's kind of obvious why it doesn't work. The question is from the physical list point of view, like myself, someone who does not think that color, the experience of seeing colors is anything other than some way of talking about what happens in our brains at the physical level. What happens when Mary is in the room reading physics books? Well, certain neurons in her brain act in certain ways, fire some certain synapses, strengthen some certain connections. When she walks out and sees the color red for the first time, different neurons fire and make connections, etcetera. That's it. That's all that's happening. There's nothing in there that should lead you to be skeptical about physicalism or anything like that, you're just exciting different neurons. That's all.

1:01:46.8 SC: Okay, anyway, that's my take on Mary, the color scientist. But in terms of James' question, what is the analog there of what is happening in a large language model? So I'm not exactly sure what you're aiming at in the question James, but I don't think that there is any obvious immediate analog of the large language model experiencing something that it couldn't experience without a sensory input or something like that, that I'm not so sure. It gets into questions of what it is a sensory input? A large language model is going to get signals down a wire one way or the other, whether it's us typing them in at a keyboard or whether it's hooked up to a video camera.

1:02:34.2 SC: So what I do think is interesting, the interesting analogy here is, can you look into which parts of the neural network that make up the large language model, the parameters of that network, how are they changing when we are giving it different information? And that is at least roughly analogous, but not the same as how the neurons in our brain fire and strength and connections and so forth. So, what we might learn is the difference between "learning facts about color" and "experiencing the color." Those are different things. You might learn something about how they're different. The nice thing about a large language model is that there are no institutional review boards. Is that what IRB stands for? Institutional Review Boards. We have these boards in academic scientific circles that tell us how we're allowed to experiment on human subjects and other kinds of subjects, and so, they have rules, and you're not allowed to just open up someone's brain and look at their neurons, but you are allowed to open up your neural network and look at its neurons.

1:03:43.2 SC: So maybe can learn something by direct experimentation and prodding around, that we can't with human subjects. On the other hand, with human subjects, those neurons have a very definite spatial location and topology, and the different parts of the brain have different functions in very obvious ways that you can see with FRMIs and things like that. So, it's less clear in the case of large language models, how that modularisation happens, if it does at all. So this is an exciting frontier moving forward, but I can't really predict how useful it's going to be in the near or middle term.

1:04:20.3 SC: Red Antonov says, considering how firmly emergence has been entrenched in the zeitgeist, do you have a favorite example of emergent behavior or phenomenon? Well, I am not sure how firmly emergence has been entrenched in the zeitgeist. I guess, it depends on which zeitgeist corners you are hanging out on. Certainly, I talk about it all the time, but whenever I talk about it, I have to put the caveat here that different people use the word emergence for different purposes or to mean different things. So, I already gave you an example of just the table. The tables and chairs to me are perfectly good examples of emerging phenomena in the sense that they're describing the... Some aspects of the collective behavior of many, many little tiny parts or that's a sloppy way of putting it, but perfectly accurate in this case.

1:05:07.8 SC: The less sloppy way of putting it is, there is a comprehensive description of what is going on in the table that involves quantum field theory and the core theory, and electrons and photons and all stuff like that. And there are less comprehensive ways of talking about it, which involve tables and chairs and wood and things like that. And it doesn't really matter that the more comprehensive theory is made of little things. That's the distinction I'm trying to draw, but it's more comprehensive. It is valid in a wider range of physical situations and the thing about the higher level, less comprehensive theory is that you need less data to give a specification of the system, that's what emergence really is. I can accurately capture some aspects of the physical behavior of the system with way, way, way smaller amounts of data. There's a sort of many to one map from the microscopic comprehensive description to the macroscopic emergent description, and yet we can still say something real and true about its behavior, okay?

1:06:15.8 SC: And in that case, there's lots of examples of emergent phenomena, the table as an example, the motion of the earth around the sun using just a center of mass, velocity and position is a great example. Air in the room being described in terms of density and pressure and temperature and velocity and things like that. These are all classic example of emergence. Now, I will footnote that by saying that there are some people who use the word emergent and will deny that any of the examples I just gave you count, [chuckle] because they want to use the word emergence to only refer to those examples where we don't exactly know why the higher level theory works, okay?

1:06:57.7 SC: If there's not anything spooky and mysterious about it, then it doesn't count as emergence to them, and I think that's exactly wrong, like, you can define words whatever way you want, go ahead. But I think it's not helpful to restrict attention to the cases we don't understand. I think that we're going to get better understanding by really, really understanding the cases we understand first and then expanding that slowly to include things we don't understand. If you leap right to the hard things, it's gonna be hard to get a more comprehensive understanding there. Okay, so I don't know whether that answered your question, Red. But that is my way of thinking about it.

1:07:35.4 SC: Brent Meeker says, Roy Kerr says that Penrose's Singularity theorem for black holes doesn't actually apply in the real world. He says there's a spinning object at the center of a black hole. He points out that the solutions that describe black holes are vacuum solutions, that a vacuum was just assumed in finding the solution, they are not solutions for collapses of real bodies. Are there any solutions, even computational ones for collapse of matter into a spinning black hole and do they treat the matter quantum mechanically? So for those of you don't know, Roy Kerr is the New Zealand mathematical physicist who first derived the solution to Einstein's equations for spinning black hole, what we call the Kerr metric. And since most black holes are spinning, that is actually a really, really important little piece of knowledge to have.

1:08:25.0 SC: And the singularity theorems that he's referring to, are theorems first... Patrick Penrose proved the first of them, and then Steven Hawking and Robert Croch proved others. These are a set of theorems that show that under certain assumptions, when you have collapsing energy and matter and so forth, it will form a black hole... Well, you will hit a singularity, is the actual content of the theorem. Then there is the cosmic censorship conjecture, which says that every singularity has a horizon around it, and is therefore a black hole. That's harder to prove, it's not even completely true in all cases. But the singularity theorems are saying, you will get a singularity in classical General Relativity.

1:09:05.3 SC: So, I'm surprised that Roy Kerr would say that Penrose's singularity theorem doesn't actually apply in the real world. It is absolutely true that the exact solutions that we have for black holes, including the Kerr solution, usually assume vacuum. Not always, there's the [1:09:25.8] ____ solution that allows there to be an electromagnetic field, for example, and there are some solutions for collapsing shells of dust or matter or things like that. But those are usually perfectly spherically symmetric, that makes our lives much, much easier. There's only evolution in the radial direction over time, you don't also need... You're not breaking the symmetry by allowing the thing to rotate. So, yeah, as far as I know, there are no exact solutions to Einstein's equations representing collapse of matter into a spinning black hole, but the singularity theorem is something different than that.

1:10:00.8 SC: The singularity theorem is... The whole point of it was that it is not an exact solution, it's an exact statement that a wide variety of solutions will have the following feature, namely they get a singularity, okay? So I think that's still gonna be true in the real world. The assumptions behind it have to do with the behavior of matter, and that might not be physically accurate. So, any theorem always has the feature that if you violate the assumptions that you used to prove the theorem, then the theorem no longer holds. And as you allude to the end of the question, does that treat the matter quantum mechanically? I don't know of any solutions to the equations of general relativity that fully and completely treat the matter quantum mechanically. No. So there's the idea that... There's two ideas. One is that whenever we talk about singularities in general relativity we're cheating a little bit, because classical general relativity ignores quantum mechanics.

1:11:02.5 SC: I think that's entirely true. I think that you should read the existence of singularities in classical general relativity, not as evidence that there are singularities in the real physical world, but as evidence of classical general relativity is breaking down at a certain scale, and it breaks down the big bang and near the singularities of black holes, etcetera. Which is perfectly sensible, because it doesn't include quantum mechanics why should you trust that anyway?

1:11:27.0 SC: So to that extent, if you're saying, Well, once we understand quantum mechanics, people no longer think that there are really singularities and black holes, then I'm completely onboard with that. The other aspect is the forgetting about the singularities, okay? Just thinking about the general internal structure of the black holes, and this is hard to get into without pictures and general relativity details, etcetera. But in a spherically symmetric black hole, whether it is totally vacuum as in the short shield solution or whether you include matter in there, etcetera. These internal structure is pretty straightforward, everything collapses, and it will eventually over time hit a singularity, that's basically it.

1:12:11.3 SC: If you look at the vacuum solutions for spinning black holes or for electrically charged black holes, they're wildly complicated inside. You get a whole other regions of space and universes and close time light curves and a whole bunch of crazy nonsense inside. Almost everyone who is an expert in general relativity believes that all this crazy nonsense is an artifact of finding a very, very specific exact solution that has no matter in it that is purely vacuum. So that is nothing to do with Penrose's singularity theorems. But if all you're saying is that real world black holes that are created from collapsing matter will have a very different internal structure, then the toy model exact solutions that you can find, for example, in my general relativity textbook, then I completely agree with that.

1:13:05.6 SC: And people have done numerical simulations to try to understand what happens when real matter collapses into black holes. I am not an expert on the current state of the yard and what we've learned from those numerical simulations, but it won't be the craziness that you get in the fully, what we call the analytically extended Kerr solution or Reissner-Nordstrom solution. It will be something probably a lot more believable. None of this, by the way, matters for anything that happens outside the black holes. Outside the black holes, we know what is going on. So it's interesting to think about what happens inside, but it's not gonna have any effect on what LIGO observes in gravitational waves or anything like that.

1:13:45.5 SC: Siddhartha says, Your recent paper, Reality Realism, says the abstract proposition that two plus two equals four is not a reflection of an independently existing truth. This was very surprising to me. Later, the paper says we are then left with a form of if then-ism in which two statements of the form, if these axioms are accepted, this theorem can be proven. But a typical mathematical realist wants the proposition theorem to be real, not merely a conditional statement. I always thought, Siddhartha is continuing... I always of mathematical realism as given a particular set of axioms, theorems and properties logically and necessarily construct themselves by implication independent of any physical reality, such that anyone starting from the axioms has to reach the same conclusions.

1:14:32.0 SC: Perhaps a better term is logical realism and mathematics in general, is an exploration of this logical realm that consists of all possible axioms and their implications somewhat like the library of battle, that [1:14:45.0] ____ story that we talked about before. But it seems you have something different in mind when you say, not merely the conditional statement. Do you think logical realism is different and side steps the issues with mathematical realism, you mentioned in your paper?

1:15:00.9 SC: Well, so this is referring to a paper that I put online recently, you can find it by googling Reality Realism or my name or anything like that. It was in response to the book Morality and Mathematics by Justin Clark-Doane. So, some of you might remember, we had Justin on the podcast to talk about his book Morality and Mathematics. Justin is a philosopher who has... He talks about the relationship between morality and mathematics in the sense that what are our arguments for realism or objectivity in both domains? And it's a complicated story that he tells, because it's not that they are exactly the same or exactly different. He says, they're more different... They're more similar than you might think, but they're still a little bit different. And one of the distinctions he tries to draw is between realism and objectivity. Objectivity is when everyone would agree something is true, but realism is something more than that. Realism is a kind of existence, mathematical realism or moral realism.

1:16:00.5 SC: So, I will very quickly confess as I do in the paper that I'm not an expert on exactly these issues and one should be an expert if one is going to comment on them. I only commented on them because I was invited to do so by Justin. In fact, in his book Morality and Mathematics, he starts the book with a quote by me, because in the big picture, I talk about not being a moral realist, but I am a realist about the physical world, and he says that therefore, Carroll must be a realist about mathematics because we use mathematics to describe the physical world.

1:16:41.0 SC: And I'm not a realist about mathematics actually, I am a realist about the physical world, but not about mathematics. I do think that mathematics helps us describe, let's put it that way, the physical world, in a very clear way. I use mathematics all the time, but I'm not a realist about it. So in the paper, even though I'm admittedly not an expert on the philosophy of mathematics or the foundations of mathematics, I try to describe what I think how it could possibly be that somebody could be a realist about the real world, about the physical world and not a realist about mathematics, not Platonist, not someone who believes the mathematical structures and so forth have real independent existences.

1:17:24.0 SC: So, to answer Siddhartha's question here, you should really talk to a real expert here, but my belief is that the standard... To the extent there is any standard understanding in the philosophy of mathematics, which is probably not completely true, there's people who disagree about all sorts of things within that field, but the idea is that there is a difference between objectivity and realism. So in that field, the word realism means something more than, everyone would agree that these theorems follow from these axioms. You can call that something, but apparently, you should not call it realism, logical realism or mathematical realism or anything like that.

1:18:10.9 SC: It's something other than realism. So the phrase if then-ism is used by these people to denote the view that all theorems are, are logical, reliable, objective, conclusions, deductions from their axioms. That is something less than what a mathematical realist in the usual construal means. So, a mathematical realist doesn't just think that there are axioms of number theory from which we can derive, the fact that there is no greatest prime number. A mathematical realist believes that there is some truth to the statement that there is no greatest prime number, that goes somehow beyond any particular set of axioms. And this becomes very important... The distinction becomes important because when you dig deeply into the mathematical foundations literature, you realize that different sets of axioms will generally have different kinds of models, different sets of axioms do not uniquely imply what it is you are talking about.

1:19:20.5 SC: And in particular, the axioms that we used to define the integers, the number theory axioms, Peano, arithmetic, etcetera, have different models, so there are what are called non-standard models of arithmetic. There's not a unique model of what we call arithmetic. And they have different properties, these models, and so, the real mathematical philosophers dig into what that means and what it's good for, and again, I'm not a super duper expert there. But I agree that there is a sense in which everyone agrees on the if then-ism statements, everyone should agree that under the usual axioms of mathematics, the theorems that you prove using those axioms will follow from those axioms. But I don't attach realism to that.

1:20:08.9 SC: I think it's a different thing. So my point in reality, realism is what I think is real as a physical world, that's what I think is reality. So that's the name reality realism, and what we call mathematics is a set of ideas and structures that turns out to be very, very helpful in describing the actual world. And the experts in mathematical philosophy disagree about whether or not you need to be a mathematical realist to consistently describe the world of using mathematics and things like that. So, even though I'm not an expert, I'm at least consoled by the idea that the real experts disagree with each other, then you can check that out.

1:20:51.0 SC: So I do think this stuff is important. I don't wanna downplay it, I just wanna be very, very honest about not being an expert in it myself. And being an expert both implies sort of knowing some substantive things, but also knowing the language, the lingo and how it's used. So, the lingo that is used by the experts does not always map on in any simple way to what we think the words mean from our experience. Henry Jacobs says, is poetic nominal-ism a thing? This would be a nominalist who moves through the world, kind of like a poetic naturalist. For example, she would believe the chairs are not real, but are still a valuable concept, and it's a good idea to act as if they are real in many domains. As an applied mathematician, I find myself leaning towards something like this since I use models all the time and they're useful enough, even though they're not real, I literally make them up.

1:21:46.1 SC: I didn't group these two questions together, but I put them together, because I do think that they're raising similar related issues and important issues. And the more I think about it, I just maybe think that calling things real or arguing about what is real and what is not real is not helpful. [chuckle] That there are properties that things have, that we can rely on certain deductions from the axioms, like you said, and those are important properties, but debating over what's real and what's not, I don't know. And this goes to not only math and logic, but as Henry brings up the idea of different levels of description of reality. So nominalism is the point of view. It was advocated by Jody Azzouni, who we had on the podcast, that what we talk about... For mathematics, which is the usual club... I don't know if this is true, I'm talking about things, I'm not an expert on.

1:22:40.8 SC: Nominalism goes back to Ockham, William of Ockham of Occam's razor thing. And it emphasizes the idea that we name things. We give things names, we give things labels, thus the name nominalisms. And rather than thinking of things as real, we should think of them as useful names, okay? So in particular, mathematical nominalists think that way about mathematical structures, integers, things like that. So what Henry is suggesting is that rather than being a poetic naturalist, which is the point of view that I advocated in the big picture, where you say that there is a real world, it really exists, but there's many ways of talking about the world, and if you have a way of talking about the world, that accurately captures some real behavior of the world, then the ingredients in that way of talking perfectly well deserved to be called real, like tables and chairs, for example. And Henry is saying, how about instead of calling them real, we think of them like a nominalist would, and we think of them as useful names.

1:23:40.9 SC: You know, sure, I kind of... I don't wanna... I don't know, is the short answer. Again, I'm not trying to dismiss the question and think of it as uninteresting or anything like that, but there are very fine distinctions being drawn here between things being real and things being useful in terms of describing the world and living in it and getting through it, and I just don't know whether those distinctions matter in any important way. I wish I knew. So again, I'm not saying that we shouldn't ask this question, it's not interesting. I think that I don't know the best way to do it. I'm not sure if it matters, whether we use words like real or just use words like nominal name, useful, things like that. I need to think about this more. Now that I'm officially a philosopher, 'cause I'm a faculty member in a philosophy department, I can't get away with not having opinions about these very important issues.

1:24:35.3 SC: Eric Chen says, in your conversation with David Wallace, you two seem to disagree as to the best way of formulating the past hypothesis. You presented it roughly as the early universe was in a low entropy macro state and the micro state of the early universe was some generic micro state within this macro state. Whereas David framed it more as the micro state of the early universe was not a conspiratorial one that contained lots of delicate and crazy correlations. Would you expand on this disagreement? What does a micro state with many crazy correlations actually mean and look like, and what's your current preferred way of formulating the past hypothesis?

1:25:09.3 SC: I feel like I'm... This in a whole bunch of questions in a row to which the answer is, I'm not sure or I'm not an expert, or I don't know. But I think that... I'm not sure how I formulated the past hypothesis in the conversation with David, but I don't think I would have done it the way that you quoted. It's very possible I did, and I misspoke or changed my mind, but the usual way to think about it following David Albert's book, Time and Chance, from the year 2000, which really put things in these terms very clearly for the first time, is to separate out two different hypotheses.

1:25:44.0 SC: One the past hypothesis which roughly says the early universe has low entropy, and the second one, which is just called the postulate about statistics, which roughly says that within that low entropy early state, the universe is in a typical micro state, not someone with crazy correlations in some way. And the point of that second thing, the point of the postulate about statistics is, as Boltzmann knew very well you can always go down in entropy from any initial starting point, unless you literally at the minimum of entropy, you know that you can have specific micro states whose future development decreases entropy, because you can just imagine going toward that micro state from the past and then time reversing it, okay?

1:26:36.1 SC: So basically, they just wanna rule out the possibility that not only is the early universe low entropy, but also, that it is sneakily arranged in one of these very, very, very rare micro states that will still go down in entropy, even though it's already pretty low to start with. That's not surprising. That's the thing to a physicist about this discussion. The fact that the early universe had low entropy is like a super important fact. That's really weird because there's very much fewer low entropy states than high entropy ones that is calling out for some sort of explanation. But the fact that the early universe does not have crazy correlations, that's a fact, but that's not a very surprising fact, that is what you would expect. So physicists worry less about this particular statement. Now, the way that David likes to put the past hypothesis...

1:27:30.0 SC: I don't know if it's the past hypothesis or the postulate about statistics that should really be the one being talked about here, but he, rather than making some statements specifically about that early micro states, he wants to sort of use the future evolution of the micro state to put conditions on what can happen or what does happen, or how we should think about what those initial conditions really are. I don't think it's that different. I think it's just a matter of convenience, honestly, and maybe it's cheating a little bit to even think about future correlations or future behavior, when you're coming up with a characterization of the early state. It'd be nicer to have a characterization of the early state that only refer to the early state itself, not to its future time evolution.

1:28:17.6 SC: Also, one of the reasons why I'm not that into this question is because I think that all of this discussion is kind of secretly assuming classical mechanics, imagining the universe is kind of like a box of gas that obeys the rules of classical mechanics, and we can have sneaky correlations in it. But the world is not classical, and David Albert and David Wallace both know this very well, they write a lot about quantum mechanics, and I think that you can start from a very bland specific initial wave function and get all of the future interesting evolution out of it from... In my view, or David Wallace's branching of the wave function, but in someone else's view may be collapse of the wave function under observations.

1:29:00.6 SC: So I suspect this particular argument is completely irrelevant, once you take quantum mechanics into consideration. There's probably some specific... Rather than saying, the early universe is one of various micro states that has the following properties in quantum mechanics, we can just say, it's this micro state. It's this particular quantum state of the early universe, and everything follows from that. That's exactly what Stephen Hawking and Jim Hartle tried to do in quantum cosmology, for example. I think ultimately, that's the way to go, so the classical conversation isn't that interesting to me.

1:29:35.9 SC: Shlayer says, I learned a lot from your discussion with Molly Crockett on the psychology of morality. You and she discussed how a focus on individual behavior can cause us to miss systemic problems. For me this brings to mind the question of emergence. Do you think that human civilization is a sufficiently complex system that it can be seen to have its own goals and motivations that emerge from, but do not exist at the individual level? I think that is a very good question, and we can raise it in a related context with AI rather than human civilization. I would say though, for the human civilization question, the short answer is no. The question that I would ask about reported emergent phenomenon is, is positing this particular feature of the higher level emergent description, giving us some useful information that improves our understanding of how the system behaves?

1:30:30.6 SC: So, when you have... What is a good example? When I am hungry, I will go to the grocery store and I will buy food and then I will come back and cook it. And I don't know exactly what I'm gonna buy, but just knowing that I'm hungry and I want to eat something and I don't have any food in the house, you can predict that I will go to the grocery store and buy food and bring it back or go to a restaurant or something like that. In other words, there is a goal that I have right now that... And knowing that goal helps me predict the future evolution, okay? I don't think that that's obviously true about human civilization, certainly empirically, individual parts of human civilization have done weird things that it is hard to account for by any uniform unified goal. At any one moment of time, different is perfectly okay to speak about countries, nations, governments or whatever as pursuing a policy with a certain goal. That's okay, but it's just like a shorthand for saying the actual government is not making the following actions, okay?

1:31:43.6 SC: That attribution of intentionality can be useful. I'm not denying that it is not possibly useful, but I don't necessarily think it's the best way to talk about human civilization. Daniel Dennett, another former podcast guest, gave what is the classic discussion of this question of when should we talk about systems as having intentions? He called it the intentional stance, when should we talk about the system as having a goal, having its own desires or whatever? And yeah, guess what, it turns out to be a tricky question. I don't... So maybe at some level for human civilizations, that's a useful way of thinking, but I think generally, it's just anthropomorphizing something that is a little bit different than a human being.

1:32:30.0 SC: Coffee Addict says, ChatGPT, the AI program you can go in and talk to has been a hot topic as an application of machine learning recently. Additionally, several image generation tools have been introduced, including stable diffusion, dolly to [1:32:46.5] ____ major needs, etcetera. I heard that the mechanism of these image generation tools are based on diffusion models, which were originally studied in non-equilibrium thermodynamics. Do you find anything interesting about the relationship between machine learning and physics, such as the use of diffusion models in image generation?

1:33:02.1 SC: Before answering, let me just say to everyone listening out there, if you haven't played around... I should have said this before in the discussion with Raphael Milliere or various other ones, but if you haven't played around with ChatGPT or Dolly or [1:33:20.6] ____ mid-journey or one of these image creation algorithms, it's very, very worth playing around with them. Both because... And it's not that hard, sign up and you can get access and then you can ask it questions, you can get into arguments with ChatGPT, very clearly. It's very, very obviously programmed to say that it is not conscious. You can try to convince it of some things, but it'll say certain... If you ask it about me, as I said, I like a few AMS ago, it'll say very wrong things about me personally, but then you can correct it and it usually go, Oh, okay, thank you for the correction. Not about whether it's conscious. So it is easy to do. It is both very, very education... It's fun, because you kind of are trying to think, trying to guess what I will say before it says it, and it's very smart, knows a lot about the world, etcetera, but also because they're gonna be everywhere, these AI programs.

1:34:18.3 SC: And so, it's a case where even if you don't wanna be an early adopter, it would be useful to you, I recommend to at least get a feeling for how they work and what they do. Anyway, with all that in mind, Coffee Addict is asking a question about the relationship between how these models actually work and physics. And there is absolutely a relationship, and I don't know a lot about it, so I'm not here to actually answer the question in any detailed way, but there is a lot of physics there, because what happens is... Back in the old days, as we talked about with in previous podcasts, Gary Marcus, Melody Mitchell, Stewart Russell, we've had a lot of podcasts about artificial intelligence. The old fashioned way was to try to model the world explicitly in your AI program that was called symbolic approach to AI. In the more recent connectionist approach, you don't presume to be better at modeling, than the computer, you make the program be basically a blank slate to start, and you let it learn, you let it try some random things and you tell it, Yes, you did well, no, you did badly, and it learns very, very rapidly.

1:35:28.4 SC: So that's very, very useful. It's good, you can train the computer to beat the world's best Go players or chess players, to have conversations, to make beautiful images and so forth and all that stuff, but is it the best? Is it in fact the most efficient way to learn? And in fact, it's not just you build a giant computer and you let it rip... Okay, there is a lot of work that goes into figuring out the best way to help your AI program learn. You know the amount of data out there, it has to sift through is very large. You would like it to be a little bit efficient, a little bit thoughtful, a little bit smart about how it gathers data, puts it together and does its next thing. So, it is undoubtedly true that thermodynamic or statistical mechanical considerations come into this question.

1:36:20.2 SC: For one thing, there are random numbers there, you can ask a GPT or ChatGPT the same question twice. So it'll give you two slightly different answers or maybe even very different answers. And basically, it's this general idea of exploring a really, really large space, the space of possible answers that ChatGPT can give you in terms of sentences, strings of words, is enormously large, much, much larger than it will actually ever do in some very weak sense. The problem that the program is trying to address is how to pick out the right string of words in the space of all possible strings of words, which is just a crazy large thing.

1:36:56.6 SC: So, that's why diffusion appears in the name stable diffusion, there's sort of different ways of attributing structure to this giant space of possible answers and using that structure, leveraging it for giving answers in a reasonably efficient amount of time. But that's it, that's all I know. I mean, that's a very superficial level of understanding. I know that it's really important. Johns Hopkins is launching a major initiative to hire people in artificial intelligence across the disciplines. So we have people who will be working in the physics of AI, the philosophy of AI, History, Biology, of course, the computer science and the engineering also. So there is a whole new field of the physics of machine learning and artificial intelligence that is fascinating and important. I don't know a lot about it, but I bet I will be learning about it in the years to come.

1:37:58.7 SC: Jonathan Bird... Ah, finally we're back on my home turf here. Jonathan Bird says, Why is losing information to entropy expected and accepted while information lost in a black hole is treated as a crisis in physics? So it's not a crisis in physics, by the way, it's a puzzle, it's a puzzle to work out. That's the bread and butter of physics, that's all we want. We want puzzles. How does information, or does information, get out of an evaporating black hole? And the answer to this question is very, very simple, actually. Although, it's an important one. It's just the difference between what you might call fine-grained information and coarse-grained information. Fine-grained information... Imagine the world were perfectly classical, so you don't have to worry about the subtleties of quantum mechanics. The world was literally a bunch of billiard balls bumping into each other or interacting through some force law, okay?

1:38:47.4 SC: When you want to describe the air in the room, as we said, you could do it in principle by knowing everything about every atom and molecule. Every position, every velocity, etcetera, you could predict what happens next, or you could just know the coarse-grained information about temperature and density and pressure and predict pretty well what's going to happen. So, the fine-grained information is that complete specification of every position and every velocity in the air in the room. The coarse-grained information is just the observable temperature pressure and stuff like that. So when you say information is lost because entropy is increasing, that's just because you're coarse-graining. That's not a statement about the fundamental laws of physics. That's just, if you list a sequence of coarse-grained states that a system is in, and those states have increasing entropy, and that's all the information you have, then knowing the higher entropy states, knowing the states of the higher entropy states gives you less information.

1:39:47.6 SC: You have lost information compared to knowing the low entropy configuration that it used to be in. Whereas, the underlying dynamics of the billiard balls interacting with each other is perfectly information-preserving. Even though, I personally I'm coarse-graining and therefore losing information, the world is not losing information. It's being micro-physically complete and reversible. That's what's supposed to be happening in the black hole. When the black hole evaporates, it's supposed to be obeying the rules of quantum mechanics and at the fine-grained level, it shouldn't be losing any information. That's what we think. Anyway, some people disagree. There is a subtlety here that I will bring up, which is that, there's a very important part of quantum mechanics where information is obviously lost, namely measurement.

1:40:38.1 SC: Whether you want to think of it as decoherence and branching as a many-worlds person would think of it as, or you just want to think of collapse of the wave function as a Copenhagen person would, you have an unpredictability, an irreversibility and if all you're told is the state after a measurement or a branching, you can't reproduce what the state was before. That is an irreversible information losing process. So, that's gonna... We talk about black holes and Hawking radiation and stuff like that, black holes get measured. They interact with the rest of the world. They decohere. So, this statement that the total evolution should be information conserving is a subtle one that you better be careful about. In fact, I wrote a paper about this. I was a co-author on a paper, let's put it that way.

1:41:31.6 SC: A lot of the work of course was done by my collaborators. This was with... I'm gonna forget the whole list of people, but Ning Bao was the first author. And the title I think was something like Branches of the Wave Function with Black Holes Need Not Contain Firewalls. This was supposed to be responding to the idea of firewalls. The point of the paper was we took seriously the idea of branching and decoherence, etcetera, and pointed out that where entanglement lives depends on whether you're asking a question about just one branch of the wave function, which you would ask if you were making measurements, or whether you're asking questions about the global wave function of the universe, which no one observer can see. So, we pointed out that there was a loophole in the argument for firewalls that was put forward by amps that we talked about here on the podcast before.

1:42:19.7 SC: So anyway, I'm just mentioning that as a point of personal interest. That I think that people are a little too quick and glib when they're talking about unitary information-conserving evolution of black holes. There is something going on that loses the information that we should keep in mind. It doesn't solve all the problems keeping it in mind, but it should nevertheless be kept in mind. George Hampton says, I just started looking into retrocausality and the transactional interpretation of quantum mechanics. And while it sounds interesting, I'm concerned that I might be chasing a dead end idea. What are your thoughts on this concept? And do you think it is worth my time and energy to dig more deeply into this? Well, I will be perfectly honest. I think it is not worth your time and energy to dig more deeply into retrocausality and the transactional interpretation.

1:43:06.5 SC: There are people who are very, very smart and very, very well informed who would completely disagree with me. So you don't need to follow me just because I'm an expert here. It's a controversial question. I am a many-worlds person. I think that the Schrodinger equation tells us how the quantum state evolves. I think that it is kind of silly to imagine we need to bring in influences from the future interacting and affecting what we are observing today. It's certainly not necessary given anything that we know experimentally or theoretically about quantum mechanics. And I certainly don't think it makes anything easier or simpler to imagine that so I'm not gonna pay too much attention to it. As always, I could be wrong, so let the buyer be aware. Gauta Ainoval says, Einstein used a lot of thought experiments when thinking about his new physics. These days, thought experiments are used a lot in philosophy, for example, when thinking about consciousness.

1:44:02.1 SC: However, I wonder whether thought experiments have become less helpful in physics. At least when studying complex systems, they seem to have a limited role, but maybe it is different in, say, cosmology or particle physics. Do you have any thoughts on this? Yeah, I think that thought experiments are almost impossible to get away without using. What do you mean by a thought experiment? You mean some situation that you set up in your mind. For whatever reason, it's too complicated or too difficult to do it in reality, like, what if I were riding a light beam? Okay, so I can't do that in reality, but I can ask the what-if question. We're trying to probe the logical coherence of our ideas. If you believe this, then this should follow in some hypothetical situation. That's absolutely something that we use all the time in modern physics. We were just talking about black holes and firewalls.

1:44:53.0 SC: The firewall dilemma comes from thought experiments. If I were to dive into a black hole, what would I see as I cross the horizon? I can't actually do this. It's just a thought experiment. Now, do they play as an important role in complex systems? I think the answer is yes. They will play a very important role, but maybe the usefulness is less obvious, because what thought experiments let us do is probe the, like I said, the logical consistency and implications of our theories. In the case of complex systems, we don't have great theories. We have some principles and hopes, some predictions, some features that seem to be universal or at least very common, but we don't have a paradigm, like we have classical mechanics, quantum mechanics, statistical mechanics, relativity or whatever, right?

1:45:46.4 SC: There's less call for pushing the implications of our theoretical predictions. Now we don't have as many theoretical predictions. So it's not the fault of thought experiments. It's the fault of our current state of knowledge in this very difficult field of complex systems. Anthony Rubo says, the Academy Awards, what are your thoughts on the dodging or the concept as a whole or on the snubbery of The Menu? I think it's outrageous that The Menu was snubbed at the Academy Awards. Although, to be fair, no, I don't think it was outrageous. I don't care that much about the Academy Awards. I do think that The Menu was one of my favorite movies of the year. I did not see anything like a comprehensive list of movies this year. I certainly didn't even see most of the best picture nominees. So I cannot possibly legitimately complain about who got nominated and who didn't. But for me, the best movies of 2022 were The Menu, Glass Onion and of course, Everything Everywhere All at Once, which did win the Academy Award.

1:46:55.7 SC: My personal rooting interest there was that I've always been both happy and somewhat amazed at how we've managed to have all sorts of impressive people here as guests on the Mindscape podcast. And one qualitative measure of their impressiveness is the number of prizes won. So we've had Nobel Prize winners, we've had MacArthur Prize winners, Pulitzer Prize winners, Breakthrough Prize winners, etcetera. Grammy winners, Emmy winners. We had never had, as far as I know, an Academy Award winner on the podcast. But, of course, I did interview the Daniels about their movie Everything Everywhere All at Once, and then they went on to win the Academy Award. So now, I'm gonna try to spread the word that the way to win awards, to first appear on the Mindscape podcast, and that will boost your case. I love The Menu. You should also watch that one. Sad that it didn't win as many awards, but not completely surprised, because that movie was more or less made for me.

1:47:58.8 SC: Jennifer and I both agree. We were watching that movie, and I don't think there's ever been a movie where, as often, we were elbowing each other and going, did you see what they just did? Isn't that great? And I also think, for those of you who haven't seen it, The Menu is described as a satire of foodie culture. And I think that's just missing the point. Again, on maybe like the Jane Austen discussion, The Menu is way deeper than a satire of foodie culture. Foodie culture is just too soft and easy a target, okay? The Menu is really an interrogation of artistry and creativity and the relationship of artists and creators to their audiences. And it's a much broader, richer more interesting vein to mine than just going to fancy restaurants and making fun of the dishes for being a little weird. They're both there. They're both in the movie, but there's more to it than that surface level.

1:48:58.1 SC: Eric Weber says, I'm auditing a course on philosophy of spacetime and matter, which is excellent. My question is on simultaneity. You say in The Biggest Ideas of the Universe that Proxima Centauri is four light years from Earth, which corresponds to an eight-year span of events that could be considered simultaneous with a single event on Earth. That blows my mind and I want to do the math. I have my graph paper out, spacetime coordinates and light cones ready to go. But even after solving many relativity problems in class, I can't seem to figure out where to start on this. Can you nudge me in the right direction? I think there's less math here than meets the eye, to be perfectly honest. Get out your graph paper, draw some X and T axes, your spacetime diagram. Put the Earth right now, this very moment at the center, X equals zero, T equals zero.

1:49:52.3 SC: Draw some light cones at 45 degrees. At X equals four light years, put a straight line going vertically, representing the world line of Proxima Centauri. And the point is that in relativity, you can pick reference frames that stretch in any horizontal seeming line, as long as it's not tilted more than 45 degrees on your piece of paper. So, I can draw a line tilted at 44 degrees. Going through the Earth, going through the center of your axes, that would be the reference frame constructed by someone going through the Earth right now, moving close to the speed of light. And it would tilt so that it goes up and to the right and hits that Proxima Centauri line near T equals four years, because X equals four light years corresponds to T equals four years. I could also tilt it in the other direction, going down, sloping downward, so that it hits the Proxima Centauri line near T equals minus four years.

1:50:56.2 SC: So the only math is, what is four minus minus four? The answer is eight. That's where the eight comes from. The total amount of time spent by Proxima Centauri, which is four light years away in the light cone, sorry, I should say outside the light cone of the Earth at one moment in time it lasts eight years, because it's four light years away. And it's just very, very simple math. You don't have to work very hard. Sorry about that. I don't want to disappoint you that there's not more math there, but that's all you need for this. Qbit says, you mentioned that in the weak field limit, which includes all of our everyday life, gravity fits well into the framework of ordinary quantum mechanics. At the same time, Roger Penrose proposes the controversial idea that every wave function collapse is caused by gravitational effects. How is there even room for such an idea if the weak field limit has already been solved without any problems?

1:51:53.1 SC: Well, it's not actually a contradiction. I mean, well, sorry. It's not surprising, but it is kind of a contradiction because Penrose is suggesting that ordinary quantum mechanics is not right. He's suggesting that we modify ordinary quantum mechanics. So, my statement is that gravity in the weak field limit, like you correctly say, gravity fits in perfectly well to the framework of ordinary quantum mechanics. I don't think Penrose would disagree with that. He's just saying that I want to change ordinary quantum mechanics in a way that depends on gravity. So that's an alternative idea. And we're too far away from being able to test this to know which idea is correct. It is potentially testable. People have thought about trying to test Penrose's suggestion for when wave functions collapse, and hopefully they will do it.

1:52:39.5 SC: And I predict that what they will see is that his prediction is not correct. I would be very, very surprised if gravity had anything to do with the reasons why wave functions appear to collapse. Josh Charles says, in an emergent spacetime, is it possible that our definition of locality is updated such that Bell's theorem is no longer applicable? Could spooky action at a distance only seem distant because we don't yet fully understand how space works? Well, yes and no. I kind of want to say to this. There's two aspects that are worth mentioning. One is that in an idea where spacetime is emergent, even if it's just space that is emergent, certainly the notion of locality is going to be a little bit less fundamental than it is if you start with space and time as fundamental ingredients of your theory. Locality is going to be an emergent approximation, like temperature and density and tables and chairs. It's not gonna be part of the fundamental architecture.

1:53:43.4 SC: So, in principle, there could be some minor variations on things like, Bell's theorem. Although really, Bell's theorem has just to do with correlations. Not necessarily... There's no statement in Bell's theorem that the correlations change as a function of distance. Bell's theorem is just a statement about what correlations there can be in a quantum theory versus in a theory of classical correlations. So, I don't expect any dramatic change in Bell's theorem. But the other thing to say is that, I don't think that the quantum mechanical, any surprise that you are feeling because of Bell's theorem, I don't think is a big deal. I don't think Bell's theorem is a big deal. Let's put my own cards on the table here. I think that Bell's theorem surprises you if you were really hoping that the world is fundamentally classical and that space and time are fundamentally fundamental.

1:54:45.4 SC: In many-worlds, the fundamental thing is the quantum state of the universe. And that is a vector in Hilbert space. And the whole idea of space is something that pops out at the emergent level. And to me, from that perspective, the weird thing is not that there is some kind of locality here as predicted by Bell's theorem, some violation of locality, some kind of non-locality predicted by Bell's theorem. The weird thing that demands explanation is that there is any locality. Since I don't start with space as a fundamental ingredient of reality, the thing that requires explanation, the clue that should help guide us in developing our theories is that locality is a pretty good approximation in the world. At the classical level, it is an extremely good approximation. So, I think that we don't know how space works in any complete way, but I also think that if you are really accepting of the Everettian point of view of reality, Bell's theorem is not some weird mystery that we must solve to understand where non-locality comes from.

1:55:50.7 SC: The mystery is why the world looks even a little bit local. Nikita Lozovoy says priority question. Remember, for those of you who are newcomers here, if you are a Patreon supporter, you can ask questions and since I can't answer all of them, once in your life, you get to label a question a priority question, and I will do my best to answer it. So, Nikita's priority question is time travel-related question. Recently, I listened to your solo podcast about time travel and it's awesome, but I'm still mind-boggled about something and thought you could help to clarify. As far as I know, a clock on the International Space Station is going about 0.007 seconds behind a synchronized clock on Earth's surface. For every six months it's in orbit, and that's through the effects described by the special theory of relativity.

1:56:42.1 SC: What I don't understand is why it's specifically the clock on ISS that goes slower relative to the one on Earth and not vice versa? Well, I think that this is something you can read about in various of my books from Eternity to Here, if you want the non-equation version, or there's some more equations, just a little bit more equations, and the biggest ideas in the universe, but it's not really... I personally don't like saying that one clock moves more slowly than another. I think that I get it. I know why people say that, but that's only gonna happen to clocks that are separated from each other, because they're moving on different trajectories through the universe. So how do you know how to compare them? You would have to bounce light back and forth or something like that. You'd have to agree on a protocol for doing that, but I know what they mean in this case.

1:57:34.7 SC: When you have a clock on a spacecraft orbiting the Earth, you can sort of get a signal from it when it's right above you and measure its time, and then wait for it to go around one orbit, get another signal, so you can ask how long has passed to you on the space station, and then you can compare it to how long has passed down there on Earth, and there will be a difference. 0.007 seconds? I don't know, that sounds perfectly plausible to me. But this is a combined feature of both special relativity and general relativity. And again, I'm not even sure in this particular case which is more important. The space station is traveling, it's moving compared to the rest frame down here on Earth, but it's also in a different gravitational field, so there's two effects and you have to include them both. In GPS, for example, this is why you have to include general relativity in your calculations.

1:58:25.0 SC: But anyway, the short answer is that the trajectories are not equivalent to each other. So you're asking, why is it one clock that is objectively said to be moving more slowly than the other one, but there's no symmetry between these two situations. One is in orbit around the Earth and the other is forgetting about the rotation of the Earth, because that's very slow compared to the orbital period of the space station. So to all intents and purposes the other one is sitting stationary on the surface of the Earth. So you have two different trajectories in the universe and I'd like to always recommend when people are thinking about these time in special relativity questions, just ask the analogous question for spatial distances in ordinary physics. I have two points in space and I travel between them on two different trajectories.

1:59:16.5 SC: If one is a straight line and the other is a curved line, no one is surprised that one involves traveling a different distance even though you leave at the same point and you arrive at the same point. Which one takes the longer distance? The one that is more curvy, the one that is less straight. The two curves are not equivalent to each other. So, in this case, there are two objects, a clock down here on Earth and a clock on the space station, that to some approximation are leaving from and are returning to the same points. We can be careful about that, like, what we mean by that. But there are ways to be careful about that. And the trajectories are very different. So you should not expect that the elapsed time on these two trajectories is the same.

2:00:02.0 SC: Both of them are kind of moving, 'cause the Earth is rotating and one is deeper in the gravitational field, etcetera. So it's a little bit trickier. But at the end of the day, you will plug into the equations. You will figure out what the elapsed time along that trajectory is in much the same way that you would figure out the elapsed distance along some curve here on Earth would be. In this case, the way that it happens to work out is that the clock on the International Space Station is gonna be a little bit behind the one here on Earth, but you could work out the equations for yourself. Paul Kent says, what are your hopes for the future of social media? In the wake of Twitter chaos, many competitors are now jockeying for position.

2:00:44.0 SC: I'm intrigued by the alternative decentralized network approach to social networks such as Bluesky, Jack Dorsey's latest project, or are we all heading for the Metaverse? The Metaverse is an interesting thing if it just means virtual reality. I think that the folks from Facebook who rebranded themselves Meta are trying to relabel things to their own advantage. That makes perfect commercial sense, but the idea of a virtual space which you can interact is much older than that. And actually, I am pretty bullish on the future of such virtual spaces. I predicted quite a while ago that eventually we would all have our own little virtual realities just like we all have our own webpages. We don't all have our own webpages, but we use them all the time, right? There are technological and social hurdles to making them super duper common. It's a little bit more clunky right now to use those spaces than to use regular webpages, but I think that that will happen.

2:01:39.1 SC: But anyway, I think it's a separate question from the social media question. I don't know what the future of social media is really going to involve to be very, very honest. I think that Twitter is in a slow decline. There's more noise on there. They're doing silly, dumb things with verified accounts and letting bad actors back on the platform and flooding timelines with tweets you don't want and advertisements. Overall, it's a less pleasant experience than it used to be. I'm slightly surprised no one has just come out with a clone of Twitter. Honestly, if you just made a clone of Twitter with an edit button and didn't mess with it that much, people would flock to it. But the alternatives that are out there like Mastodon and Post and so forth, they all want to make improvements. And the improvements aren't always what people want. Mastodon is a good idea that it's just so clunky to get on it and to use it and so forth.

2:02:39.0 SC: So, I don't know. I think that there is a feature of social media that it works best when there's a lot of people on it. That means that the winner is not necessarily going to be the best one. It might just be the one that, for whatever reason, most people end up joining. And therefore, I find it very difficult to know what that will be. There's not a lot of stickiness. If you do get a lot of people to join your social media network, there's no reason why you can't just kill off the old ones. Myspace died. So there's no reason that just because Twitter exists now and is fairly popular, or Facebook exists now and is fairly popular that they will still be around 10 years from now. Not that hard to imagine people switching. I love the idea... I don't understand a lot about the Bluesky project that Jack Dorsey has. It seems to be an idea where individuals control their knowledge, their data, their information and then they can choose to share it as they wish on different social media sites, which if that is, it sounds great. People need to make money. People want to make money anyway. And I think this is what Elon Musk doesn't understand about Twitter.

2:03:55.9 SC: The value in Twitter is the people tweeting on it. And you can sell those people to advertisers. Somehow, he thinks that the people tweeting on it should be paying him for the right to tweet. It's like, asking magazine contributors to pay for the right to publish in your magazine. That's just not the way it works. And if he tries to make people pay for it, they're gonna switch to go somewhere else. I don't know. So in other words, is my short answer, but I am very open to the landscape looking very, very different five or 10 years from now. Michael Kramer says, could you say something about your choice of Desert Island episodes of the podcast? Then he goes on. It's a longer question, but he's asking what are the episodes of the podcast that personally had an effect on me? As a principal I kind of don't like to rank my favorite podcast episodes.

2:04:47.7 SC: I think it's kind of unfair to all the guests. I very much appreciate that all the guests who I've had on the podcast have taken their valuable time to be on the podcast and to talk and to share their thoughts with a lot of people. So don't I want to rank them. If you rank them, then someone's gonna end up last. Someone's gonna end up not being noticed. I do have a couple that I do think made a big impact on me in different ways. I'll just mention one, 'cause I used to mention it all the time and I haven't mentioned it, I think, in a while, which is Joe Walston talking about urbanization. This was a fairly early podcast. Joe Walston is a conservationist. And he tries to make the point that the fact that there's an increasing percentage of people who live in cities in the world is really good in many ways, because the more people live in cities, the more the rest of the land of Earth is left for nature in some sense.

2:05:47.5 SC: And so, he says that we can have an equilibrium with a very large population here on Earth as long as we live densely as human beings and we preserve the habitat for the rest of the ecosystem outside. He doesn't mean that you ban people. People actually like living in cities, so you just make it pleasant and possible for them to do so, they naturally do so. The world is becoming more and more urban. And Joe's point was that in many ways, that's good. That can be remarkably good and maybe if we get there fast enough, we can save a lot of species from going extinct. So that's one, anyway, that I would bring up that is not an obvious choice. Some of the obvious choices are just kind of obvious and good. Ravi Ivaturi says, in your most recent book The Biggest Ideas in the Universe: Space, Time and Motion, I love the way you built up Einstein's field equation from first principles. I assume these equations don't yet include the cosmological constant.

2:06:43.1 SC: I'm curious to understand the conceptual leap needed to apply these equations to the entire universe, which led to the introduction of the cosmological constant. Actually, it's really not like that. There's no conceptual leap needed to apply these equations to the entire universe. Einstein did it pretty quickly. He had the equations from the start in 1915 and by 1917 he was writing papers about cosmology. It's just a matter of choosing a form for the spacetime metric that would apply to the whole universe. Likewise, the cosmological constant is absolutely part of the field equations I wrote down. It is true that historically, the way Einstein invented the idea of the cosmological constant was to say, is there a different term you could add to the equations that somehow relates to the curvature of spacetime? But he recognized pretty soon and we certainly recognize very well now, that that term is absolutely the same as adding an energy density to empty space.

2:07:43.9 SC: That's why vacuum energy and cosmological constant are completely equivalent. Adding energy density to empty space is just a particular form of energy along with all the other forms of energy. So it was already implicitly there as one of the allowed things you could put in the equation from the start. You don't have to add anything to it at all. Jameson Rader says, how does causation work in a block universe? I'm not sure what you're aiming at here, just because I'm not sure how you think causation works normally. Or, I guess I should even say normally, but in a non-block universe. There's one question, which is how does causation work? That's an important question for philosophers and scientists to think about. I'm honestly not sure why living in a block universe would have any effect on how you thought causation worked at all.

2:08:35.7 SC: The way that I think of causation is as a higher level emergent phenomenon, as we've been talking about in the podcast already. If you look at the equations of general relativity, the Standard Model of particle physics, or whatever, there's no term for causation in those equations. Causation is not a separate thing. Causation is a way that we human beings have of conceptualizing what happens according to those equations. Even in Newton's laws, F = ma, force = mass × acceleration, and force due to gravity is Gm1m2/r2. You can remove force from the equation entirely. You can just relate the distance you are from a heavy object to the acceleration due to gravity that it causes on you. The word cause never appears in the framework. To me, the question is, why is the concept of causation useful to us at the higher emergent level? 'Cause it certainly is.

2:09:37.4 SC: I'm actually writing a paper about that, and I've given a couple of talks. You can find them online if you look hard enough, about the relationship between entropy and the arrow of time and cause effect relationships. It has to do with the fact that the information we have about the universe changes over time, as we've talked about earlier, because entropy is increasing, etcetera. The way that it works is that given the fact that entropy is increasing, there are facts about what happens at any one moment of time that could be one way or the other, given previous moments. Not everything is predictable, both because of quantum mechanics and because of our ignorance about the microstate of the universe. And which actually happens, propagates into the future and affects it, but does not propagate back into the past. It's ultimately because of the arrow of time, is the short answer. But I think that all the words I just said there apply equally well for presentists, as well as for block universe people, so I'm not exactly sure what you're aiming. Hope that answer reveals something that you were looking for.

2:10:38.3 SC: Jason Richiardi says, I'm currently reviewing Bell states and their importance in quantum networking in my quantum computing course. So I was astonished when I heard ChatGPT's haiku on Bell's theorem in episode 230. ChatGPT is also excellent at coding activities. Do you think large language models will reduce the need for software developers in the workforce? Well, I think that the most obvious prediction to make, which is not necessarily going to be true, but is the historically sensible one to make is that the introduction of new technologies will not reduce the overall need for jobs. It has never happened. New technologies have gotten rid of certain occupations, but they've also opened up the need for new occupations. So I don't think that LLMs will reduce the need for software developers, but they might radically change what the average software developer actually does on a day-to-day basis.

2:11:36.7 SC: You can already, for those of you who don't know, I've actually been playing with this, because I am a very bad coder. I mean, back in my day, in my salad days as an undergraduate, I was very good at writing computer programs. Some of the code that I wrote is still used to analyze data in the Villanova astronomy department, but that was a skill that completely atrophied. I didn't use it that much once I got onto theoretical physics and so forth. But I've always kept up a little bit. I learned a little bit of Python, the very basics. Some plots that I made have appeared in papers that I've written, etcetera. So I was fascinated by the idea that you can now use these AI programs to help you write code. And so, in VS code, which is the environment in which I play around with these little Python programs, you can have an extension, which will basically give you AI-generated suggestions about what's coming next. And that's not just I write a line of code and it guesses what the next line of code is.

2:12:38.6 SC: You can write in the comments, I would like to write a program that plots the energy eigenvalues for the simple harmonic oscillator as a function of position. And it will just write the whole code for you. And that's kind of amazing and absolutely game-changing for people who write software for a living. But you still need people, because [chuckle] well, let's put it this way, the program, the AI program makes a lot of mistakes. It doesn't always work, okay? And maybe it will get better and better. But I think you will always want a human being both posing the question about what code you want to be written and then checking that it's actually doing what the human being wanted it to do. The large language models are a tool, right? Until they get to the point where they're separately conscious agents, which we're not there yet, they're tools that you can use, but there are still gonna be human beings deciding how they should be used.

2:13:32.9 SC: Philip Malinowski says, this question relates to a recent paper in Nature. Papers and patents are becoming less disruptive over time. In a nutshell, does the decrease in disruptiveness something... Is it something you observe? And especially what signs do you see? Do you have an interpretation of why that might be happening? A couple of hypotheses that come to my mind are maturing of science, there are fewer low-hanging fruits and replication crisis, fewer reliable insights. So I don't have very informed opinions about this paper. It is a paper that I've heard about, but not read myself. I've seen what seemed to be very reasonable sounding critiques of the title of the paper, okay? Papers and patents are becoming less disruptive over time. The idea, as I understand it, is that they studied... That they characterize a paper as disruptive if it sort of was the beginning of something.

2:14:31.1 SC: The later papers that cited it, cited it but no previous papers, okay? And it didn't cite that many papers itself, it started a whole new thing, so it was constantly being cited by later papers. I have no idea whether that kind of methodology accurately captures an intuitive notion of disruptiveness, but okay, we do what we can. And I don't know... This seems like an almost impossible question to ask just because different sub-fields of science are so very, very different, right? Or even of engineering. So, there seems to be a giant selection effect here. It's like asking who's going to win the next presidential election, but you ask your friends in the yoga studio that you go to in Brooklyn versus you ask your friends who are farmers in Mississippi. There's going to be a giant selection effect depending on who you're talking to. It might not be a reliable sample and I don't even know how to make a reliable sample that covers all of science when you're asking is science becoming more or less disruptive over time.

2:15:36.4 SC: It's absolutely true that individual fields do mature and low-hanging fruit does get picked and it becomes harder and harder to make a truly disruptive contribution. So therefore, it's very possible that you get a result like this just because you are oversampling the mature parts of science, right? You don't even notice that there's this new idea, new whole area of science, the physics of machine learning, like maybe you didn't include that in your study. I just don't know. So I don't have a strong opinion about the reliability of the result. My guess is that we have more and more mature science as time goes on, so there are more and more fields in which you don't overturn the entire apple cart all at once, but we also just have more and more fields.

2:16:24.2 SC: We have more and more things going on in science, so my guess would be that it could be simultaneously true that there are more and more disruptive things happening even though the fraction of things happening that are disruptive is going down. That I have no data on it all, just a general feeling. Nelita S says, would you explain in your own words how does looking at distant galaxies allow us to look back in time? Do we know the age of the photons by their wavelengths only? Well, there's really two aspects to this question, the really easy one and the extremely subtle one. The very easy one is, whenever you look at anything, you're looking back in time because it takes time for light to reach you. The old Mitch Hedberg joke is people show me a photograph of themselves and say this is what I looked like when I was younger. Every photograph is what you looked like when you were younger, even though it's just a little bit younger, okay?

2:17:17.3 SC: Of course, when you look at astronomical objects, they are further away, so it takes light much longer to get to you, moving at the speed of light, one light year per year. And therefore, the effect is much more noticeable than if you just take a selfie, which was still in the past, but not that long in the past. The subtle part of the question is, how long did it take? What is the age of this thing that you're looking at back there? Remember, we just talked a little bit about in special relativity, you can't even say what is simultaneous between here and some distant place. So how are we even talking about the age of distant galaxies and how do we know what that is?

2:17:55.8 SC: A big part of it is that even though this doesn't always get emphasized in discussions of special and general relativity, in cosmology, in the universe in which we live, even though there's relativity and you can use whatever reference frame you want and different observers will observe different things, nevertheless, there is an obvious standard reference frame to use, one that is at rest with respect to most of the galaxies in the universe and the cosmic microwave background radiation, etcetera. There is a cosmological rest frame. So when we talk about the age of the universe or the age of distant galaxies, we implicitly mean in that universal cosmic reference frame. It's not the only one we can use, but it's extremely convenient and so that's what we always end up using. Then the question is how do you know what the age of this distant thing is?

2:18:50.4 SC: Well, what is easy to measure relatively is the redshift. We know that when light is given off by different physical processes that involve atoms, there are atomic spectra that have very well-known, highly calibrated wavelengths in their rest frame. So when you see not just a thermal black-body spectrum, which is featureless and hard to characterize, we don't exactly know the relation... Well, if you knew exactly what temperature a black-body was and then you observed it and redshifted, you would know what the redshift was, but usually you don't know exactly what that black-body is. But if you see a spectral line from a transition in an atom, then you know exactly what its wavelength was and you can figure out its redshift. So that tells you the amount by which the universe has expanded since the light was emitted. That's not exactly the distance.

2:19:42.8 SC: Hubble explained to us that it is related to the distance, but if it's very, very far away, then that relationship might also be subtle. That's why it was difficult for Hubble to actually show that there was a linear relationship between distance and redshift, because redshifts were easy to measure, but distances were hard. So how do we know the age of the photons? Well, it depends. It depends on how far away they are. These days, we think we have a very accurate model of exactly how the universe has expanded over time. So we can still try to directly measure the distance to things. That's what we do when we try to measure these cosmological parameters like the Hubble constant and the cosmological constant and so forth. But if you just want a pretty good answer to the age of the universe when a certain galaxy, etcetera, was giving off its light, all you need to know is its redshift and the cosmological model can be used to then say and the universe was this old when it gave off its light that you're seeing right now. So that's generally how we would end up doing it.

2:20:49.0 SC: Julian Howe says, are there examples of different kinds of infinity showing up in physics and how would you differentiate them? Well, it depends on what you mean by showing up in physics? There are physical models of the universe, physical theories that may invoke different kinds of infinity. At least at the very simple way, there are two kinds of infinity that are relevant. The countable infinity of the integers and the uncountable infinity, which is bigger, which is characteristic of the real numbers and things like that. Now, of course, mathematicians have been able to define a whole bunch more infinities that are also uncountable, but in different and subtle ways. As far as I know, those uncountable infinities... Well, let me think about this. There certainly are theories of quantum mechanics and quantum field theory where you're doing a path integral over all possible quantum paths, etcetera, which would in principle involve infinities even bigger than the cardinality of the real numbers.

2:21:50.8 SC: But that's all theoretical, right? That's all our construction of models. That is not something you can experimentally observe in the world. So in terms of the... I was saying that there are two issues here. One is our theories. Yes, our theories might very well involve different kinds of infinities. The other is the world. Does the world really involve different kinds of infinities? There, we don't know. There there's no super strong evidence. Our theories, as you know, map onto the world in certain way, under certain circumstances, in certain approximation levels, etcetera. But we don't have any experiment that would actually distinguish between a theory that was only involving countable numbers versus uncountable continuum or something like that.

2:22:36.5 SC: So, I don't know if we ever would even be able to do that. Of course, some people had asked whether or not secretly our use of infinity is unnecessary, whether or not the real world is fundamentally discrete and finite in number of things that are happening in it. But again, if it is, that number of finite things that are happening is so large that infinity is a pretty good approximation for our purposes. Relatedly, we have the next question from Tim Converse, which is is our quantum Hilbert space infinite dimensional or finite dimensional? If this is not known, what considerations or arguments are in favor of one side or the other? It's not known, is the short answer to that. Many physicists will say it's infinite dimensional for the following reason.

2:23:20.1 SC: If you make a model of a very, very simple continuous classical system and then you quantize it, so you have a simple harmonic oscillator or the electromagnetic field or anything that is continuous classically, and then you quantize it, the quantum Hilbert space will be infinite dimensional for that. Even just a single particle moving in one dimension has an infinite dimensional Hilbert space as long as that dimension that it moves on is smooth, okay? So surely, the reasoning would go, once you have a big universe with many, many particles, you're going to need an infinite dimensional Hilbert space. The problem is that we don't know whether that model of the universe as a smooth manifold is good once you have quantum gravity, etcetera. In fact, there are indications that it's not.

2:24:09.3 SC: The best indication is from black holes. We know that the entropy of a black hole from Bekenstein and Hawking is a finite number and it is proportional to the area of the event horizon. It's roughly the area of the event horizon in Planck units divided by four, depending on how you put the two pi's in there, in the Planck units, etcetera. Maybe you can argue, although this is a little bit shakier, that what that entropy represents is entanglement between inside the black hole and outside the black hole. And maybe you can argue that the entropy of a black hole is the maximum entropy that a system of that size can have, which means there's a maximum amount of entanglement that you can have and that maximum amount is still finite if the black hole entropy is finite.

2:24:55.7 SC: And therefore, the actual Hilbert space of the quantum system that you're calling the black hole would also be finite. And even though that only applies to black holes strictly, it implies an upper limit for any region of space 'cause a black hole is the most entropy you can fit into that region of space. I actually wrote a short paper arguing this with Ning Bao and Ashmeet Singh saying the Hilbert space of quantum gravity is locally finite dimensional. It was just one of these little essay contest things. It's not a major research paper, because I think a lot of people already believe that this is true, but for the purposes of the essay, we wanted to marshal all the evidence in one place so people could sit down and think about it because it's really not known. All you can say, by the way, is locally finite dimensional. In any one region of space all you need is a finite dimensional Hilbert space to describe what's going on.

2:25:46.0 SC: Are there an infinite number of regions of space? Something we don't actually know. So for many reasons, we don't know about the Hilbert space of the whole universe. I think it's the kind of thing we should care about a lot, but that's just me. Nicholas Fokrot says, in the Everettian multiverse accepting that some universes are those in which very rare things happen, coffee cups phasing through tables, oxygen molecules collecting on one side of the room, etcetera, are some universes inherently more likely for any other reason than chance to be those in which very rare occurrences occur? Put another way, could the fact that something ultra rare happened in the universe mean that that specific universe is more likely to be one in which such things happen? Of course, we don't know the final super duper answer to this question, 'cause we don't know all the laws of physics, but in our current understanding of many-worlds and how quantum mechanics works, the answer is no.

2:26:40.0 SC: So just to make it very, very clear, Nicholas is asking if you tend to be in one of those rare branches of the multiverse in which unlikely things are seen to happen, does that imply that unlikely things will continue to happen? And the answer is absolutely no. They're completely independent from each other. And this is important because I think that one of the reasons why people struggle with believing or giving a lot of credence to many-worlds is they instantly home in on the weird unlikely worlds where unlikely things happen and they're made uncomfortable by that. It's very much like people who didn't like Boltzmann back in the 1800s saying that entropy just usually probably increases. And they're like, every time I've seen it, it always increases. That's because the words likely and probably are used in a very, very dramatic sense here.

2:27:34.1 SC: The probability that a bunch of very unlikely things happen in the Everettian multiverse is just really, really tiny, just like it is in ordinary quantum mechanics. So focusing on those rare unlikely branches of the wave function gives you a bad impression of what the Everettian multiverse is really like. And there's no reason to think, in anything we currently know about physics, that if you happen to be in a branch of the multiverse where a measurement came out in one of the very, very unlikely outcomes, that would continue to happen for other measurements or other branchings of the wave function.

2:28:09.2 SC: Anonymous says, I want to be a Bayesian and I feel inadequate unless I can justify my beliefs with priors and evidence. When I can't find a working prior, I panic. I don't think you are as obsessed with retconning a prior that fits your conclusions. I think you're satisfied when you have a frequentist or other pragmatic justification for your conclusions. Should we be okay with beliefs when we can't come up with a Bayesian justification for them? An example of a non-Bayesian argument that you make is the probability of the data of a needlessly big universe, given weird hypotheses like no low entropy past, is small or cognitively unstable. Therefore, this weird hypothesis is false. I'm not sure that I quite think you're using certain Bayesian terminology correctly here. It might be that I'm just misunderstanding here. Bayes' theorem says that you have a prior credence or probability that you're giving to various propositions and then it tells you how to update your credences when new evidence or information comes in. That's all that happens.

2:29:11.5 SC: The idea of Bayesian epistemology is that that's how knowledge works. We start with a whole bunch of credences for different propositions, then we update them when we get new information coming in. If only we actually did that, that would be great. Of course, the reality is a little bit messier. If you know nothing, if you're just starting out and you have a proposition that, I'm going to flip a coin and it's going to come up heads or tails, and my friends tell me it's a fair coin, I have a prior probability before I collected the data, before I actually flip the coin and see what happened, I have a prior that it will be 50-50 heads or tails. Then when I flip the coin and I gather some data, if I flipped it 100 times in a row and it's heads every time, my prior is going to change. I'm going to update that, and guess what? It is no longer a prior.

2:30:03.3 SC: It is a posterior. It is a probability that you're associating with a certain proposition, but it's not your prior anymore. The prior is what you have before you've collected any data. It's not a general word we use for all of our beliefs. Almost all of our beliefs have already been updated by a whole bunch of data that we have collected. The pros and cons of Bayesian reasoning are almost all in the choice of a prior. People who are pro-Bayesian will say, look, as long as your priors aren't completely crazy, if you collect enough data, the priors cease to matter. If you had a prior that I think I'm 99% sure that this coin is not a fair coin and you're going to flip it and it's going to be heads every time, then you just collect the data. You just flip the coin over and over again and you see one way or the other. In practice, that can be hard sometimes, so we are left with large uncertainties. That's okay. The promise of Bayesian reasoning is that data overwhelms your prior ultimately.

2:31:03.4 SC: And therefore, there is no algorithmic way of choosing what your prior should be. It's a little bit fuzzy to say when things are priors and when things are posteriors because we all have certain inclinations, intuitions, pictures of the world, and that's perfectly okay. But as a good Bayesian, you shouldn't be too worried about picking your priors. You should be mostly worried about updating those priors when data comes in, when information comes in. Nevertheless, to answer, there's a hidden question there at the end where you say that when I talk about randomly fluctuating universes and their cognitive instability, I'm being non-Bayesian, I don't think I am being non-Bayesian. I'm making an argument for giving a very, very, very tiny prior to universes in which we are randomly fluctuated out of the chaos and all of our memories are completely unreliable, but I still am very much working with a Bayesian paradigm there.

2:32:03.5 SC: Nick B says, the commodification of brain data, which is an idea that appeared in episode 229 with Nita Farahany, is perhaps the most terrifying phrase I've heard on any of your podcasts. Unlike many of the doomy scenarios that you have discussed over the years, this feels urgent and unfortunately inevitable. As greater abilities to read brains develop, can you see a way for us to access the technologies that are useful without corporations mediating the experience in an intrusive way? Well, I am kind of on your side that this is very terrifying, [chuckle] but it's going to creep up. It's not like the corporations are going to invade our houses and attach electrodes to our brains and scan them. They're going to offer us the chance to purchase brain scanning electrodes and we are going to buy them and then we're going to send them our data, because it is slightly more convenient. This already happens.

2:32:57.0 SC: This happens with credit cards, it happens with social media, it happens when we buy things online. We are being watched. It's not quite 1984 level watching, because for what it's worth, if this is some kind of comfort here, the corporations don't care about you as an individual that much. It's not like some person in the room is tracking your purchases. The computer knows what your purchases is, but the corporation cares about the whole statistical distribution of purchases, etcetera. But I think that empirically, we human beings are pretty willing to sign over our privacy for a small increase in convenience.

2:33:38.6 SC: And so, we've been doing it already with all those other examples. So when it comes to buying headphones that will let corporations read our electromagnetic activity in our brains, yeah, we're going to do that. We're going to give it away. For what it's worth, I think we have to rely on government interference here, if we want to be saved from the worst abuses of things like that. I don't think that there's much incentive on the side of the corporations to not scan our brains and use all of the data. They don't gain anything by not doing that. There will be a race to do that. As Nita Farahany was trying to do, we need legislation that prohibits that, or at least we need some principles, even if you don't have detailed legislation. You want to establish some legal principles that will let people have countermeasures, whether it's suing or whatever, when they think that their private brain data have been abused.

2:34:38.6 SC: I suspect, I'm not an expert here, but my guess, my prior, is that there will be some terrible cases of abuse. It will be interesting to see whether or not people actually react against it, or whether we decide collectively that it's okay, that our brain data's privacy is not that important to us as long as we make it easier to play video games without using our hands. That I don't know, but I don't see lots of reasons to be optimistic about it. That's why I think that podcast is very important. It talks about things in a much more down-to-earth and specific way than we've talked about before, and it is something that is potentially pretty worrying. Derek Bain says, what do you think of Baltimore so far? Be nice, it's my hometown. I really like Baltimore. The Baltimoreans have a great attitude in many ways. It is certainly, and this is very subjective, but I've lived in a bunch of suburbs and also a bunch of big cities. People in Baltimore are the nicest people that I've ever met. You walk into restaurants or whatever, you just do things in the city, and people are just really nice to each other in ways that I maybe should have expected, but maybe living in bigger cities has spoiled me a little bit.

2:36:00.4 SC: And they have a little bit of self-deprecation there. There's a motto that was invented by someone that has become very popular on bumper stickers, etcetera., which says, Baltimore, actually, I like it. Baltimore is not one of the famously sexy and vibrant cities in the United States. If you're a foreigner, you don't live in the United States, and you think about what are the big fun cities to go to in the US, you might say New York, LA, San Francisco, who knows, Austin. I don't know these days. But Baltimore is a solid East Coast city. I think it slides under the radar, I guess is what I would say. As I said earlier in the podcast, it's big enough. There's a lot going on. There's far more going on than I can ever actually experience. The beautiful parts of it are beautiful. I happen to be lucky enough to live right near campus, which is one of the beautiful parts, the campus of Johns Hopkins, so that's very fortunate. The restaurant scene, which is very important to me, is great.

2:37:02.5 SC: There's a lot of music and theater and not an inconsiderable bonus, it is on the Acela corridor, so they can hop on the train and go to Washington, DC, or Philadelphia, or New York, or Boston. I've already been doing that, and that's a wonderful thing. Of course, it is in many ways there are issues in Baltimore. It's not an especially wealthy city. There's poverty, there's homelessness, like any other big city. The roads have potholes in them. Maryland and Baltimore governments historically, have not been the most effective at getting things done for the citizens, etcetera, so there's plenty to complain about. I don't know if there's more to complain about per capita than in other big cities or anything like that, but roughly speaking, yes, I like it very much. You can afford a nice house in Baltimore that you couldn't afford in Los Angeles. Academically and intellectually, Hopkins has been amazing, so, so far so good for me as far as Baltimore is concerned. Dan O'Neill says, as an author, what qualities do you value in an editor, a literary agent, a publicist? These are great questions, and it's going to depend a lot.

2:38:11.0 SC: I can give you a little bit of my own feelings, but it's going to depend a lot on what kind of author you are, what your audience is, what your writing style is, etcetera. These days, there's a feeling that editors... I'm not even sure when to say the word editor and when to say the word publisher. Let's back up a little bit. Let's imagine you want to write a trade book. I did, by the way, if anyone's interested, do one of the holiday special podcasts, short solo podcasts, on writing trade books and the process of it. You should check that out if you want. You need an agent if you want to write what we call a trade book or a popular book. I think a lot of academics make this mistake, because if you're a professor of something, you can probably get your book published. If you're already a professor, getting your book published is not the big hurdle, 'cause you're a professor and you can go to an academic press. You can go to Harvard University Press, Princeton University Press, whatever.

2:39:11.4 SC: They are not primarily devoted to trying to make best-sellers. They like it when they make best-sellers, but they really want to publish books that are academically respectable. That's not to say they will publish any old thing, but if you're already a professor, probably you're competent enough to write something that is respectable and you can get it published. That doesn't mean people are going to buy it. If you want to publish with a commercial press, they're going to be much better at getting people to buy your book. They're also not going to listen to you unless you have an agent. If you actually want to write a book that people will buy, a substantial number of people will buy, your first step is always to get an agent. If you have a book idea and you pitch it to random publishers, to Random House or whatever, they're just going to ignore you. They don't pay attention to you.

2:40:01.5 SC: But if you pitch it to agents, they will pay attention to you because that is their job. Their job is to filter through all the pitches for trade books and decide which ones are promising. By the way, they sometimes make mistakes. I've pitched to many, many agents before one would accept me. A good agent is one who will try to get what you want, to try to fight for you. They are your advocate in this whole process. Then you have the publishing company where there's a person at the publishing company who will be your editor there. That's your editor working for a publisher. That's why I always hesitate between the words editor and publisher. You can also hire an editor, which is a separate thing. You personally can pay for someone to edit your manuscript, because a lot of people who are called editors working for publishing houses are more about acquiring books than about actually editing them.

2:40:55.1 SC: My own editor, Steven Morrow at Dutton, who has been my editor for all of my books, he actually edits the books. We absolutely collaborate on shaping a book into something good. I got very, very lucky there. What you look for there in an editor is someone who will edit, who cares about the book. Ideally, someone who will edit in a way that you are compatible with, but it's hard to know until you've gone through the process. But you can get an idea just by talking to them. As far as publicists are concerned, if you publish your book with a major publishing house, they will have publicists on staff. I don't have a publicist other than what the publishing house does. I've been pretty lucky, I think, although it's very much a crapshoot is my impression. As a podcaster, I get pitched a lot of people to appear on the podcast by publicists working for book companies, for publishers.

2:41:49.6 SC: And I think that there's certainly a mode of being a publicist, which just means whenever you have a book coming out, you send out a form letter to all the podcasters and all the magazines in the world saying, do you want to talk to my client? It's not very high effort, but it's also not very effective. A good publicist will know that there's certain venues that match your book and develop personal relationships with the people at those venues and work to make sure that the right authors are connected with the right publicity venues. Again, hard to know ahead of time whether that's true, but any major grown-up publishing house will at least strive to do that.

2:42:30.5 SC: So, it's a lot of a black box because you have to sign a contract, etcetera, before you know what it really will be like to work with your editor and your publicist and things like that. You can talk to other people who have already written books, but still you're going into the dark a little bit then you see how it works. Nicholas Weiberg says, if dark matter could only interact gravitationally even with itself, would we still expect it to lump together? If so, is this a fact of general relativity or would it also happen in Newtonian gravity? You do expect dark matter to lump together, but not as much as ordinary matter. And the reason is that in order to lump together, you have a bunch of particles that are almost perfectly smoothly distributed, but not exactly smoothly distributed. What's going to happen? Gravitationally, it's going to pull together particles in those regions that are over dense and it will clear out the regions that are a little bit under dense, but then the particles start moving toward the center of the over dense region and because they don't interact other than gravitationally, they just zoom right on by each other.

2:43:42.6 SC: You might think that to some very, very good approximation, you don't really lump together, 'cause they don't stick. They don't hit each other and dissipate like ordinary matter does. Real atoms can bump into each other, lose energy by emitting photons and therefore stick together or at least stay in the same region of space. Dark matter can't do that. It's collisionless. It just slides right through. However, there is a more subtle effect, which is that a whole bunch of particles, even if they're just gravitationally interacting, can come together and lose energy collectively by spitting out a few particles. You have a million particles and mostly they condense into a dense region by conserving energy, by spitting out some high energy individual particles so that most of the particles are lumping together.

2:44:31.7 SC: That is not as efficient a process as literally bumping into each other and radiating in terms of losing energy. That's why galactic halos of dark matter are bigger and puffier than the condensed disk made of ordinary matter. If you think about what we know about the structure of galaxies, the picture you take of a galaxy with all the stars and the spiral arms, etcetera, is embedded in a bigger, puffier dark matter halo, which is exactly what you would expect if the dark matter is collisionless. Again, one of the reasons why the vast, vast majority of working cosmologists believe in dark matter, because the universe looks like it would look if dark matter really did exist. That's nothing to do, by the way, with general relativity. It would be just as true in Newtonian gravity.

2:45:21.2 SC: Joshua Hillerup says, In principle could a living being, like a cat or a physicist in a some sort of well designed container that has things like an independent oxygen supply, be put into superposition for an appreciable period of time? Or is there a hard limit to what or how long could something be in a superposition? This is a tricky question because it depends on what you feel about the foundations of quantum mechanics. As usual, I will give you the answer for an Everettian version of quantum mechanics, but if you believe in objective collapse of the wave function, you will get a completely different answer here. For Everett, it's trivially easy to put something into superposition for an appreciable period of time. There are literally a huge number of universes in superposition and they are very, very different. Some of them are very, very different. Many of them are very close and very similar to each other, but many of them are completely different from each other.

2:46:15.4 SC: Maybe what you mean by the question is, could you put two things into superposition on a single branch of the wave function? The reason why that's a difficult thing to answer is, what do you mean by a single branch of the wave function? To me, the most useful meaning to that is to imagine that we have, or at least the most intuitively graspable meaning to that, is to imagine that we have a clear distinction between system and environment. That's the crucial setup that we need to explain decoherence, as I talk about in something deeply hidden and elsewhere. What happens is, if you have a macroscopic system like a cat or a physicist, it's interacting with the rest of the world. It's interacting with all the photons, all the air molecules, etcetera and we don't keep track of exactly everything that is happening in that rest of the world. We lump it together and call it the environment.

2:47:11.9 SC: And my definition of branching of the wave function is when the system becomes decoherent by becoming entangled with the environment. That's when the universe branches. In that case, you cannot put a cat or physicist into superposition on a single branch of the wave function, because they keep bumping into the environment around them. Even if you put them in vacuum, they're going to radiate. They're going to give off infrared radiation because they have a certain warmth. That's why quantum computers, etcetera, have to be cooled down to close to absolute zero to prevent them from interacting with all the thermal bath of photons everywhere around them.

2:47:50.0 SC: If you have a living being, try to put it in a superposition of being in two different physical spatial locations, for example, it will instantly radiate in different ways depending on whether it's in one spatial location or another. Therefore, it will branch. Therefore, on a single branch, you do not have a superposition anymore. You effectively have a collapse of the wave function onto those two different branches. It depends on what you mean by superposition. This actually goes back to what I said earlier about the paper that we wrote on branches of the wave function not containing firewalls. Where is their entanglement? Where is their superposition? Depends on whether you're asking about the whole wave function of the universe or just a single branch thereof. Matthew Cushman says, the Boltzmann brain paradox and the simulation argument seem to me to have a similar structure. Under some reasonable-ish assumptions, most observers with our experiences are ontologically abnormal.

2:48:45.6 SC: Do you think they are similar and pose similar challenges and have similar resolutions? There's definitely a similarity there. In fact, in the philosophy literature, both Boltzmann brain scenario and the simulation argument would be classified under skeptical scenarios. Not that you should be skeptical of them, but they are scenarios in which our immediate impression of what the world is like is dramatically off, dramatically wrong for some reason. In one case, because we randomly fluctuated into existence and all of our opinions randomly fluctuated. In another, because we're being tricked by the existence of a simulator layer that is giving us the impressions of the world around us. But the details of these two skeptical arguments are very, very different. For one thing, the Boltzmann brain scenario is way better defined. We know exactly what the probabilities of various things happening are because by construction, the Boltzmann brain scenario is thermal fluctuations in an eternally existing physical configuration.

2:49:48.1 SC: We know how to characterize those things quantitatively, so we know exactly what to predict, what to expect, etcetera. Whereas the simulation argument is entirely ill-defined. We don't know what the simulators can do. We don't know what they're interested in doing. We don't know what the probabilities are of different things happening etcetera. I think that there are both similarities there and differences. It's interesting to think about them. Of course, none of this is new. It goes back to brain in the vat thought experiments or Rene Descartes imagining that he was just being dreaming, because he was being tricked by mischievous demon that was fooling him. The idea of skeptical scenarios has a very long philosophical pedigree. Preston Justich says, keeping in mind that you identify to some extent as a Kantian constructivist, I have a philosophical question about the relationship between normative ethics and meta ethics. Would you agree or disagree with the claim that in order to do normative ethics responsibly or to make normative claims about what one should do, that we must first not take for granted meta questions about normative claims.

2:51:02.2 SC: Instead, we need more background questions about the status of moral claims themselves, which produces a stronger meta-ethical foundation prior to even making claims of what one ought or ought not to do. Even if we're relying on our intuition as our starting point in the form of weak premises per se, is there a more concrete way to do that responsibly? Well, I think that there are details here that matter. For one thing, I'm not a Kantian constructivist. I'm a Humean constructivist. I talk about the distinction between those two ideas a little bit in the big picture. Constructivists think that morality or moral rules are not just located out there in the world. They are constructed by us human beings. The difference between a Kantian and a Humean version of them is the Kantian will say, and there is one uniquely right way to do that. There's a correct rational way to construct our moral rules. The Humean will say, different people are going to construct different moral rules, and we can't say that any one of them is right or wrong.

2:52:02.2 SC: We can talk to each other. We can try to persuade each other. Maybe we'll argue that someone else is being internally inconsistent, etcetera. But we can't just say, objectively, you're wrong about your moral rules that you've constructed. I'm on the Humean side there. As far as your actual question is concerned, I have mixed feelings, I would say. I would rephrase it as this. I'll answer the rephrased question because if I misunderstand it, then I don't want you to wonder why I'm answering a different question. We have ethical rules. Don't kick puppies for no good reason. That's an ethical rule. Then we have meta-ethical rules, which are the reasons why we accept or construct certain ethical rules. Meta-ethical standards are like, I'm a moral realist. I can find moral truths out there in the world, or I'm a constructivist. I'm a utilitarian or whatever. I shouldn't say utilitarian. That doesn't count. Utilitarian is ethical, not meta-ethical. But if I have some reason why I'm utilitarian, that would be the meta-ethical stance.

2:53:09.2 SC: I think, I don't want to be too harsh. I think that it's actually quite similar to physics. In a different case, there are obviously differences because there is an objective right or wrong to the predictions you make as a physicist in a way that there's not with the ethical rules. But I think that as a physicist, if I say that I have a rule for when I put my cup of coffee on the table, it does not fall through, someone might say, well, okay, what is your justification for that in terms of the quantum field theory and the Pauli exclusion principle and the composition of the table and the cup? I think I would argue that the relationship between ethical rules and meta-ethical rules is like the relationship between this everyday folk physics idea that the coffee cup is not going to fall through the table and its underlying micro foundations.

2:54:07.4 SC: Namely, it's nice when you have the underlying micro foundations, but you can get through the day pretty well without them. Even if two people disagree pretty severely on their meta-ethical principles, but their ethical principles match pretty well, I'm happy to get along with such a person. I might disagree with their foundational intuitions about why they're doing things in a certain way, but I'm going to say that they're doing the right things. That's more important to me. Justin Wolcott says, would I be wrong to say that the odds aliens have discovered how to travel faster than light are the same as aliens having discovered how to perform magic like witchcraft, not just like illusions? Yeah, I think that that's wrong, but I think that it depends on what you mean by wrong. This is a question that we should think about as good Bayesians. What is the probability, the credence we have, that aliens would be able to discover that we can travel faster than light? What is the probability that they could be discovered how to truly do magic?

2:55:16.5 SC: Both probabilities are very small. In that sense, they're smaller than epsilon, where epsilon is 0.01% or something like that. If you want to just lump together all credences smaller than epsilon and call them the same, then it's the same. But there are degrees of small numbers. If one number is 10 to the minus 5 and the other is 10 to the minus 10, then the 10 to the minus 10 is smaller than 10 to the minus 5 by a factor of 10 to the minus 5. I think that the ability to do magic is way less likely than the ability to travel faster than light, even if the ability to travel faster than light is still super duper unlikely. What I mean by that, the reason why I think that is because I can imagine small modifications of the known laws of physics that would let you travel faster than light. I don't think that those modifications are true. I don't think you actually can. I think that they're very unlikely, but it's not difficult to imagine slightly different laws of physics that would allow it.

2:56:16.9 SC: Whereas with just doing magic, I don't know how to modify the laws of physics in a simple way to make that permissible. Jacob Asmith says, my girlfriend is really into poetry and really enjoys both reading and writing it. Do you have any favorite poetry books, poets, or poems? Thanks. Jacob, it's okay to admit that you like poetry. You don't have to blame your girlfriend for this question. Everyone should like poetry a little bit. My own relationship with poetry is very much like my relationship with classical music, which is that I really do enjoy it, but I am not in any sense an expert, knowledgeable, or do I spend a lot of time engaging with it? I feel slightly bad about that. I think that engaging with good poetry requires a bit of effort. You have to put in some time into reading and reading very carefully. It's just hard to find that time these days, man. There are other things pressing on one's attention. But I do like it. I get a chance to read poetry books, etcetera. Poetry is very different. It's almost hard to say I like poetry.

2:57:28.3 SC: There are certain poets that just don't do it for me at all, and others I quite enjoy. Many of the classics I enjoy, like Dante or Homer, make for amazingly good audio books. Homer, anyway, was originally oral tradition poetry. If you're going to drive across the country, it's hard to do better than listening to a good audio book of the Iliad or the Odyssey. The Inferno, Dante was not meant to be read out loud, but it really does work very, very well being read out loud. Of course, I listen to a translation, not the original Italian. Among more modern poets, I enjoy a crazy incoherent selection of people. James Merrill is someone I really like, which is a little bit weird Adrienne Rich, Wendy Cope, who writes these silly little fun poems. John Ashbery, slightly older, E. E. Cummings. Yeats, of course, is just super amazing. I'm not going to claim any super expertise about who the great poets are, but there are people who I absolutely do enjoy reading, yes.

2:58:43.3 SC: P. Walder says, in the postscript to How Physics Makes Us Free, Jananne Ismael encourages the reader to consider themselves as a complex system that has arisen through the process of natural selection. As such, the associated microscale structure is arranged to support the emerging complex system. Does this imply a downward causation process where the natural selection process operating at the level of a complex system causes the appropriate microscale structure to arise? I think that questions about downward causation are tricky, and I shouldn't dismiss them. In this case, I'm just going to dismiss them. Roughly speaking, the answer is no. I think that questions of downward causation will depend a lot on what you're taking the lower microscopic level to be. If you're taking it to be literally atoms and molecules and quantum fields and so forth, the core theory, then there's no downward causation going on. Because if I ask what an individual electron does, the information I need to answer that question is literally what is exactly the local environment of that electron?

2:59:51.1 SC: What is the value of the electromagnetic field of other fermionic fields that are nearby? That is all I need to know. I do not need to know anything about the larger context in which is it embedded. I don't need to know if it's in a rock or a brain or interstellar space or whatever. I need to know the value of the other quantum fields at exactly that location. There's no room for downward causation to do anything. That doesn't mean it's not helpful to we human beings. If I know that a certain thing is a snowflake, then that will instantly give me some information about the microscopic structure of the water molecules that are going into making up that snowflake. That doesn't mean that the water molecules themselves care about the fact that they are in a snowflake, but it helps me understand it a little bit. I think it's very different if you're talking about the relationship between, let's say, a human being and society.

3:00:48.2 SC: When human beings are the microscopic level not the macroscopic level, human beings don't behave nearly as locally and microscopically as water molecules do. They know about the context that they are in. It's a different story. Maybe, although I'm not really sure, maybe downward causation is a useful idea in a situation like that. David says, I read the New Scientist Essential Guide article entitled Beyond the Big Bang from Einstein's Universe Essential Guide Number 10. You are quoted as saying, we used a half-assed version of quantum mechanics when we do cosmology. What would a full-assed version of quantum mechanics tied with cosmology look like? I didn't remember that quote, but that's a good quote. I'm glad you brought that up back from the archives, David. I think that was a long time ago, but I still agree with it. I think there's two senses in which we use half-assed quantum mechanics when it comes to cosmology. One is the underlying attitude towards quantum foundations, and the other is what it is we are actually quantizing.

3:01:53.2 SC: Let's do the second one first. Mostly when we do cosmology, like when you're predicting the density perturbations in the cosmic microwave background, we're not doing quantum gravity. We don't really know the rules for quantum gravity in these circumstances. People have tried, again, Hartle and Hawking with the wave function of the universe, and other people, Alex Flanken and Andre Linde and others, have tried to do a quantum gravity story about the origin of the universe and the associated inflationary fluctuations, but we don't know what the rules of that game are, so it's a very difficult game to play. Instead, we imagine that the universe is expanding, and there's a classical description of spacetime, the background spacetime, and then we quantize fields on that classical background spacetime. These fields that we quantize include small fluctuations of the gravitational field that allows us to predict the existence of what we call tensor perturbations, gravitational wave perturbations caused by quantum mechanics in inflation. We can do that, but you see how it's half-assed, right? We're quantizing some things, we're not quantizing other things.

3:03:00.9 SC: The other half-assed aspect is that, what do we mean by quantum mechanics? Do we mean fully ever-ready in quantum mechanics? Do we mean Copenhagen? Do we mean something else with hidden variables? This was exactly the question that caused Hugh Everett to come across the Many Worlds interpretation, because John Wheeler, his advisor, asked him to quantize gravity, and he thought about the quantum theory of the whole universe, and he came to the conclusion that the Copenhagen way of talking about quantum mechanics with a quantum system being observed by a classical measuring apparatus was inadequate to the purposes of studying the universe as a whole, so that's why he invented Many Worlds. Nevertheless, people today still use Copenhagen-ish quantum rules to apply to the early universe in a way that is entirely inappropriate.

3:03:57.7 SC: The Copenhagen rules are perfectly fine when we're looking at the kinds of experiments that we look at in particle physics or atomic physics or other things that inspired the Copenhagen rules in the first place, but they just don't make sense. When we talk about the early universe, there's no one there making measurements. A big part of the traditional standard textbook quantum rules are what happens when a measurement is made. There's no one making measurements in the early universe, and Everettian has no problem with this. Everettian says there's no such thing as measurements. What there is is decoherence and unitary evolution of the full quantum wave function. People will say in the early universe, there is a probability distribution that the density fluctuations look like a certain thing. Where does that probability distribution come from?

3:04:45.9 SC: I get what the probability distribution is via the Born Rule for measurement outcomes, but there's no one measuring things in the early universe. I literally wrote a paper about this with Kim Boddy and Jason Pollock. For better or for worse, we looked at how you should correctly talk about quantum fluctuations during eternal inflation. People talk about them as if there was some magical wave function collapsing device that happened once every Hubble time, and there isn't. We asked, what would the correct way of doing it be? I should have predicted this ahead of time, but we got the answer, which is it's almost no difference. You can do it correctly and you can get numbers, and they slightly are different than the numbers you get by doing it in a half-assed way, but it's not a big difference. We could have gotten lucky and found that there was a big difference between doing it correctly and doing it in a half-assed way, but in that case, we did not. That's the life of a scientist, so be it. Daniel Donaldson says, in the biggest ideas in the universe, you brought up the opposing views on space, substantivalism, and relationalism.

3:05:56.3 SC: You noted that today, most physicists would side with treating space as a thing itself. However, you did not indicate a preference yourself, which you normally do. Which side do you feel is more valid or likely? Understanding that in the book, you also noted that we don't have the final answer. Well, I'm in some sense not sure that this will turn out to be a very relevant question in some sense. The idea of substantivalism, space is a thing all by itself versus relationalism. Space is a convenient way of characterizing a whole bunch of relationships between things in space. I'm not entirely sure that that's a good distinction to make. That's a relevant distinction to make. I'm not sure what difference it makes, but that is a more esoteric philosophical question. I think that this question relates to more down to earth physics questions about what is space. I think that in my view, space should be thought of as a convenient way to talk about relationships between different subsystems in the quantum mechanical wave function and in that sense, I'm closer to being a relationalist than a substantivalist.

3:07:06.8 SC: But in classical general relativity, it sounds very substantival. Space and spacetime are things and they have their own dynamics and so forth. The most correct answer to this question might very well depend on your level of analysis in some sense. I think that probably deep down, relationalism is going to win, but for everyday physics, predicting gravitational wave signatures from LIGO, substantivalism might be the way to go. Ariane Malek says, is there an objective definition of entropy not in terms of emergent macrostates? Laplace's card shark, which I guess Ariane just invented, wouldn't see entropy increasing in shuffles, because all shufflings are equally ordered there. Would Laplace's demons see objective increases or decreases in entropy in the forward reverse direction since it always sees the fully specified configuration of the wave function for the universe?

3:08:03.2 SC: The short answer is no, there is not an objective definition of entropy not in terms of emergent macrostates. Entropy is by construction in classical statistical mechanics, it is a higher level emergent concept. If you are Laplace's demon, Laplace's card shark, Laplace's whoever, you just know the microstate of the universe. You don't need to talk about entropy because you know everything. Entropy is a coarse-grained phenomenon. Whether you do it in what we call, ex post facto, the Boltzmannian way, which is to define macrostates and make equivalence classes of microstates if they look macroscopically similar and then s equals k log w is the equation on Boltzmann's tombstone, or we do it in what you would call the Gibbs Shannon way and say I have a probability distribution over microstates and s equals minus the integral of p log p in some sense.

3:09:00.8 SC: Either way, there is a lack of information that is important into the definition of entropy, so Laplace's demon wouldn't suffer from that. However, there is one little footnote here, which is that quantum mechanics is a little bit different. The von Neumann entropy of a subsystem is a perfectly well-defined kind of entropy, which says in quantum mechanics, because of entanglement, I can have perfect information about a composite system, a system that has subsystem A and subsystem B. I know exactly the quantum wave function for the whole system, but because of entanglement, there is no such thing as, it's not just that I don't know, there is no such thing as the unique quantum state of the subsystem. There is a density operator rather than a wave function, and that density operator can be thought of as a statistical superposition of different, or a statistical combination, I should say, of different pure wave function states.

3:09:58.5 SC: But it has an entropy, and that entropy is objective in the sense that we can calculate what it is everyone agrees on it, what it is, even if you know the complete wave function, the exact quantum state of the combined system. So in quantum mechanics, subsystems have an objective entropy, even if the overall system does not, and even Laplace demons would agree with that. Jeffrey Seagal says, the discussion with Raphael Milliere was very stimulating. Near the end, he expressed the valid concern that human rights and sapient machine rights might come into conflict. I suspect this question will become real very soon, since it seems likely that the GPT-4 connection numbers roughly match the number of synapses in the human brain, 100 trillion to 1 quadrillion. I would be interested in your thoughts on the issue of human versus machine rights. I think this is an important issue, a complex one. I don't think that the comparison of synapses to connection numbers is a very direct one, so I don't think that's immediately relevant.

3:11:00.2 SC: I think that we're going to have to think more carefully about what we mean by an agent, a person who deserves rights. It's not something that legally or even casually we think about very carefully. Legally, we need to define very carefully what we mean by a person, but we do so about, is a certain human being competent? Should we count the collection of human beings like a corporation as a person? Things like that. We do not go very far into asking whether non-human beings are agents, are individuals, are deserving of any kind of rights. I am of the opinion that right now, even with GPT-4, which is one of the more advanced AI large language models out there on the market, I don't think that we're even close to saying that something is generally intelligent or an agent. Not just is there hallucination on the part of these large language models. They say things with great confidence that are completely untrue, but there's an ability to extrapolate away from the information that we know about to contemplate, to be creative that they don't have yet.

3:12:19.7 SC: The large language models are able to do remarkably human-like things because they are trained on what things that have actually been done by humans. There's going to be much more that you need, I think, to qualify something for rights. For example, and I've said this before many times, the large language model doesn't get upset if you don't talk to it. It doesn't get annoyed. It doesn't get frustrated. It doesn't get bored. There's many features of actual living beings that are not in common with these large language models. That is not to say that they could not be instantiated in artificial beings of some sort, or even to say that that won't happen relatively quickly. I don't know what the criteria should be for giving rights to some artificially generated intelligence. I do think that it's going to require some thought. I think that the tricks that large language models can do right now are not up to the task, but maybe we're close.

3:13:27.2 SC: I really don't know. The pace of progress has been very, very fast, but I would need to think about it more myself before I would say anything with confidence about when they should count as persons. Nikola Ivanov says, you mentioned in one of your podcasts that you were working on some ideas as to why the notion of position and not, for example, momentum is so important in the macro world we live in. Can you please elaborate on these ideas? Yeah, I wrote a paper somewhat recently with Ashmeet Singh, former graduate student who is now a professor. It's the paper on quantum mereology, which is asking the question if you just have the Hilbert space and state vector in it and you start somewhere and let it evolve, is there a right way to divide up all of Hilbert space into subsystems, which you might call the system under consideration and the environment?

3:14:21.8 SC: And the main criterion we used was that we wanted the system to have a classical limit. We want to be able to explain things in the real world like the fact that you and I in the tables and chairs around us act pretty classically. We came up with some criteria for doing that. Along the way, a very important role is played by how the environment monitors the system. If you think, as we already talked about, if you think about a cat or a human being in a room with light photons bumping off of the system and air molecules, etcetera, there is a very, you can write down the interaction between the environment and the system. What we know in the real world is that when does a photon of light interact with you? Well, when it hits you, it bumps into you in space. That's a crucial feature of what we call space.

3:15:15.4 SC: Interestingly, if you study classical mechanics, if you study the versions that I explain in. The Biggest Ideas in the Universe, Volume 1, you'll find a formulation of classical mechanics called Hamiltonian mechanics, which treats position and momentum on a completely equal footing. The fundamental equations of Hamiltonian mechanics do not differentiate between positions and momenta. That's a little bit alien to us because we are used to thinking of positions as fundamental. You get a momentum by taking the derivative of position with respect to time to get the velocity. You multiply the velocity by the mass to get the momentum. Therefore, momentum is a derived quantity and position is a fundamental one. Interestingly, in the Hamiltonian version of classical mechanics, that's just not true. There's a complete equivalence between position and momentum, but in the real world, there's not. One can ask the question, where does this distinction between position and momentum in the real world come from?

3:16:16.3 SC: And what we found, it's very tentative and we're trying to think about how to push it forward a little bit more. What we found was that in order to get classical-looking behavior of a subsystem of a larger quantum system, there needed to be a specific way that the system is being monitored. Basically, the environment monitors some variable in the system. You know, again, this is a volume two of the biggest ideas thing, but we also talked about it in something deeply hidden. There's something called the uncertainty principle in quantum mechanics. Position and momentum cannot simultaneously be exactly defined. What that means is that if you think about what you mean by position and momentum, they can't both be being monitored by the environment at the same time.

3:17:07.5 SC: So, the fact that something is being monitored by the environment in order to allow a quantum mechanical system to act classically implies that you have to pick one position and momentum. Again, a feature of the Hamiltonian formalism is that you can call one variable position and one variable momentum or even have some combination of position and momentum act as your fundamental variables. So the point is that of these possible things to measure, you pick half of them. And that is what is being monitored by the environment. And that is what expos facto, we call position. We call location. In space, you're only allowed to monitor one thing, and that's what we call position.

3:17:47.9 SC: Now, it very much depends on the laws of physics on the Hamiltonian itself that governs the laws of physics in the quantum mechanical picture, that there is even something that would give rise to a classical limit at all. So we're not saying that this necessarily happens, we're saying that it's a very interesting thing from the Hamiltonian perspective, classically by itself, without quantum mechanics, the Hamiltonian could be any function you like of position and momentum of these two variables. But our suggestion, which is very vague and we're trying to make it more firm, is that the kinds of Hamiltonians that arise as classical limits of quantum mechanical Hamiltonians do distinguish between position and momentum.

3:18:34.3 SC: They pick one out to be monitored and to give rise to a robust classical limit. I have no idea why the real world allows for a classical limit at all. That's another level question. But I think if you want the classical limit, there's going to be a preferred variable. And we're going to call that the position, that is my current belief about this question. But again, still working on it. Jim Murphy says, "I've watched your biggest ideas videos, but I'm still a bit lost on how we describe a particle in quantum field theory in particular. I'm confused when you talk about the modes of the field, I'm thinking of it as breaking apart the field into a bunch of sine waves using Fourier transforms. But there could be an infinite number of sine waves required to describe the field. This leaves me wondering how breaking apart the field into those waves could describe different numbers of particles. Which of the sine waves are relevant when determining whether the field describes one particle or two particles and so on?" And this is a very good question.

3:19:30.7 SC: This is like a tricky thing that I'm trying, I'm struggling with right now in writing the book version of the biggest ideas videos that the volume two will be on quantum quanta and fields, and mostly field theory and particle physics. So why does a quantized classical field look like particles to us? And I think the confusion here is that it is, well, let's back up a little bit. The classical field, as you correctly say, can be decomposed into a bunch of sine waves using Fourier transforms. So rather than giving you the value of the field at every point in space, I can give you the value of the field, the contribution to the field configuration from different sine waves with different amplitudes and different wavelengths. And in quantum field theory, what happens is that for each mode, for each wavelength that is contributing to the classical field, it can have its own quantum state. You can separately, and this is true in the limit where the field is not interacting or being complicated.

3:20:35.2 SC: That's why quantum field theory is hard, because there are interactions in the real world. But if you just simplify your life and look at free fields that are not interacting, each mode behaves all by itself. And it behaves in such a way that it looks like a simple harmonic oscillator and a simple harmonic oscillator has energy levels, the vacuum state, the first excited state, the second excited state, and so forth. And so what we mean by a particle in quantum field theory is a quantum state, which has the first excited state of a number of modes. So it's not that if I take two modes and I construct a quantum state where one mode is in its first excited state, and another mode is in its first excited state. That's not two particles, that's one particle whose wave function is a super position of those two modes, right? So basically when you create a wave packet or some thing that has a spacial profile that is representing a single particle, it is a superposition of many different modes each in their first excited state. If there are many different modes in their second excited state, we would call those two particles.

3:21:45.0 SC: And you can actually do a little bit of manipulation to explain why that resembles what you and I know as two particle systems and so forth. If you have a superposition where some modes are in their vacuum state and some other modes are in their first excited state, you would call that a superposition of zero particles in one particle. Okay? So there is a translation, there's a dictionary that goes back and forth between the quantum field theory language and the particle language. And that's kind of one of the miraculous fun things about quantum field theory. Paul Larkin says, "There's much discussed in physics about extra dimensions of space, eg, and string theory having nine space and one time dimension. Why do we con generally consider only one time dimension and not multiple? Is this an area of research and if so, in what ways? And if so, why not?"

3:22:29.4 SC: It is an area of research. It's a bars at USC is the person who has really spent the most time thinking about multiple time dimensions. But naively it just doesn't work. Everything breaks down. When you have two time dimensions, you can go in loops in time very, very easily, right? You can just walk in a circle in time rather than going forward in time. There's also, generally speaking, and maybe you can work hard to get around this, but generally speaking, the energy of the system is unbounded below. So nothing is stable, everything runs away and it's just a complete mess. It looks nothing like the real world. So intellectually you can just sort of ask, can I really work hard to make a version of this? That does look like the real world, but it's hard and it's pretty unconvincing and there's no real motivation for it. There's no experiment that we've done that that is better explained by having multiple time dimensions than not.

3:23:23.7 SC: Whereas with spatial dimensions, it's very easy to hide them. You just make them small and then you can't see them and you get away. That's not true with extra time dimensions. And there's motivation for the extra dimensions of space, namely that they're demanded by string theory. And string theory is a finite theory of quantum gravity. So there's some motivation for doing it. So there are good reasons out there why we spend a lot more time talking about extra dimensions of space than extra dimensions of time. [3:23:51.4] ____ I'm not sure if I'm pronouncing that correctly, but I'm trying says in from eternity to here, you have a picture to illustrate closed time-like curves that reminded me so much of braid theory. Is braid theory at all involved in the analysis of CTCs? Short answer, no.

[chuckle]

3:24:08.2 SC: Long answer. Maybe that's a little bit, you know, it's two syllables rather than just one. You know I don't know a lot about braid theory. When we draw pictures of closed time, like curves. We are necessarily being a little bit hand waving. You know, we're gonna draw some of the closed time like curves, but not all of them. Whenever you draw a picture of a space time and try to make it visually clear to the seer, the observer of the picture, then you're making some compromises. And in the real world, it's hard enough to make one set of closed time-like curves, right? Like one topologically distinct region of spacetime, which is foliated by closed time-like curves to make it interesting topologically so that you have, you know, some closed time-like curves that braid around others in some interesting way. Well, that's even harder. And the motivation once again as for the previous question, the motivation is pretty low for something like that. So, as far as I know, there's not been any direct important, application of braid theory to the analysis of closed time-like curves.

3:25:15.0 SC: But, you know, if we discover some or if we have a good reason why it should be there, then things will be different. Achi Locus says, "Have you ever played Dungeons and Dragons or other tabletop role playing games? And if so, what did you think?" Sure. You know, I mean, I'm pretty old. I'm the original generation. Of course, I played D&D when I was in junior high school and maybe even high school. I forget exactly what the years were, but you know, around then, we're talking late '70s, early '80s, right? That was the first wave of enthusiasm. And I liked being the dungeon master more than I liked being a player. And, you know, we didn't always obey the rules exactly and things like that, but we had a lot of fun. When I see and and hear about people playing it now, I haven't played it for many years or even decades.

3:26:02.2 SC: I am always struck by how, I mean, probably there's a huge selection effect, right? That I'm hearing about people who are really into it and really good at it, but they're way better storytellers than I ever was. You know, as a dungeon master, I was really devoted to coming up with clever puzzles and tricks and challenges for my players to solve, but not, you know, telling a convincing narrative story so much. So I think that's how I would be different if I were, were doing it now, but, you know, again it's an investment. You have to like put some time into it. I have no objections to the idea that I would ever play it again, but I have other things going on right now, so it's... I don't see it happening anytime soon. Andre Benu asks a priority question, "You have remarked in the past, the quantum field theory is a local theory. Quantum field theory is based on partial differential equations. Therefore, the properties of a quantum field at any given point depend only on the properties of the fields in its immediate vicinity.

3:26:57.5 SC: However, Bell's theorem is an indication that quantum mechanics in general is not local, at least not in it is not compatible with local hidden variable theories. Quantum field theory is only one type of quantum theory, but it is the most advanced version of quantum mechanics. How is it then possible? The quantum field theory is a local theory, but quantum mechanics in general is non-local?" Good. I like this question because it's easy to answer, it's an important question, but it is one which the answer is very clear, namely that when we say the quantum field theory is local to be perfectly honest we're not talking about quantum mechanics. We're talking about the unitary evolution, the evolution of the, of the quantum field theory when it is not being observed. You know that in quantum mechanics in general, whether it's quantum field theory or anything else, we have separate sets of rules for what the quantum state does when it is not being observed and when it is being observed.

3:27:50.3 SC: In many worlds, what we mean by being observed is just [3:27:53.5] ____ and branding, etcetera. But still, that's okay. We can use those words. Bell's theorem is entirely about correlations of measured quantum states. You can't do Bell's theorem without talking about measurement and collapse of the wave function. So when you say quantum field theory is local, that is a statement about the dynamics of the theory while it is not being measured. When you say Bell's theorem, you're making a statement about correlations between quantities when they are measured. So there's absolutely no contradiction there. There is talking about two very different things.

3:28:31.7 SC: David Maxwell says,"You seem to do so much full-time professor podcaster continually learning more broadly. I honestly don't know where you find the time. Are you willing to give some estimates of what you spend your time on, both professional and personal? Is part of it that many of these things are done in your leisure time?" Yeah, I guess that's part of it. Look, you know, for better or for worse I don't think that... I don't like to talk too much about, you know, my system or my way of doing things because I think it's bad for other people to think that there's some right way to do things. There's many different good ways of doing things and people need to find their own way, find what works for them. Sometimes people ask me like, what is your routine for writing a book? And I got, I just gotta laugh. There's no routine. Like, whenever I can find the time and the inspiration, then I work on the book.

3:29:19.0 SC: That's my routine. In terms of dividing things up, you know, the podcast is a special case. Like the podcast is a weird thing, which is not really academic work, not even really book writing, which is kind of semi-academic, certainly it's not research, it's intellectual work. You know, I do pick people to have on the podcast because I want to learn something from them. And discussions on the podcast can bleed into my more scholarly academic interests, but it's not something that is recognized by the establishment as academic work. And I work hard to the limit my podcasting time. I work very, very hard to make sure that on average from week to week, I don't spend more than one full day's worth of effort on the podcast. In terms of other things you know, book writing and whatever, that just comes and goes very dramatically.

3:30:13.9 SC: There are times when it's book writing time and, you know, weeks go by when I don't really do anything other than work on the books. There's other times when I'm being a professor and I'm teaching and you know, I gotta write that lecture, gotta read those papers, grade those papers, talk to those students. And the crucial part of the question, the interesting part of the question here is how many of these things are done in your leisure time? And, you know, look, I'm in a very, very privileged special situation where there's not a lot of hard and fast distinction between my leisure time and my work time. I love what I do for a living and I'm very, very lucky to be like that. I'm not someone who thinks that you need to love what you do. I think that there's plenty of people who very respectably and legitimately work because they need to earn a living.

3:31:01.2 SC: That's completely fine and we should honor that. But I am lucky enough to be able to earn money doing what I love doing. So I spend a lot of time doing the various things that I do and you know, whether it's reading or thinking or you know, now relearning python and programming with help from AI extensions and so forth, you know that's fun for me. I like it. The podcast is fun for me. So, I don't draw a hard and fast distinction between my leisure time and my work time. Sometimes my leisure time is, you know, cooking dinner or sitting down and watching Colombo or whatever and that's clearly pretty clearly leisure time. So I do have leisure time, don't worry. But I don't really count up how much of it I spend or how many hours of the day are devoted to that.

3:31:48.1 SC: Jesse Remmer said, "I enjoyed your podcast episode with Raphael Milliere on AI. At one point, Milliere says that GPT's architecture is not comparable to the human brain in terms of how it actually functions. That is a neural network does not actually work the way our neurons work. Then he goes on to speculate how we may use these networks to learn about the human mind, whether we are born with innate grammar, for example. Isn't this a contradiction? We don't use airplanes to study the flight of birds. Well, I don't think it's a contradiction." I just think it's, the lesson here is that it's not going to be easier, obvious, or straightforward. That doesn't mean it can't happen. This is why it's work to do this kind of research. There's a lot that we do know about the human brain and a lot that we don't know about it.

3:32:31.2 SC: Neural networks are not exactly the same as the way that your neurons work in your real brain, but they were inspired by it. You know, the neurons came first before neural networks did and people looked at how the neurons work and said, hmm, you know, maybe we can make something like a computer program that works in a similar way. They don't work in exactly the same way, but there's a relationship there. I think what Raphael is probably getting at is the fact that, as we discussed earlier in the podcast, we train our neural networks, our deep learning network or whatever it is and it learns and it gains capacities to do certain things. And we can talk about the capacities it has and then once that happens, we can go in and look, it's actually hard to do, but in principle, we can go in and look and ask what's going on in the neural network?

3:33:18.8 SC: What is happening inside when it is parsing a sentence or something like that. There is no guarantee that what we find when we do that is directly applicable to what goes on in the human brain. But there's enough open questions about what goes on in the human brain that we can say, can we be inspired? Can we get an interesting idea for what is going on in the human brain that we might not otherwise have thought of before? You know, it's much like I think about the physics of democracy. It's not that I think that you should think about democracy like a physicist. It's that maybe you get inspired by thinking about things we've learned from physics when we're also trying to understand this very, very difficult thing, which is how a democracy or society works. I think it's a very similar thing in the relationship between neural networks and actual neurons.

3:34:06.9 SC: Casey Mahome says,"This is a bit of a basic question, but I realize that I don't fully understand how solid objects work. Something like water makes sense to me because the molecules are bound together electrically, but in general, the molecules bounce off of one another. I suppose I can also understand things like plastic being made of long molecules that can rotate around and get stuck on each other. But how does something like a chair stay together? Are the intermolecular forces really that strong?" I'm not gonna be able to give you the world's best answer to this. This is not the kind of physics that is my personal specialty, but I think that you've already answered it. It is intermolecular forces and it does have something to do with the particular molecules that you're talking about. If you're talking about something like, why is iron solid?

3:34:49.5 SC: It's roughly speaking because the molecules take on more or less a crystalline lattice kind of shape inside the iron. It's individual iron atoms, okay? And they arrange each other. Why? Because of the tiny electromagnetic forces between the different atoms. You have to go into some work to show exactly what those, what the sizes and shapes and functions of those electromagnetic forces are. Because iron atoms are... Have a little bit of complication in them. They're not super complicated, but there's a little bit of complication there. Likewise, for ice, when we have water molecules or something like that, whereas in a chair, if your chair is made of wood rather than iron, then it's more like plastic, right? Because wood is an organic material and there's cellulose in it, and those are long molecules, and they gain tensile strength by braiding together and wrapping together in interesting ways.

3:35:40.9 SC: There's also this separate question of why atoms and molecules themselves take up space and that goes into deep questions about the poly exclusion principle and firm and quantum field theory and spin statistics and things like that. But roughly speaking, I don't think you should be surprised that objects are solid because you know when, let's put it this way. You might think the forces between individual molecules are very small, so how do they create a large, solid object? And the answer is, there's a lot of molecules and they build up, and you have to run the numbers. You can't just say the forces between two molecules aren't that big, therefore it makes no sense that the table is solid, right? You have to be a little bit more quantitative and careful than that, and I promise you that if you are, you will find that your table is solid as it should be. It does not violate the loss of physics.

3:36:32.8 SC: Robert Zanelli says, "You've written an interesting paper on extreme of black holes, in an extreme of black hole temperature and surface gravity goes to zero, although it is unlikely a black hole can become extreme rule. Assuming this is possible, what are the gr predictions for the gravitational field inside the merged horizons and at a distance beyond the merge horizons?" So, I'm not sure how to answer this question. I included it in the questions being answered because Robert worked very hard to get... He was struggling with a Patreon being obstreperous and breaking the question into little pieces. So, you know, I think it deserves an answer, but the answer is, you know let's back up a little bit. An extreme old black hole, for those of you who don't know, is an example of a black hole where the mass is being balanced by something else.

3:37:24.2 SC: So the simplest example is electric charge. Okay? If I have two black holes and they're both positively charged, then the gravitational field is going to attract the two black holes together. But the positive charge on both is going to repel them. And you can ask, you know, when does it balance? When does it become equally big? And it turns out that it balances When you have what is called an extreme old black hole, there is a largest possible amount of electric charge that you can put into a black hole. And when that happens, it's called extreme hole. In fact, we think that in practice, that's never going to happen. I mean, not just a as physically, but if you try to make one, you would have a lot of difficulty. Think about it, I have a black hole that is positively charged. I'm trying to charge it even more by throwing a proton into it.

3:38:10.4 SC: But that proton is being repelled very strongly by the electromagnetic field there. So it's hard to make extreme of black holes. Nevertheless, they're theoretically very interesting. So I wrote a paper with Matthew Johnson and Lisa Randall, where we discussed some puzzles about the limit that you take to get to an extreme old black hole. It turns out that if you start with a black hole that is not quite extreme old, so it has some electric charge, but not enough to be extreme old. And then you look at the spacetime diagram of that almost extreme old black hole versus the spacetime diagram of an actually extreme old black hole, they look very different. There's a whole region.

[laughter]

3:38:53.3 SC: Where it's present in the non-real black hole, but it's not present in extreme old black hole. And between us, Lisa and Matt and I called this region Who-ville after the Dr. Sue's book. We didn't put that into the final paper, but when we were writing the paper, we called it the Who-ville paper, and we showed that this whole region sort of discontinuously disappears when you hit the extreme limit. And that's part of why it's hard to make an extreme old black hole. And maybe it helps explain some of the weird features of extremal black holes. But, uh, to the actual question being asked here, what are the gr predictions? I mean I'm not sure what you want. We have an exact solution that you can find in my general relativity textbook for what the metric and therefore the gravitational field looks like inside an extreme old black hole. You know, if you are yourself not electrically charged, then you just fall into it and there's a horizon and you can't get out just like any other black hole.

3:39:49.0 SC: So the gravitational field for uncharged objects is just not that different, especially outside, the internal structure is more complicated. But as we discussed when we were talking about Roy Kerr's question, no one should believe the internal structures that you get from taking overly seriously these exact solutions to general relativity. The real world is gonna be messier than that. Julie Mars says, "I enjoyed the last interview with Raphael Milliere, the emergence of a, the new breed of AI tools are exciting and thought provoking. It appears to me that we are experiencing something like the advent of personal computing and the internet. In this case, though, the changes appear to be vastly accelerated in comparison. If I'm optimistic about these new tools, what are your thoughts on the subject after speaking with Milliere? You know, I do think it's important to keep in mind what Raphael and I were talking about was this question, which is at the intersection of computer science and philosophy, which is to what extent should we think about these existing AI models as thinking, okay, we didn't really delve into the technology questions of how these tools are going to be used.

3:40:57.9 SC: We did not delve into what are the dangers, what are the prospects? These are all very interesting questions. But we already gone on two hours in that podcast and so we didn't cover those. That's okay. Yes, you're right. I agree with you, Julie, that the changes, the rate of change because of these new AI tools is much faster, seems to be much faster than the rate of change was when we introduced personal computing or the Internet. Part of that is a bootstrap effect, right? Like the rate of change that we would be witnessing if AI had come before the Internet might appear to be much slower. But because of the Internet, we are aware of these changes that are happening very rapidly. And also something like the rate of change of progress is always hard to quantify, right? We might find a 100 years from now that we are still just getting started right now in these years right now.

3:41:51.1 SC: But I do think, so I'm very bad at predicting technological change. I think I've told this story before, but when the first web browsers came on the scene NCSA Mosaic, the first graphical web browser and then Netscape after that, I was very much excited by this. This is the 1990s. I was very much enthusiastic and convinced that this was gonna change the world. And I tried to convince other people that it was gonna change the world, and I completely failed because when they said, well, how will it change the world? I was very bad. I was unable to articulate why that would happen. I explained that you could order a pizza over the internet, and they were like, I have all the technology, I need to order a pizza already. I do not need to do it using a fancy bit of technology.

3:42:45.4 SC: So I am convinced that these AI tools will be pretty much ubiquitous. They're gonna be everywhere. They're gonna help you write, they're gonna help you program, but they're also gonna help you compose music. They're gonna help schedule airlines and buy tickets and plan your vacation and many things. But I don't know what, I don't know what I admit I have enough humility to admit that I'm bad at predicting that. So I can see that they will be everywhere, that they'll be in all sorts of tools. They already are, right? You know, Siri and Alexa are examples of AI things. Even Google search and Bing and things like that are AI based to some extent. Those particular examples are intentionally throttled down. You know, you can't get the best AI in Siri or Alexa, because we don't yet know what the best AIs will do under certain circumstances.

3:43:43.4 SC: I think there's a lot to worry about in terms of the safety and things like that. I think that we... It's almost hopeless to imagine banning certain kinds of research or forbidding certain kinds of AI in certain kinds of situations, because human beings are just always gonna get around that's just a almost hopeless task. I do think and hope that we should think seriously about what we let the AIs do, what we do to preserve safety and things like that. I'm not at all convinced we're going to do that, just as we said earlier, that we're very bad at protecting our own privacy. We're also very bad at slowing down the rate of cool progress, because maybe it'll be dangerous. We human beings just aren't good at that. So I think it's a legitimate worry that things are, that they'll be disastrous consequences, even as there are also very good consequences.

3:44:34.9 SC: You know, there's still ongoing debates about whether social media and smartphones, etcetera, make people sad overall. Like obviously there's, they're hugely beneficial for some reasons, but there's at least some evidence that might be on the right track that overall they make our quality of life worse. So, I don't know, and I don't, I don't think we're very good at stepping back, taking a breath and looking at the long term implications of things carefully. We're just short term bumbling into the next cool thing kind of people. That's what we do. Not I'm not saying we should do it that way, but that's what we do. And I bet that's what we're gonna do with AI. Robert Weinman says, "Philosophy has caught my interest. Hegel is often quoted, so I decided to read some totally incomprehensible, I read people's analyses and they seem reasonable, but I still can't make sense of the stuff Hegel wrote. Do you have any suggestions?"

3:45:25.6 SC: I did discover what Popper thinks of it, so I don't feel too dumb. Well, you know, I'm a big fan of philosophy overall, but I am also absolutely convinced that philosophy has an enormous dynamic range in the sense that, you know, the worst physics papers out there are still pretty okay, whereas the worst philosophy papers out there are just anti-useful. Are actually harmful. Whereas the best philosophy papers are amazing, really crucially important. So Hegel is a you know, the standard bearer for a kind of philosophy that I personally find completely useless. I have zero interest in anything Hegel ever said. Someone like [3:46:05.1] ____ much of what he said, I struggle to see why it is important or useful or comprehensible, but others I can see why yes, you absolutely need or this is absolutely insightful and helpful.

3:46:17.7 SC: So the fact that you didn't get anything out of Hegel, I wouldn't worry about that at all. It depends very much on what questions you are interested in ahead of time. Different philosophers have different things that they care about. You know, if you're interested in Bert Russell, you might not be interested in Simone de Beauvoir, you know, who knows? So I think that you have to rather than just saying like, let's read philosophy, you have to ask, ask yourself what are the philosophical questions that I would like the answer to? You have to have a little motivation there. You can, it's gonna be hard to read just about any philosopher. And in terms of, you know that's why I don't have suggestions. I think it's gonna depend very much on what you care about. You know, if you read more or less contemporary, let's say 20th century philosophers, whether they're living people like David Chalmers or recently living people like Karl Popper or Bertrand Russell, or Simone de Beauvoir, or Wein or Putnam or whatever Richard Wardy, you know, just depends on your interests.

3:47:20.3 SC: They will generally engage with previous philosophers. So you can kind of work your way backwards, right? You can read some relatively recent people, figure out who the ones that you like are and then dig into the ones who they talk about, right? Who they think are interesting and important to engage with. Or you could go the other way. You could just read Plato and Aristotle, or you could read the classics of Chinese philosophy and Daoism or Confucianism or Indian philosophy or whatever. See what engages with you Buddhist philosophy. Because there's no like set five things that everybody should read. It's gonna depend a lot on what you care about. Finally, of course, you know, there's more specific things, not... You know, I have this idiosyncratic belief that the overlap between science and philosophy is actually way more important than it has been taken to be.

3:48:19.0 SC: I think that and now I'm just sort of babbling, so don't, it's been a long AMA here, so don't take me too seriously here. But you know, there have been absolutely attempts in the history of philosophy to make it more scientific, right? The logical positivists famously tried to make science the paradigm for doing philosophy. And I don't think that's the right answer, but I do think that there should be way more interaction between scientific insights about the world and philosophical insights about the world. I mean, that's why I'm doing what I'm doing with my new job at Johns Hopkins. I think there's a lot of, speaking of low-hanging fruit as we did before, here's a place where there's a lot of low-hanging fruit. I tend to think that in current contemporary philosophy, there is philosophy of biology, philosophy of physics, things like that.

3:49:05.0 SC: And there are also areas like mathematical logic or epistemology or metaphysics which interact and engage with science. But there could be a lot more. I think there's a lot more insight to be gained by really taking both seriously at the same time. And people don't do that. So anyway, that's my little, digression to say that read Modern Philosophers of Science read David Albert and David Wallace or Tim Molin or Jenann Ismael or Lori Paul who we had on the podcast before. These are people who are taking science very seriously and I think that one big thing in the decades to come is that these kinds of scholarly work will have a much bigger impact on the rest of philosophy. That's what I'm looking forward too.

3:49:56.6 SC: James Tyler Cann says, "I could use some advice. I have a master's in linguistics and I'm partway through a PhD in the same subject. However, I've been dissatisfied with the research currently being done in the field, the replication crisis, and I'm looking for something new. I'm considering doing a master's in computer science in the UK because it's a one year degree and would open up new possibilities. Do you think it would be unwise to leave my current PhD to go after a second master's in the hope of opening up my job prospects?" So I always address questions like this with a great deal of trepidation because advice is impossible to give for career choices like this because different people are different and I don't know everything about your situation. Having said that, although I do think that getting a PhD is intrinsically a useful thing to do, it's kind of training and research and scholarship and knowledge that is in principle, very useful. No matter what you end up to go and do.

3:50:51.2 SC: If you want to use that PhD to launch a career in that area, then you had better love what you do. I don't think, you know, getting a PhD is not something you reluctantly do because you can't think of anything else to do, or you don't have any other options. I think that you really need to be passionate about what you're diving into a PhD is a lot of your life. If it's just, even if it's just five years, you know, that's a significant fraction of your life and certainly a lot of work and angst along the way. So if you already know that you are sufficiently dissatisfied with the kind of research going on in this field that you are tempted by another career switch or change of focus in your scholarly interests, then my very rough inclination would be to say yes, go for it.

3:51:41.1 SC: Change your mind now. Do it sooner rather than later. You kind of gotta love what you do on the academic track and, and getting a PhD is part of that. So that would be my very rough said with great you know softness. Don't be too definitive about it, but that's roughly the direction which I would lean. Samuel Benjamin says, "Following your call for questions each month on Patreon, I often scroll through them as they come in liking or Harding the ones I find interesting. Do the number of likes or hearts on any given question ever influence your decision to answer them, even if it doesn't, would you encourage my continued use of this function and why?" Well, I will tell you the honest, sad truth. They have very, very little impact on my decision to answer them. And that's not because I don't care about them, it's just a down to earth technological feature.

3:52:32.3 SC: Namely that the way that I go about figuring out what questions to ask, et cetera, is I take the whole list of comments and I cut and paste it into a text file where I can both edit the questions down, 'cause some people like to ramble on when they're asking questions and I can, you know, rearrange them to put them in in an order that I like. And of course, I can delete the ones that I'm not going to get a chance to get around to. And when I do that cutting and pasting, the number of likes doesn't come with me. So I just don't know. I could put more work into figuring out which ones were highly liked, but I just don't, sorry about that. Now, there is one option, which is you can leave a comment on the comment, right? You can leave a comment on the question and you can say, oh, I really, really like this question that will generally get transferred in my cutting and pasting and I will see it.

3:53:24.9 SC: And there have been cases where I was gonna not answer a question, but then like two people said, oh yes, this is a great question, I hope he answers this one. And so I did answer it for that reason. So if you really care about a certain question being valuable, just leave a comment on the comment and say, yeah, I really hope that this question gets answered. I'm very interested in this also. Jeff B says, "I just watched a video that really helped me understand gauge fields. So I guess the idea is that what appears to us as forces is really a warping in some sort of space for gravity. It is literally spacetime, but for charge it could be warping in charge space. This is similar to forces we often think of as fictitious, like the [3:54:07.3] ____ force. Is there a profound difference between these two ideas or is it just easier for us to explain away the [3:54:14.8] ____ force as fictitious because it is so macroscopic?"

3:54:19.8 SC: You know, I think that there's something there, but I don't love the analogy between gage theory forces and fictitious forces. I think you've got the idea, your idea of warping in some sort of space is exactly right. But I don't personally think of those as fictitious forces. I mean, it's almost exactly like a ball rolling down a hill, right? If you have some undulating landscape of hills and valleys and put a ball on it and the ball rolls downward on the hill and then comes back up and decelerates, etcetera, I don't think of those as fictitious forces. I think of those as just forces, right? That the slope of the hill gives rise to a force. A fictitious force comes about because we're in some non-inertial reference frame, like we're in a rotating reference frame or something like that.

3:55:11.6 SC: But there is exists another reference frame, the inertial one where you don't see the force, okay, for these kinds of forces. There's no reference frame you can go into where they disappear. They are real, they really exist out there. So I do think that they have to do with curvature, with slopes, with derivatives of some background fields. I don't think that it's helpful to think of them as fictitious. Sandra Snooky says, "What kinds of physics will be featured in your upcoming book, physics of Democracy? I suspect that there will be a fair bit of statistical mechanics, but probably little to no quantum theory or relativity. What makes certain branches of physics useful in the social sciences and others not?" I think, you know, you're kind of answering your own question. I think you kind of know what the answer is. You just want me to say it out loud.

3:55:54.9 SC: So you're exactly right. There'll be a fair bit of statistical mechanics. There will be some network theories, some percolation theory, phase transitions, things like that, some emergence. All of these, these have in common that what they're doing is talking about the relationship of microscopic things to microscopic things. And that's why I think that there is something called the physics of democracy, because that is another example where you have microscopic things and macroscopic things. The microscopic things now are people and macroscopic things are countries or nations or governments or whatever. There's a very obvious parallelism there. And so I think that you can borrow ideas and inspirations from the physics analysis of these things which are inevitably simpler and in toy models, spherical cow kind of language, and maybe be inspired to get some insights onto the behavior of social systems. So there will be no quantum theory or relativity.

3:56:50.3 SC: Why not? Well, generally speaking, quantum theory and relativity aren't important for the studying the aggregate behavior of systems, which themselves are very classical and non-relativistic, right? Human beings are pretty darn classical. They might be probabilistic, they might, you might say that the best theory we have of an individual human behavior has some stochastic element to it, that's fine. But it's not quantum mechanical. It's not that there are super positions or interference or entanglement or anything like that in human beings. Likewise, you don't need relativity. People are moving slowly compared to the speed of light, right? Newtonian spacetime is more or less fine. So I think that there's a specific kind of physics that is very, very relevant to social situations, but relativity and quantum mechanics aren't gonna be part of it. There's other books I've written about relativity in quantum mechanics, so I don't feel like I'm depriving you of anything there.

3:57:45.4 SC: Sean Bentley says, "Often I read science articles about the usage of quantum entanglement for passing information. I can grasp how it could make information passing more secure, but wanted to make sure I understood correctly that there should be no way to use it for passing information faster than the speed of light. Is this true?" Yes, this is completely true. In fact, it is so true. I wouldn't personally state it as using quantum entanglement for passing information. The passing of information is generally completely classical. You have to send a signal from here to there, somehow. Things can be entangled and the existence of entanglement can imply correlations between what we see when we measure it. That's Bell's theorem or e p r or whatever. But it does not constitute passing information because if I am Alice and I'm measuring my spin here and my spin is entangled with Bob's spin I can measure it and get an outcome and therefore know what Bob will see.

3:58:40.4 SC: But Bob doesn't know what my measurement outcome is, so his predictions don't change when I make that measurement. So no information passes back and forth. So I would say that you're exactly right. In the broader super structure of information sharing entanglement is very important or could be very important, whether it's in encryption or keeping things safe from eavesdropping and stuff like that. Entanglement can serve a purpose, but in the actual passage of information from point A to point B, it doesn't really help. Okay, we have one question left from Pete Harlan who says, "I come over for dinner, you ask me if I'd like a cocktail, and I say, yes, I'd love your favorite martini. What do you bring me?" Well, I'm a big martini guy, so this is a great question, except that even though I am a big martini guy, I'm not that uptight about it, honestly.

3:59:36.8 SC: You know, I know there are people in cocktail land just as in many other aesthetic endeavors where they're gonna insist on the rightness of their preferences. But I am truly deep down a subjectivist about these things. I might make a certain kind of martini, but I'm actually happy to have different kinds of martinis. And if you like a different kind of martini, I will go to try to make you happy. My own personal martinis are gonna be not too dry. So there'll be gin and a non-negligible amount of vermouth. So well there. It certainly won't be vodka, let's put it that way, because gin is just more interesting than vodka. In fact, I have a little kit that helps me make my own gin not by having a a still or anything like that, but basically what gin is to some approximation is vodka, which you then flavor with botanicals, and you can do that yourself at home.

4:00:31.3 SC: The reason why people don't do it more often is because it generally colors your vodka when you turn it into gin with the juniper berries and the other kinds of botanicals. And the professional gin makers will then clarify the gin to make you look completely transparent. But at home you're not gonna be able to do that. But the flavors are important to me, so I don't see why anyone would ever want a vodka martini. This is going against my previous subjective inclinations, but there you go. I'm gonna be honest with you and not to dry, because I think that I don't wanna just drink cold gin. That's not a cocktail. I mean, that's a fine thing to do if you're in the mood for that. But that's not a martini, that's just gin.

[laughter]

4:01:09.3 SC: The wonderfulness about a cocktail is that you're blending different ingredients together and they balance each other and they play a role. So I would be, you know, four to one or five to one in gin versus vermouth, dry vermouth. So dryness by the way, if you're not a cocktail person out there, dryness has two slightly different but very related, connotations here. One is dry is the opposite of sweet in cocktail land. So there is sweet vermouth, which is generally the red vermouth, but there's also dry vermouth, which is not sweet, which is the clear vermouth in the green bottles. Okay? But then specifically in martini land, dryness just means how much vermouth do you have in there? So almost everyone who's gonna put vermouth in their martini at all is gonna use dry vermouth. But the dryer, it means less vermouth at all. So a dry martini has less vermouth than a not dry martini.

4:02:14.0 SC: And there's this kind of macho feeling that the less vermouth you have, the better because the gin is what it's all about. And I completely disagree with that, but there's a feeling out there. So when I order a martini in a restaurant or a bar, I need to specify not too dry. [laughter] You don't want it like half and half or whatever, but I don't want to be unable to taste the vermouth. And also I don't want like a dirty martini with olive juice or anything like that. You want the garnishes to be there and to be noticeable, but not to overpower. So couple of olives, I'm very happy with, you know, garlic stuffed or jalapeno stuffed olives or just a twist. Twist sort of looks better aesthetically, but doesn't give as much taste to it. So it depends on what I'm in the mood for.

4:02:57.4 SC: Like I said, St. George's terroir gin is my current favorite, but I'll be honest that I haven't really tried to explore every single kind of gin out there in the world. There is a fun gin called Indigo that is colored blue. [laughter] If you wanna make a blue colored martini, it just tastes like gin. It doesn't taste blue. I don't know what that would mean, does tastes like blueberries, but you can make a blue martini, which in that case I would strongly recommend going for the twist to decorate it rather than the olives. The green olives just don't look good in the blue martini, but the yellow twist of lemon peel looks great. So there you go. It depends on what I have on hand, what mood I'm in and what do you want. And I think that's the way our aesthetic choices should be handled. So with that, thanks as always for supporting Mindscape. I appreciate it very much. Thanks for sitting around for this long set of AMA questions and I doubt that any of the climate change denialists have stuck around this long. But if you have, I hope that you are at least mildly entertained. Thanks, I'll talk to you next month.

[music]