AMA | September 2026

0:00:01.4 Sean Carroll: Hello, everyone. Welcome to the September 2026 Ask Me Anything edition of the Mindscape Podcast. I'm your host, Sean Carroll, and we should start today's AMA with corrections. [chuckle] I should do like a corrections column, just like the New York Times has to do, or something like that, right? Usually I try not to have many mistakes that need correcting, but had a couple of whoppers over the past month. I'm not gonna blame advancing age, but perhaps the onset of the new semester, we're now teaching again, it's the new semester starting, has distracted me from getting things correct. So, one obvious blunder was in the episode with Vitor Cardoso, I introduced him and put his institute, the Center for Gravity, in Copenhagen, which I then mentioned was in the Netherlands, which is not true. That is not where Copenhagen is. Copenhagen is in Denmark. Amsterdam is in the Netherlands. I've been to both cities, they're both very nice, but they're also different. They're not the same. And so that was just a little brain fart there. I did manage to catch it, but only after it went live, so I tried to replace it. And what happens is, once you publish the podcast episode, it gets distributed and put into people's caches and memory, and it takes a long time to update it. So I could update the YouTube video right away... I say video, but of course it's just audio. But it takes a long time for the actual audio regular podcast to promulgate to all the right places. So, sorry about that to Vitor, to the people of Copenhagen, and the people of the Netherlands and Denmark. Sorry about that.

0:01:42.0 SC: The other one is more interesting and a more subtle mistake on my part, but still a pretty egregious blunder. I'm supposed to know about these things. It was based on the AMA question from Michael Bright last month. And people did catch this also. I will mention Ashley Driver and Jeffrey Siegel, both of whom asked AMA questions pointing out my goof. They're too polite to just say, "You goofed," but I did. So here is the setup, and I'll just try to explain it and get it right. You may have heard of the double-slit experiment in quantum mechanics. It says if I take a source of electrons and I fire the electrons at a barrier with two openings, two vertical slits as openings, and I have a detector screen on the other side, what will I see? If I fire the electrons one at a time, according to quantum mechanics, the electrons are described by wave functions as long as we're not observing them. And therefore, if the electron goes through both slits, on the other side, it will emerge from both slits as a wave... As two different waves that can then combine and either reinforce each other or cancel each other out. And therefore, what we see at the detector, for every electron, it's just a dot, it's just where you detected the electron signal. But if you accumulate many, many dots by doing this over and over again, you see an interference pattern. You see bands where there's lots of electron dots, bands where there's almost none. And that's interpreted as saying that there was a wave going through the two slits. This is evidence of the wave-like behavior of quantum matter.

0:03:16.8 SC: And then the spin is, or the twist, I should say, if you try to look for which slit the electron went through, the left one or the right one, that means you've made a detection, you've made a measurement, and the wave function collapses in the usual way of talking about it. And once that happens, the electron only goes through one slit. It goes through the one you saw it go through. And therefore, coming out the other side, it is a wave once again. You're done detecting it because you didn't stop it, you just sort of noticed which slit it went through. But it's only going through one slit, so there's nothing to interfere with, there's now only one source. And what you see is a pattern on the detector screen which is just a single-slit pattern, a single blob if you want. No interference bands, no oscillating, yes, no sort of bands of electron dots. And so that's one very famous thought experiment in quantum mechanics, which was first a thought experiment, now eventually it's been done. The other famous thought experiment that Michael's question combined with it is the EPR thought experiment. Einstein, Podolsky, and Rosen, they say construct two particles that are entangled with each other, let them go very, very far away, and measure one, and instantly the state of the other one is collapsed to whatever it needs to be to be compatible with the entanglement. You can't send signals using entangled particles because the person who would detect the second particle doesn't know that the first one has been detected and observed to be in a certain state. And therefore, as far as they're concerned, the probability of getting whatever measurement outcome they might get is just whatever it was before. So no signal has been sent.

0:05:02.8 SC: So Michael says, let's combine both of these. Let's imagine an EPR-like experiment where you have two particles that are entangled and moving in opposite directions. One goes toward Alice, one goes toward Bob. But rather than just detecting them, Alice and Bob have a double-slit experiment, each of them do, right? And they would both go through and maybe make interference patterns. And the question was, if Alice, at her double-slit experiment, starts detecting which slit it goes through, then does the wave function for Bob's collapse, and now you know which slit it's gonna go through, and therefore the interference pattern disappears? And my mistake was... Well, so what I said was the interference pattern will disappear. And I gave a perfectly sensible argument for why that's true, because it will. But the problem is it was never there, that's what I got wrong. There never was an interference pattern in that case. And the reason why is because the particle sort of already entangled with the rest of the world by the setup, right? When you do the original version of the double-slit experiment and the particle just goes through two slits and you "detect" which it goes through, what that means is that you've entangled that particle with the rest of the world, in this case, your detector for the location of the electron. As long as that entanglement is there, the electron is no longer in a pure state all by itself. It's connected to the rest of the world. And that's what destroys the interference pattern on the other side.

0:06:38.4 SC: Here, I was so focused on what happens if you start measuring what happens when the particle goes through the slit that I didn't stop to think, what if you don't measure? What if you just let it go through? Here, the particle's already entangled with the rest of the world by the setup of the question. And therefore, what it emerges on the other side of the slits is not in a pure state that can interfere. So you already don't get an interference pattern. You've basically pre-detected, pre-measured, even if you don't know the outcome of the measurement. By putting the electron or the particle moving into an entangled state, you have implicitly measured which slit it goes through, and therefore you've destroyed the interference pattern. And that's good news. This was pointed out to me by an email from a physicist friend, [0:07:28.7] ____ Frondsman, but also AMA questioners this month, Ashley and Jeffrey. So, it's good news because if you could do this, if you could actually have an interference pattern and turn it on and off just by detecting particles at one of the detectors and therefore have the interference pattern go on and off, appear and disappear on the other one, you could send signals faster than the speed of light. And that is bad. And there are theorems that say you can't do that. So, maybe... You should always be careful about these things, right? Maybe your theorems are not right, or maybe your theorems are assuming something that is not applicable to this situation. In this case, the theorems are fine and the implications of the theorems are completely correct. You should not be able to send signals faster than the speed of light, and therefore this wouldn't work. So I apologize to Michael for bungling his question. The answer is you're never gonna see any interference patterns either way, whether or not you personally start looking for which slit the electron goes through in that particular setup.

0:08:36.8 SC: All right, well, here we are, a whole new month of AMA questions. More opportunities for me to make mistakes and bungle, but I do promise that if the mistake is straightforwardly factual, I will try to let you know about it and try to fix it. If it's a mistake of judgment or I like a different TV show than you do, then I'm not gonna fix that. You just have to live with those kind of more subjective kinds of disagreements. That's okay. Thanks as always to the Patreon supporters of Mindscape. They're the ones who make this AMA possible and they're the ones who ask the questions. You can be a Patreon supporter of Mindscape. Just go to patreon.com/seanmcarroll. For a few bucks a month... I still need to switch the payment schedule, sorry about that. It's gonna happen soon. Big disruption, sorry about that. It won't really be a disruption for anybody. And then the benefits of being a Patreon supporter are that you get to ask the questions for the AMAs, and also you get ad-free versions of the podcast, as well as just the feeling that you're supporting something that you like and giving some money back. So many, many thanks to everyone who does that and has been doing that for years now. It absolutely keeps me going. So with that, let us all go.

[music]

0:10:08.1 SC: Eric says, "Would an Alcubierre drive be able to escape a black hole?"

0:10:14.7 SC: Okay, so this is a sneakily good question. I don't mean sneaky in the sense that we shouldn't expect it to be good, but at first when I read it, I was almost tempted to not include it among the ones that I'm gonna answer just because I didn't think I had much interesting to say about it. But the more I thought about it, there is something interesting to say, even if the interesting thing to say is that I'm not quite sure what the answer is. Let me go back for those of you who don't know what we're talking about here. Back in the 1990s, I think it was, Miguel Alcubierre, who is a Mexican physicist, proposed what is known as a warp drive. So the idea is... Back in the '90s, we were arguing about wormholes. People had thought about wormholes, and Kip Thorne and his collaborators had pointed out they could be used for time travel. They're sort of a shortcut using the curvature of spacetime in general relativity to get from one place to another faster than you might think. Now, the interesting thing is, for the wormhole idea, you're never traveling faster than the speed of light, right? What you're doing is you are making use of the flexibility of spacetime in general relativity to sort of trick yourself into thinking that the speed of light is different than it really was. The speed of light is always the speed of light. It's one light-year per year. It can never change from that. But if spacetime itself can curve in different ways, the number of miles per hour you gotta go might be different than you think to get from one place to another.

0:11:48.1 SC: So in other words, if I know that Alpha Centauri is four light-years away, I think naively it's gotta take me at least four years to get there by a clock that is sitting back here on Earth. But a wormhole provides a shortcut. It's not going faster than the speed of light, but it's changing the distance to Alpha Centauri because I can go through the wormhole and it's like magically appearing much faster than I might have thought. So it looks to the person on the outside, like you're moving faster than the speed of light. That's just because they don't understand or they haven't thought about the actual path that you took. So Alcubierre realized, this is all great, but you don't need a wormhole to do it. You don't need to change the topology of spacetime. All you need to do is change the curvature of spacetime. So, he wrote down and published a paper explaining that if you can imagine a spherical bubble where the metric, the tensor field that tells you how spacetime is curved, takes on a very, very particular form, and this very, very particular form lets that little bubble move through the ambient spacetime in a way that, again, if you didn't know any better, you would think was moving faster than the speed of light. What it's really doing is shrinking the distance that you're actually traveling. So to someone outside the warp drive bubble, it looks like you're moving faster than light, but from inside, you're actually barely moving at all. It's the spacetime around you that is moving and you're just kind of following it.

0:13:17.8 SC: This is not trivial to do. In fact, it's probably never gonna be done, the actual construction of such a warp drive, for reasons that people have debated, mostly using the fact that it involves negative energies or exotic matter, whatever you wanna call it. Something that people don't talk about as much is that it requires an enormous amount of energy. Even... You know, who knows how to control negative energies? If you could actually do that, we think that in quantum field theory, you can sort of get a little tiny flicker of negative energy occasionally, but we don't have any controlled way of conjuring up negative energies on the regular and putting them to engineering uses. And anything... Just because gravity is a very weak force. Remember, the whole Earth, all of Earth, which is a lot of matter, exerts a really tiny gravitational force on you, right? You're able to stand up and walk around using your own muscles, even though the entire Earth is pulling you down. So the amount of energy you need to warp spacetime in any noticeable way to make something like a warp drive is absolutely crazily enormous. So this is just not something that is even conceivable in terms of technology as we can currently think about it.

0:14:37.6 SC: But okay, nevertheless, let's put all those things aside and answer Eric's question. "Could you use it to escape a black hole?" So my initial feeling was no, you can't do that because, again, it only looks like you're traveling faster than the speed of light. You're not actually traveling faster than the speed of light, and to escape a black hole, you need to actually travel faster than the speed of light. But here's why the question is a little more subtle than you might think. So just to fill in a little bit what I just said, a black hole is a region of spacetime with a horizon around it. A horizon is basically a barrier... It's not really a barrier, you can cross it one way but not the other. It's a boundary, that's the word I'm looking for, a boundary that says that if I'm on one side of the boundary, I can't get to the other side without moving faster than the speed of light. If you're inside the black hole, you can't escape to the outside without moving faster than light. And so you might very naturally ask, could an Alcubierre warp drive bubble help you change that picture? And here's the reason why it's a hard question to answer, because what Alcubierre did was write down a metric. And this is something, I guess I have to explain how general relativity works here. Einstein's equation of general relativity, I presume you've all read volume two of The biggest Ideas in... Sorry, volume one of 'The Biggest Ideas in the Universe: Space, Time, and Motion,' so you're all familiar with Einstein's equation, right? But I can remind you, maybe if you've forgotten, there's a left-hand side to Einstein's equation which says, "Here is what the curvature of spacetime is doing."

0:16:16.1 SC: And then there's the right-hand side of Einstein's equation, which says, "Here is the source of matter and energy that is making spacetime do that." So if you don't have any restrictions at all on what the sources of matter and energy are on the right-hand side, every possible spacetime is a solution to Einstein's equation. You just give me the spacetime you want, I calculate the left-hand side of Einstein's equation based on that, right? And then I posit that there maybe, should be, could be a right-hand side that makes the equation solved, right? In other words, there is always some distribution of matter and energy in the universe that could give you any spacetime curvature at all, unless you start putting restrictions on what kinds of matter and energy are allowed. And that's what is generally done in general relativity textbooks or proofs of theorems like the Penrose-Hawking singularity theorems. They need to assume some properties that energy-momentum are not allowed to have in order to prove anything at all about possible solutions to Einstein's equation. So Alcubierre didn't do that. He just helped himself to whatever he wanted. And so all Alcubierre did was say, "Here's the metric I want. I can tell you what the energy would have to look like. It's gonna involve negative numbers and things like that. I have no idea how to get it, but if it could be gotten, then we could make this warp drive." And it's a little bit even more sketchy than that, what is actually going on, because it's a solution to Einstein's equation for the curvature of spacetime. It's kind of an all-at-once solution.

0:18:04.5 SC: Very often in physics, what we'll do is look for what's called an initial value problem. Like I give you, "Here's the baseball, here's its location, here's its velocity, tell me how it travels on a parabola-like trajectory through the air before I can catch it." Okay? That's a very typical setup. This is not what's very often done in general relativity. In general relativity, often, not always, but often, we just sort of give you a solution that is four-dimensional from the start, and that's what Alcubierre does. So you might ask, given the setup of the Alcubierre warp drive, is there a way to sort of get the behavior you want from a well-defined initial value formulation? That is to say, rather than just saying, "Here's the four-dimensional spacetime," you say, "Here is a configuration of matter and energy at one moment of time, and I'm gonna give you some dynamical equations that tell you how that matter and energy evolve, and then I'm gonna solve all the equations for both matter and energy and spacetime in the future and see that I get a bubble of warp drive moving purportedly faster than the speed of light from the external point of view." So that is not done. We don't know how to do that. As far as I know, no one has really shown how to do that with sensible forms of matter and energy for the Alcubierre warp drive. So to really answer this question, "Could you escape from a black hole?" You would need to set up those initial conditions. And a black hole is a weird thing because the definition of a black hole involves not just the moment of time, but the future. It's a statement about the future. It's a statement that if I'm in this region of spacetime, at no point in the future can I escape to the outside, right?

0:19:53.7 SC: So setting it up as an initial value problem is a difficult thing to do. Usually, again, in general relativity, we'll start with some well-defined simplifications, things like that, and we'll get a useful thing. This is why numerical relativity is hard, because solving these initial value problems in general relativity is very difficult. But the upshot of all that long discussion I just gave is that I don't know whether you could imagine in a well-formulated context making the equivalent of an Alcubierre warp drive and escaping a black hole. I strongly suspect the answer is no. But I actually, if I'm honest and trying not to make mistakes anymore, I have to say I don't think that we understand, and certainly I don't understand the initial value setup in a way that would allow me to actually answer that question. There you go. Sorry for a slightly unsatisfying answer.

0:20:44.5 SC: JJP Master asks a priority question. So every Patreon supporter gets once in their life to ask a priority question, and I will do my best to answer it. Typically, I don't get to answer every question that's asked, maybe half of the questions that are asked, just because there are limits on how much talking I can do. So the priority question forces me to try to answer it. I mean, there's always a philosophical question about how one should deploy one's priority questions. Because if you ask a question and I don't answer it, and you're welcome to ask it again and again and again, but if I keep not answering it, you should get the message that in my judgment, my answer would not be that interesting, either because I don't know something about it, or I have nothing interesting to say, or whatever. So if you do ask a question and I don't answer it, there's two options. One is you can just think to yourself, "Well, probably there was not anything interesting to say." Or the other option is you can start asking a priority question if you think that's so important to you that you want me to try. I'm happy to do that.

0:21:48.0 SC: So here's this one, "In your last AMA, you expressed sympathies for full non-indexical conditioning when doing anthropic reasoning. However, that approach seems, by my lights, to undercut the force of your appeals to a cosmological multiverse to explain anthropic coincidences or fine-tuning. Compare two multiverse models. One predicts that every pocket universe without fail will be life-permitting, while the other predicts that only some such universes exist."

0:22:16.3 SC: So, by the way, I'm pausing here. I presume that this kind of use of the word "universe" is really like "region of space," and that's how cosmologists typically use it. So we're not talking about the many-worlds of quantum mechanics or something like that. Like, a big old pocket universe could be billions and billions and quadrillions of light-years across, but it's still tiny compared to a very, very, very big multiverse out there.

0:22:41.9 SC: Okay, so JJP Master's question continues, "It seems like fully non-indexical conditioning would strongly favor the first model, the one where every universe is hospitable to life, as we cannot appeal to a selection effect to level the playing field. However, the first model would simply push back the anthropic question, why are life-permitting laws or constants necessary? Rather than resolving it, as does the second model. Would you agree with this assessment, or do you think there is something that permits the relevant kind of anthropic reasoning used to argue for a multiverse from cosmic fine-tuning, given an FNC approach?"

0:23:20.0 SC: So for those of you who don't remember when I did talk about this, fully non-indexical conditioning is the idea that you shouldn't close your eyes and pretend you're a typical observer in the universe. You know you're not a typical observer. You know all sorts of things about you that make you not typical, and there's no reason not to use those things, not to be allowed to reason on the basis of them. So fully non-indexical conditioning is an idea... The label comes from Radford Neal. Philosophers have discussed similar things. Philosophers and scientists never talk to each other about these things. And the idea is just, except for indexical facts, which is to say, if there are multiple copies of me, I don't know which one I am, okay, that's a true uncertainty, you're allowed to think yourself typical within the set of all people who look exactly like you. Okay? That's fine. But there's no reason to forget everything else you know about the universe and think you're typical within the set of all observers. And so the question is saying, if you do that way, how can you recover traditional anthropic reasoning? Wouldn't you favor... If I understand the question, I'm not completely sure, but if you favor a universe where every place is hospitable to life, isn't it just more likely, I'm guessing this is the idea, it's just more likely to exist in that universe than one where some regions of space allow for life but not others?

0:24:48.9 SC: So there's two things going on here, two parts of the answer here. One is there's a very standard... Standard in the sense that everyone agrees on it, part of Bayesian reasoning, which is you need a prior on these different theories. So if you had two theories, one of which predicted many, many, many regions of space, all of which were hospitable to life, and another one which only predicted a tiny fraction are hospitable to life, and your priors on both of them were equal, they seemed equally reasonable in terms of the physical mechanism that could create these kinds of universes, et cetera, then I'm perfectly happy preferring the one that has more living beings in it, right? That just is saying that that universe is more likely to support an observer like me. I don't actually have any problem with that. I think that the general reason why you appeal to anthropic reasoning is because in order to have a theory where an observer like you would definitely exist without a multiverse, or equivalently, if there were many regions but they all allowed for the existence of life, those theories seem finely tuned and have low priors, right? So I can just invent a theory where the cosmological constant is small and go boom, I'm done. I've invented a theory. The problem is if I didn't know the value of the cosmological constant, I would have put a small prior on those theories. And if I have a multiverse with different values of the cosmological constant everywhere, I have no reason to put a small prior on that. There's no fine-tuning there. It's just random stuff happening and still allows for the existence of life. So you have to take the priors into consideration when you do this sort of comparison.

0:26:32.1 SC: The other thing is there is something which I'm not sure... I don't think it is in Neal's original paper, but it is in the paper that Isaac Wilkins and I are working on ourselves, which is sort of a saturation effect where we don't update your priors by how many observers are like you. We update our priors by the probability there would be at least one observer like you in this universe. So if your universe is relatively tiny, that does get penalized compared to a universe which is very big and has more places for observers to appear. That is something that we accept, and we actually think that that is a perfectly reasonable way to go. There's something called the presumptuous philosopher problem in anthropic reasoning, which the traditional way of doing it without saturation, which just puts the prior proportional to the number of observers like you, that has the huge problem that there's an infinite preference for infinitely big universes, right? Even if the universe is big enough but finite, big enough that there should be many, many observers like you, that's nothing compared to the infinite number of observers like you in an infinitely big universe. So you just say, without looking out the window, you go, "All right, I know I'm not in the finite-sized universe. I have to be in the infinite-sized one." And that seems a little bit presumptuous, right? So we don't want that to happen. But I do think that if there's two universes with different non-zero but non-one probabilities of getting at least one observer like you, that it's okay to prefer the ones that have a slightly bigger probability of getting an observer like you. I don't think that that's too presumptuous. I think that's just doing good science.

0:28:21.3 SC: Ken Wolf says, "In your talk with Luis Bettencourt, you discussed how the size of the city that a person lives in can change their behavior right down to things like how fast they walk. I could not help but think of a previous episode where Mark Berman made a distinction between directed or top-down attention and involuntary or bottom-up attention. Could it be that a feature of larger cities is that there are simply more things which are forcing us to direct our attention, leaving less time and opportunity for involuntary attention? Could that be a source of these changes in behavior?"

0:28:51.3 SC: It could be. That is possible. As you know, that's gonna be my answer. Mark Berman and Luis Bettencourt are colleagues at the University of Chicago. They collaborate together and they talk to each other, so I'm sure they have thought about this. My impression is that the people like Luis Bettencourt and Geoffrey West, who is his collaborator also, who thought about the city scaling laws and the superlinear and sublinear behavior, et cetera, are less concerned with the mechanism and more concerned with just figuring out that it happens. I think... Or at least their kinds of mechanisms are not down at the psychological level. They're at the network level of people bumping into each other. I think they would say you walk faster in a city with a large population because the entire rate of all processes is faster. There are more people, there are more interactions, more things happening. It just sort of speeds you up in your day, including just walking down the street. You gotta get somewhere, you gotta get someplace to do. So I don't think that they bring attention into it in that direct way, but this is why the social sciences are both interesting and challenging. Maybe those attention aspects also have something to do with it, and maybe that's a competing explanation. And then you have to figure out how to test one versus the other. Can you do some random controlled trials where you figure out, aha, it's because of people's attention patterns rather than people just rushing to do the next interesting thing. I truly don't know, but that'd be interesting to figure out.

0:30:26.5 SC: Erkon Sertelli says, "Looking at the history of physics, it seems that often mathematical concepts invented decades or centuries earlier turn out to be crucial for describing new physical theories. How do theoretical physicists approach mathematical concepts that do not yet have any real-world use case? As in, is it common for physicists to keep up with newer, more abstract maths in case they prove useful for giving inspiration or offering new perspectives on existing problems? Or are they mostly ignored until a particular problem creates a need for that kind of mathematical framework?"

0:30:58.0 SC: Well, you know, there's different kinds of physicists out there, right? And they work in different ways. And I think that most physicists don't try very hard to follow developments in new mathematics. Most physicists are interested in physics. And this is a hobbyhorse. This is something I say over and over again because everyone loves to emphasize the importance of mathematics in physics, both sort of historically or philosophically, but also just pedagogically when you're learning physics and you're trying to become a good physicist. It is perfectly clear that there's a lot of mathematics that is very, very useful in becoming a good physicist. But people get a little carried away with that, honestly, and they start thinking that physicists should follow mathematics because it might be useful or whatever. And there are some people who do. There definitely is a sort of subclass of physicists who love to play around and learn new mathematics and then put it to work, see if they can find a place where that new mathematics could be useful for something. But I think that most physicists have the math they need and they're interested in solving physics problems, right? Most physicists are not string theorists or highly mathematical people. They're trying to figure out what makes superconductivity or why sand moves in a certain way or how a magnetic field pushes around a plasma. And these are all things where we know the math, we have to figure out the physics. So there's a small fraction who do that. But for the most part, physicists will wait for the need for new math before actually diving on that bandwagon.

0:32:32.8 SC: Killian says, "If you were innocent of a serious crime and the evidence needed to establish the truth was available, would you rather 12 human jurors or 12 AIs decide your fate?"

0:32:47.4 SC: So I think right now, if this was 2026 we're talking about, 12 human jurors, no problem. I would not even hesitate to pick the human jurors. And this is because... I mean, and I say that 2026 because maybe it will get different. Maybe the AIs will get better at it. But this is again, my own personal experience using LLMs, which fits in 100% well with how they're trained and what they're doing. LLMs are really, really good at conventional things, at things where people have talked about them a lot. This is why LLMs are especially good at computer programming, because a lot of computer programming is doing a task that someone else has already figured out how to do, right? And it's in the LLM's training data, and the LLM just has to apply it for you. But from my own use of LLMs, which is often in areas of theoretical physics that are not well established, right? I'm trying to figure out new things, that's where the LLMs are not any good, right? If they don't have data in their training set that is easily interpolatable into the relevant conditions that they're asked about, then they can stumble and make things up, and they can often act very confident in their wrong answers. Okay? So I worry that if I were innocent of a serious crime, that the LLMs would just look at other people who were accused of such serious crimes, figure out that usually they were guilty, and find me guilty. It's not quite that simple, but you know what I mean, they would fall into a pattern. Whereas I have more trust in the human beings to be able to look into the specialness of the particular situation, which is what is called for when you're on a jury. So as of right now, give me the human beings any day.

0:34:37.6 SC: Adam Mills says, "If gravity is a wave, can information be passed through it similar to other waves? Could we theoretically manipulate gravity waves the same way we manipulate radio waves? If that's possible and we had the technology to do so, what applications could we choose to use gravity waves for as opposed to using another wave?"

0:34:56.5 SC: So gravity waves are waves just like electromagnetic waves. Information can be passed through them just like other waves. It will never be a useful, plausible way to convey information, basically because... Two reasons, actually, I gotta get two reasons because they're both sort of equally important here. The most obvious one is that gravity is super duper weak. Okay? When we detect signals at LIGO, the Laser Interferometer Gravitational-Wave Observatory, it's true that they're a billion light years away, so they're very, very far away. But they're created by two black holes, each 30 times the mass of the sun, spiraling into each other. Okay? That is not something you're gonna build in a radio. You're not gonna have two 30-solar-mass black holes in your radio or even one solar-mass black hole in your radio. It's very, very hard to generate even a little tiny gravitational wave by moving around ordinary amounts of matter. You could do the calculation if you wanted to, go look up in a book, look up the quadrupole formula and ask, if I have a dumbbell and I move it back and forth really, really quickly, how much gravitational waves do I create and how easy would it be for them to be detected? Very, very, very, very, very hard. Remember, LIGO, the amount by which the gravitational waves that they detect actually jiggles the mirror is much less than the size of a single proton. They're only able to measure the accumulated effects. They have many, many photons in the lasers and they look for very, very tiny changes in the location of the crest of the laser.

0:36:40.1 SC: So it's not... Again, and that's a four-kilometer-long tube. You're not gonna carry around a four-kilometer-long tube in your pocket either. So just as a practical reason, as a practical matter, it just doesn't work. But there's another, the other aspect, which again is equally important, is that for electromagnetism there are positive and negative charges. For gravity, there's essentially only positive charges, because mass is, or energy, if you want, is what generates gravitational fields. And mass is positive. You don't get negative mass particles generally, modulo what we just talked about, they can flicker into existence momentarily, but you don't have a negative mass particle just sitting there in front of you. And what that means is that you can't cancel out the gravitational field and manipulate it in very, very fine ways like you can an electromagnetic field. A radio or the computer in front of me is this intricate dance of all sorts of positively charged and negatively charged particles constantly manipulating electromagnetic fields by canceling them out, generating them, canceling them out again and all this stuff. So that's why you can do so many interesting things with electromagnetism, because you can manipulate it in sort of pinpoint ways. Gravity is dumber than that. All you can do is make it, right? If you have two 1-kilogram masses and you bring them together, you have a 2-kilogram mass. That's it. There's nothing else you can do. You can't manipulate the inside of your masses to make it seem like it's less than 2 kilograms or anything like that. So for all sorts of reasons, gravity waves are just much, much less useful than electromagnetic waves.

0:38:20.1 SC: Peter 42 says, "String theory says reality is fundamentally made of tiny strings. How do these strings connect to each other? Are they bunched up like spaghetti or do they sit neatly in a crystal lattice? Is there any truly empty space between them?"

0:38:33.7 SC: So I'm gonna answer this question... I'm gonna address this question, I should say, because I can't answer it very well, and I'm kind of tempted to say because no one can answer it very well, but I say that with 100% awareness that maybe somebody can and I just don't know. So string theory is an interesting thing. Remember, it was not invented to be a theory of gravity, right? It was not invented by people going like, "How does spacetime work?" They were actually just trying to understand features of strongly interacting particles, a relationship between the masses and the spins of various strongly interacting particles. They said, "Well, maybe if the particle is modeled as like a little segment of string that is itself spinning, I can do this." And only later did they realize that in a completely consistent version of string theory, you had particles, you had strings that acted like gravitons, that acted like massless spin-2 particles that coupled to the energy and the momentum of other particles in the theory. So string theory predicts the existence of gravity, okay? But it predicts it in a weird way. And now you have to sort of... String theory has advanced quite a bit since the '70s or '80s where it was first talked about, but the initial formulations of string theory were strings propagating through spacetime. And you didn't construct the curvature of spacetime out of the strings. You proved that in order for the strings to consistently propagate through spacetime, that spacetime has to be curved in a certain way, and that certain way is predicted to be Einstein's equation for general relativity. So the spacetime through which strings travel has to obey the rules of general relativity. But spacetime is there as an ingredient, right? So it's not that... So when you say truly empty space between the strings, what does that mean? I don't know what that means. It's empty space, but there's still a curvature, there's still a metric, there's still some geometry.

0:40:35.3 SC: But also this language of strings moving through curved spacetime is kind of old-fashioned, okay? It is what is called perturbative string theory, where you just take strings and you let them bump into each other and you create the equivalent of Feynman diagrams and you look at them that way. But it was never the whole story. They always knew it would never be the whole story. You need something more profound. This is why people like Joe Polchinski could write papers with titles in the 1990s like "What is String Theory?" Even though string theory was invented in the '60s and early '70s, by the '90s they're still asking what is it? And still now they're asking what is it? We have explicit examples where you have things like the AdS/CFT correspondence, where there's a duality between two theories. One of those theories is string theory. String theory in ten-dimensional anti-de Sitter space, compactified so that it's really five-dimensional anti-de Sitter space times a five-dimensional sphere. So that's string theory. And the theory that it is dual to, so completely equivalent to, is just an ordinary non-gravitational quantum field theory in four dimensions, four spacetime dimensions. So that theory has no strings at all. Okay? So this is a version of string theory without any strings in it, on one way of talking about it. And they're related by holography and it's non-perturbative, and it's all very complicated. So I think that "we don't know" is the short answer to here what the strings actually do.

0:42:10.6 SC: The right way to think about string theory is it wasn't discovered... And this is not at all unusual in the history of physics, it wasn't put forward as a complete, well-formulated thing, right? It was put forward as like a set of rules that work in a certain regime and you're trying to figure out the full theory. This is exactly how quantum mechanics was invented. If you go back to like Planck and Einstein and Bohr and de Broglie, they were all making up these ad hoc rules when they invented quantum mechanics. And it wasn't until Heisenberg and Schrödinger in the mid-1920s, 25 years later, that we were able to have a fully blown theory of quantum mechanics. And even then, of course, you all know it's really not a fully blown theory because the measurement problem is still sticking around. So we don't exactly know what the right way to think about string theory at the deepest possible level is. That's why we gotta keep thinking about it.

0:43:06.8 SC: Emmett Francis says, "I've been fortunate enough to receive Sean Carroll advice here at a couple previous critical junctures in my academic career, thesis writing, starting my postdoc. And things seem to have worked out so far." [chuckle] I'm glad they have. Thank you, Emmett. I think it's much more your fault than mine, I'm quite sure. "Now I'm approaching my first cycle applying for faculty positions and I'm curious if you have any pointers you would be willing to share. For context, I'm mainly applying to engineering departments in the US. My background is bioengineering, but any thoughts on the faculty application process in general are much appreciated."

0:43:42.8 SC: Yeah, there's this standard problem that we train people at every stage to do a certain thing, but that's not what we want them to do at the next stage where their training is supposed to come into play. So, of course, when you ask for advice about applying for faculty jobs, there's really two kinds of advice. One is getting a faculty job and the other is being a good faculty member if you get the faculty job. And those are two different things. For the getting of the faculty job, I think most of your work is already done one way or the other. If you're already applying, you have a CV, you have publications and whatever. The departments will hire on... I'm hesitating myself here because it's not 100% true. What I want to say is they'll hire on hope. They will hire on the hope that you grow into the best version of yourself, right? And in the back of their minds, whether they say it or not, they're thinking, "Well, if they don't, we'll just not give them tenure and we'll be rid of them," right? So they're willing to roll the dice a little bit. But what that means is you have to sort of make it clear what that best version of yourself is. How big can you be? How great can your research be? What is the highest achieving version of the scientist you want to become? And you want to make that clear. That includes the research you do, but also, are you enough of a self-starter to take initiative, to apply for grants, to set up a lab if that's relevant to what you want to do, to advise graduate students, to start new courses? All of these things that you weren't expected to do as a postdoc, now you're gonna be expected to do them. And so giving the impression that you'll be good at doing all those things as well as keeping up your research at whatever level you can. Of course, the most important thing is to have really, really good research, and I can't help you with that. By now you do or you don't.

0:45:40.4 SC: In terms of being a good faculty member, there's again 20 different things that go into it. This is the weird thing about being a faculty member. Everyone knows that it's really hard, it's very busy, you have a lot of demands on your time. It's still surprising how busy you are, how many things... It's not any one thing. It's not like, "Oh, there's this new responsibility you have." No, there's a hundred new responsibilities you have. They're all individually very small, but they can overwhelm you if you let them. So you gotta put a lot of effort into prioritizing, keeping your eyes on the prize, making sure that you're carving out time for whatever it is is most important to you while juggling a million different balls in the air and things like that. But look, it's fun. If you actually are at the stage where you get invited for an interview or something like that, consistent with what I said earlier, I would say this is not the time to be self-effacing. This is not the time to be humble, right? Try to explain to them why you are the best person for them to hire. Don't be shy about doing that. You don't want to be arrogant either. You don't want to say other people are idiots. You certainly don't want to say the people in the department looking at you are idiots. You want to act very, very interested in their work, but you also want to act like you are the right person for them to hire, which means that they can't do better than hiring you, right? And you have to walk that line as carefully as you can. So good luck with that, Emmett. We're all rooting for you here at the Mindscape podcast.

0:47:09.6 SC: Nikola Ivanov says, "In your effort to recover the classical world from the universal quantum state and the Hamiltonian, which features of classical structure emerge and which are assumed in choosing how Hilbert space is divided?"

0:47:23.1 SC: So this is referring to the ongoing project we have to just conceptualize quantum mechanics from the start as a bare-bones theory of a quantum state vector moving through Hilbert space according to some Hamiltonian. And the idea is from that bare-bones description, things like space and particles and location and locality, these all emerge from that bare-bones description. And that's it. So the answer is no features of classical structure are assumed in choosing how Hilbert space is divided. You're supposed to imagine going through all the different ways to divide up Hilbert space so that you can recognize them as, "Ah, this is giving me some emergent classical description." Now, I think that most Hamiltonians moving in Hilbert space in different ways don't give rise to any classical description at all. So there's some assumption about the appropriateness of the form of the Hamiltonian. There has to be the possibility of a classical world emerging before you go looking for what classical world does emerge. And we're still trying to figure out what is the best way of pinpointing what that criterion actually says. But that's okay. So we're trying to assume as little as possible. That is what we're trying to do, and we'll see how it goes.

0:48:41.8 SC: Roland Weber says, "I recently learned that Emily Wilson, former Mindscape guest, is working on another translation of the Odyssey. More syllables per line, relaxed line limit, making it bigger and better than her first. She's already more than halfway through. When she succeeds in this endeavor, will it be epic enough to bring her on as a repeat guest?"

0:49:00.8 SC: Emily Wilson was a great guest, I thought, here on Mindscape. And of course, she's gone on to great things since then. Great things described as a lot of public recognition for her relationship to the movie that came out, the Odyssey, directed and written by Christopher Nolan. I saw the movie. I love the movie. I thought it was great. It takes a lot of liberties with the Odyssey. So I can absolutely imagine why or how, if you're a translator of the Odyssey, if you're a professional classicist who has lived and breathed that world for a long time and tried to make it as accurate but also as contemporary and relevant as possible, that you would not like the much more free-form version of an adaptation that Christopher Nolan did. And that's legitimate. And so Emily Wilson had real problems with the Odyssey, and she wrote about them, and that caused a lot of foo-foo, and who cares, really? Let a thousand different versions bloom. It is... I don't want to say that too glibly because I do think that there's an interesting issue when you're doing an adaptation of one work into another work, which is, of course, when you do an adaptation, you have to make it your own, right? You're not just trying to do a shot-by-shot remake of some previous movie or something like that. Otherwise, there's no reason to do the adaptation. Just watch the original, right? You're adapting it, you're making it different. That's okay. On the other hand, I do absolutely have sympathy for people who think that you should not change the most essential features of what made the original distinctive. So, for example, if you were going to make a movie about Sherlock Holmes, and you said, "I know, what if Sherlock Holmes was stupid and was not a very good detective? That's my adaptation." I don't think that makes any sense. You shouldn't do that. Why would you do that with Sherlock Holmes? The whole point of him is he's a good detective, right? So you're allowed to make your own changes, but you shouldn't miss the entire point or, again, what makes the original very distinctive.

0:51:13.6 SC: And I suspect... I didn't read Wilson's disagreements in the London Review of Books and elsewhere, but I suspect that that's the kind of thing that she thinks, because Nolan did create a motivation for a lot of Odysseus's actions, which I don't think were in the Odyssey. And I think that they worked. I think that as a movie, it worked great. But it wasn't the Odyssey. It wasn't the Odyssey as Homer wrote it. And so I think that someone like Wilson would be naturally mad at it. But it doesn't matter because we don't have repeat guests here on the Mindscape podcast. So even if people do really great things, which all of my previous guests have done, there's too many other people I haven't. Until we get to the point where I've had interviews with more than half of the people on Earth, then I think there's no real reason to have repeat guests. There's plenty of interesting people still out there. On the other hand, I'm super thrilled that she is doing another adaptation, a translation of the Odyssey, because that's a really cool thing to do. It really helps drive home the fact that there's a million choices that you make when you do a translation of anything, especially a translation of a poem where the rhythms of the original language might not be the rhythms of the language that you're writing in. So there's... Right from the start, there's an infinite number of choices you're making. And just to show how even just one person with presumably the same inclinations and knowledge base and whatever can create two very different translations of the same ancient work is a fascinating project. I like it. I'm interested to see how that goes.

0:52:57.4 SC: Okay, I'm gonna group two questions together. One is from Kevin James, who says, "Is there any relationship between dark energy and the many dimensions of string theory or any other theories that include greater than three spatial dimensions?" And then Ivo Majors says, "Cumrun Vafa recently gave a summary talk on his dark dimension proposal. He argues for a dynamic dark energy field and a fourth spatial dimension, which appears to us as dark matter. He also shared encouraging first signs of evidence from the DESI, Dark Energy Spectroscopic Instrument, survey. What do you think are the prospects of this being right? It sounds like the craziest idea since quantum mechanics."

0:53:35.7 SC: So Kevin's question is answered by what Ivo is referring to. Absolutely, there could be a relationship between dark energy and the many dimensions of string theory. What happens is that the short version here is that when you have extra dimensions of space and you curl them up so they're invisible, what invisible means is that you can't, as a large macroscopic person, directly perceive what is going on in these dimensions. And one way of thinking about them is in quantum mechanics, or really in any theory based on waves, there's a relationship between energy and wavelength, which is an inverse relationship. Short wavelengths mean high energies, and long wavelengths mean low energies. So the idea of these curled-up dimensions is generally, not always, but generally that the dimensions are curled up very, very tiny. And they're curled up so tiny that anything that is changing on the scale of those dimensions would appear to us as a very high-energy particle, higher energy than we've ever seen at the Large Hadron Collider or anywhere else. And therefore, you don't see the dimensions directly, but you do see them indirectly because the size and shape of the extra dimensions feeds into the parameters of the low-energy particle physics theory, the effective quantum field theory that you have at low energies. So things like the cosmological constant or the Higgs boson mass or whatever can depend on the size and shape of the extra dimensions. And in some theories, you could hook it up so the size and shape of the extra dimensions changes gradually from place to place or time to time, and then that would show up as a new field. And that field might be the dark energy field or the dark matter or whatever. And this is what Cumrun Vafa is thinking about. We had a podcast episode with him not too long ago. You could go back and listen. He did a very good job of giving a sales pitch for that idea.

0:55:35.7 SC: But just to back up a little bit, a lot of Vafa is an extremely successful string theorist, very mathy guy, like many string theorists are. What is he doing messing around with particles and astronomical data and things like that? The answer is... Actually, I shouldn't... I'm gonna say this, but again, I haven't looked up the history in detail here. But in 1998, we found that the universe was accelerating, right? Something that we attribute most straightforwardly to a positive cosmological constant, a positive vacuum energy in empty space. Now, string theorists had been saying for a while that positive vacuum energy is not possible in string theory. So you might think that discovering a positive vacuum energy makes people go, "Ah, string theory is ruled out." [chuckle] But that's never how science works. What should happen, and I think that in right-thinking people did happen, is that your credence in string theory being the correct answer went down a little bit when we found that the universe is accelerating, because it is not necessarily impossible, but hard to accommodate a positive vacuum energy in string theory. And I think, and this is the part of the history that I'm not completely sure about... Well, sorry, I'm skipping a step. So the one part that I am sure about is this problem helped inspire, not was 100% responsible for, but helped inspire work on the string theory landscape. This is the idea that, okay, you can curl up extra dimensions, maybe you can curl them up in many different ways, and maybe those different ways correspond to different versions of low-energy physics. And maybe inflation... There's a lot of maybes here, as you can tell, makes all of those different possibilities real in different parts of the universe. So there is no once-and-for-all theory in string theory that says the vacuum energy is number X. Rather, there's a distribution of values of the vacuum energy in different regions of space, and we're gonna live in a region where it's small.

0:57:42.5 SC: And so the landscape of string theory is a metaphor that is supposed to tell us all of the different low-energy theories that you could get by starting with string theory and then curling up extra dimensions. And people realized that one spinoff of this idea is that there's gonna be some low-energy theories that you could just write down as a low-energy person, as someone who didn't know about quantum gravity or whatever, but that you could never derive from string theory. So on the one hand, there's many, many theories you could get at low energies that you could derive from string theory, but maybe there's many, many that you can't. So maybe not all low-energy theories are created equal. And the low-energy theories that are not derivable from string theory are dubbed the swampland, the set of low-energy theories that have no way of being consistently completed at high energies with quantum gravity. And so maybe... So there's been a bunch of conjectures about what theories are swampland-like and which ones are not. I have not followed all that stuff. But one strong idea that many people have, I think that Vafa has, is that the simple-minded idea of a constant cosmological constant that never goes away is in the swampland, is not a theory that you can get from string theory. But maybe you can fit the data in different ways. For example, if it's dynamical dark energy, if the dark energy is really changing a little bit with time. So that's compatible with string theory and with the string theory landscape. So I would say... I know string theorists don't like to think this way, but I would say that the discovery of the accelerating universe should lower your credence in string theory a little bit. I would also say that if we go on to discover that the dark energy is evolving with time rather than constant, then that should raise your credence in string theory a little bit. I think that's a perfectly legitimate thing to think. I don't think that the evidence we have right now one way or the other is very definitive. So I think it's okay to wait and get better evidence before we get too excited.

0:59:53.0 SC: Keith asks a priority question and then says, "Only if I want feel like it though." I being me, Sean. That kind of undermines the point of a priority question. [chuckle] You can just ask a question and I will answer it if I feel like it and have time. If someday only like three people ask questions, I will ask all of them. That's how it works. Priority questions are you force me to do it, so I'm gonna do this one. "From an engineering perspective, I really appreciate the new 'Four Faces of Information in the Natural Sciences' Nature Reviews Physics perspective paper. In Figure 1, I'm wondering if I am reading the labeled subjective-objective axis correctly, distinguishing information as a useful pragmatic construct from information as something real. I initially read subjective in the probabilistic sense. In engineering, we can be agnostic about whether P(x), a probability distribution, is interpreted subjectively or objectively, as said in Box 1. Perhaps this is the mechanical engineer in me talking, but if subjective partly implies useful or operational, it seems kind of ironic or something to place thermodynamics on the opposite side. Ultimately, what exactly is the vertical axis meant to distinguish when calling it subjective-objective, and why do the engineering and thermodynamic faces fall on opposite ends of it?"

1:01:10.5 SC: So this is referring... Keith's referring to a paper that I was a co-author on that just came out called, as he says, "Four Faces of Information in Natural Systems." Fernando Rosas, who is a scientist in the UK, Portuguese, I believe, originally, computer scientist, information theorist, led this project with a bunch of co-authors to talk about defining information in different ways. And it's really a descriptive paper, not a prescriptive paper. We're not telling people what to do. People can get upset when they think we're telling them what to do and they don't want to do it that way. We're trying to help people navigate the fact that different subfields of the natural sciences use the word information in very different ways. And so there is a figure, Figure 1 of the paper, attempts to organize all of these different ways in which we use the word information. The paper is built around the idea that there are four faces of information. And the idea... It's Fernando's idea, it's supposed to be four ways in which information presents itself to the scientist. And the four are engineering, statistical, thermodynamic, and ontological. This does not map directly onto the kind of people who talk about information in this way. An engineer might talk about the statistical face of information or the thermodynamic or whatever, but it's the role that information is playing in the particular context in which it's being used. So engineering is the word we actually use to describe Shannon information, the use of information in communication systems. Statistical is the sort of the statistician's version of information where you're just talking about different probability distributions and different relationships between them, and things like mutual information can be used to do that and how well you fit the data, so forth. Thermodynamic face of information is the information that is used to characterize the relationship between energy and entropy and free energy and work, limits of control and computation, things like that. And then there's the ontological face of information, which is the idea that somehow information is what something is made of, maybe the underlying substrate of reality itself.

1:03:32.7 SC: And so we... Well, we, Fernando invented this cute 4x4 table where he put the four different faces. We originally just had the four faces. Actually, that's not even true. We originally just had the three faces. [chuckle] When Fernando recruited me on the project, there were only three faces of information. And I realized that there was a kind of sliding together of two different versions of information, sort of the thermodynamic way that we use to talk about thermodynamic systems, but then this what we now call the ontological face, where it really is "it from bit," information makes up the universe kind of thing that string theorists or whatever would talk about, quantum information theorists. So we separated it out. It became four faces of information. And once there were four, it was obvious to think of them as a 2x2 grid, right? And it's not... I'll admit, not a perfectly comfortable fit, but the vertical axis is objective versus subjective. Horizontal axis is concrete versus abstract. So concrete is engineering and thermodynamic. Abstract is statistical and ontological. Subjective is engineering and statistical. Objective is thermodynamic and ontological. Again, I wouldn't take the axes overly literally, they're supposed to help you get a feeling for the kind of ways in which these different faces show up. Subjective, I'm not quite sure how to think about what Keith says when he says, "if subjective partly implies useful or operational." I don't think it does. That's not what is meant by subjective, at least not in my mind. Subjective versus objective is, everyone agrees on what this is, that's objective; or different people might assign different meanings or probability distributions or whatever to the same thing, right? Like taste in ice cream is subjective because some people might like vanilla more than chocolate, vice versa. But everyone objectively believes that vanilla ice cream is closer to white and chocolate ice cream is closer to dark colored, right? So there's a difference between subjective and objective that way, but they're both super useful or operational. And thermodynamic is supposed to be concrete and objective. I mean, that is exactly the origin of thermodynamics itself in the thinking about steam engines and entropy and so forth, and only later on did we realize that information is involved there. So anyway, anyone can read the paper if they want. 'Four Faces of Information' just came out. Sadly, and this was not my idea, but sadly it is behind a paywall at Nature Reviews Physics. But if you look hard enough, including on my Bluesky account or Fernando's Bluesky account, there is a link to an open version where anyone can read it if they want.

1:06:32.2 SC: Kira K. Says, "I am 28 years old and have an immense passion for physics that I never acted on. Now I find myself wanting to go back to school to ultimately get my PhD in astrophysics or theoretical physics, and I currently only have an associate's degree. If I started now, I would achieve this by the time I'm in my mid to late 30s. My question is, realistically, what do you think the job outlook would be like for someone like me finishing at that age? I worry that my chances at working in academia would be slim and the job market would be too competitive for me. I've seen differing opinions online on whether it would be truly worth it, so to speak, and even my brother, who's a physicist, seemed to think it's a bad idea."

1:07:09.0 SC: You know, this is... I'm gonna say something I always say for these kinds of questions, I can't actually give you an accurate answer because everyone's situation is different. And especially in a question like this, what are the chances that you'll make it as a physicist? That depends much, much more on your talent and your drive than it does on your age. Your age matters a little bit, but if you're the next Einstein and you're a genius, you're gonna make it even if you're a little bit older than everybody else. If you're just not that great at physics, or if you're just not that into it, or if you don't take education and training very well, then you won't do so well. And I have no way of knowing how well you do on those various scales. I wouldn't discourage someone from pursuing a PhD or something like that in physics or astrophysics just because they're 28 years old. That's not a cut-and-dry reason not to do it. But I do think you should just have your eyes open and be realistic about what the chances are. Like, how good are you at physics? How really passionate are you about spending late nights studying it? And I like to, roughly speaking, tell people that getting a PhD in physics is a worthwhile thing even if you don't end up becoming a professional physicist. It's good training for all sorts of things. And the chances of becoming a tenured professor of physics after you get your PhD are always small. They're never zero, they're always small. Things like your age and other things can sort of push the probabilities up and down, but they're not deterministic. You're not just gonna say, "If I do this, I can't succeed." So ask yourself how much it would be worth to you to really dive into doing research in physics at the level of cutting-edge research in whatever area it is. The effort that that takes, would that be intrinsically rewarding to you even if you didn't get a job? And also, what are your chances of getting a job? Again, being a little bit older might be a little bit of a disadvantage, but it's so little I wouldn't even think about it that much. There's other factors that are much more important there.

1:09:20.1 SC: Chris Kaltwasser says, "In a recent episode, you spoke of a very accomplished researcher as someone worthy of winning the Nobel Prize in physics, but who doesn't share your acceptance of Many-Worlds. You uttered something close to, 'What he lacks is a bit of courage here.' I got a chuckle out of that. I love Many-Worlds, but I'm also a fan of Alfred North Whitehead and his process philosophy. However, his most prominent contemporary spokesperson, Matthew Segall, thinks Many-Worlds is bogus. On the other hand, I'm sympathetic to Segall and Chalmers in their belief that consciousness or feelings may be fundamental. Reflecting on this sort of intellectual wandering, I found that I want to fine-tune my compass. What distinguishes intellectually courageous speculation from unjustified speculation?"

1:10:05.2 SC: Well, I don't think it's about speculation, really. When I talk about being courageous, it's somewhat tongue-in-cheek. So again, don't take this overly seriously here, but courage is not about having a speculation and imagining that it might be true. Courage is having a theory already. So you have some credence that the theory is correct or not, but let's imagine that your credence is pretty high that it's correct. The courage comes in in accepting the implications of that theory, not in speculating about new theories. When I say that certain people lack the courage to really take Many-Worlds totally seriously, I'm thinking of those people who already treat quantum mechanics essentially as if Many-Worlds is true. They just don't want to say, "And there are all those worlds," right? So they believe the equations; they just don't want to accept the implications of those equations. And I think that this is something the history of physics has shown us over and over again is a mistake. Some speculations will turn out to be right, some speculations will turn out to be wrong, but one easy way to make a mistake is to accept a theory but not its consequences. And I think that that's something that for quantum mechanics and Many-Worlds is a tempting way out of the apparently absurd implications of what the quantum mechanical equations are trying to tell us.

1:11:27.0 SC: Nalita S. says, "You mentioned in your interview with Brian Greene..." This was a YouTube interview I did, it's not a Mindscape thing. I think that's what Nalita is referring to. Brian was a guest on Mindscape, but I think this is the more recent YouTube thing for the World Science Festival. "You mentioned in your interview that your Catholic school years left you with plenty of stories. If you could share just one, which experience best illustrates how your younger self learned to question authority, examine beliefs critically, or separate conviction from evidence? I'd love to hear the human story behind that part of your intellectual journey."

1:12:03.7 SC: Yeah, I do have lots of stories, but I don't know if they're especially illuminating stories. There's some silly ones. I'm trying to think of relatively illuminating ones. I was already pretty atheist internally when I got to Villanova as an undergraduate, but I became more convinced of my atheism. I became more willing to say, "I'm an atheist," when people ask me about my religious beliefs. And part of that was being exposed to other people who did. Role models are important, like it or not. Just seeing that it can be done is crucially important to doing all sorts of things, whether it's being a vocal atheist or whether it's just going to college, right? Seeing that someone like you can do it does really matter. So by the time I was there, I was never... I don't think I ever tried to be obnoxious about it. I was surrounded by plenty of people who were not atheists. Let's put it this way, a standard charity, like an undergraduate charity nonprofit group that got a lot of support, was Villanovans for Life, the anti-abortion group, right? The campus was not especially conservative or liberal one way or the other. There were people on both sides, but there was a presumption, like, of course you should be against abortion because we're good Catholics here. I did organize... I was very active in the Honors Program and in the astronomy department, of course. The Honors Program did lots of little things. In fact, I helped pioneer a sort of... I think we called it the Friday Colloquium or something like that. But some weekly or monthly get-together where there'd be a talk for the Honors Program people to talk to themselves or to bring in external people.

1:13:44.3 SC: That just hadn't existed before. And I edited the Honors Program magazine for a year and things like that. So we put on this panel discussion as part of one of these programs on abortion, actually. And it was kind of interesting because the people generally... It was not an extremist school in any way. So we had a panel discussing, I think that there was some recent decision from the Supreme Court. This is the early '80s or the mid-'80s. There's always a recent decision by the Supreme Court that is relevant to be talked about. And so I forget exactly what the subject was. But we had a priest on the panel. We had someone who was arguing in favor of abortion rights. We had a biologist. I remember the biologist, who was like a popular biology professor at Villanova, and he said, "Why do you want me on this panel? We don't need... This is not a biological question. This is a moral question, or a political one." And then he got on the panel and he listened to what other people said, and he said out loud, "Boy, we really needed some biology on this panel, 'cause some people don't really understand how the biology of all this really works." So that was useful. But I liked... The priests were generally pretty good. The priest on the panel, I have no recollection of what he said, but at some point he was talking and someone from the audience said something, and the priest guy says, "You know, I worry that my being here as a priest..." and he was in the uniform, right, with the collar and everything, he said, "I worry that I'm being talked to and responded to as a priest rather than as a person saying things." So he took off his collar as a rhetorical move to say, "Just listen to my words. Listen to what I'm saying." And I thought that was very nice and very educational. So we had different useful things in various ways.I'm trying to think of any others. I took lots of courses where we talked about the existence of God, and I always argued that God did not exist, and that was fun. But I never really got into any real trouble. The closest I got into trouble was there was one incident, and I forget the details now, but there were a couple of times.

1:15:58.4 SC: So not only is it a middle-of-the-road Catholic institution, but it's at the height of the Reagan administration, okay? So the country as a whole was feeling conservative in various ways. And when that happens, you get various incidents along free speech issues. I know that there's people who've convinced themselves that the left is the real threat to free speech, but I know it's not true. I've seen it in action. So there was some event where the administration shut it down because they didn't like this, and this caused some controversy on campus. And at the time, I was the president of the Villanova Political Union, which was supposed to be some kind of political debating society on campus. It kind of failed. It kind of sort of almost went... Almost disappeared before I was there. So even though at some point I was the president, there were only two members. So it wasn't very impressive being the president. But I proposed a debate. In fact, I proposed that I would debate one of the priests, one of the Augustinian priests who was the dean of students or whatever, about whether or not this particular cancellation was a good idea. And the guy accepted the invitation and then he backed out at the last minute. And that was sort of frustrating. We never did have that debate. There were various letters to the editor that appeared in The Villanovan, the student newspaper. And the only thing remarkable there is that I wrote a letter and they misspelled my name, so it was spelled "Stan" instead of "Sean." They turned the E to a T. So some of my friends jokingly gave me the nickname Stan for the rest of my time at Villanova. So overall, nothing very dramatic, nothing very hilarious for that matter, but just I was looked upon tolerantly as the weirdo atheist, and I think that was fine. I've been in plenty of other situations where almost everyone around me was an atheist, so it's always good to be on the other side of the fence.

1:17:54.8 SC: AJ says, "Is there any non-theoretical evidence for there being a cosmological multiverse?"

1:18:00.3 SC: Well, I'll give you two answers. The quick answer is if what you mean by that is some observation, something in a telescope or a microscope or whatever that says, "Oh, look, here's the multiverse," then no, there is no evidence of that form. But the real answer to this question, unfortunately, needs us to think deeply about what you mean by evidence. I know no one wants to do that, right? They just want it to be easy. When I say no one, I mean scientists. Philosophers love to do this, but scientists just want these things to be cut and dried. All the training, all of the selection for picking scientists is a selection for people who like things that are cut and dried, who don't want to have all these little fiddly nuances. They just want words to mean what they mean and move on with it. And in questions like this, that's just not available. Evidence is actually a subtle concept. And this is something that is very directly, once again, fitting into a Bayesian way of thinking about theory choice and theory updating. Whenever you make any observation of the world, you are allowed to use that observation to increase or decrease your credences in all the various theories that might have different likelihoods for those observations being made. So that can be extremely indirect, right? So for example, if I invented a theory that was really, really compelling, like, "Oh yeah, this theory just totally works," and this theory predicted a single universe with exactly the right value of the cosmological constant that we currently have, then my theoretical invention would shift people's credences by quite a bit. Right? Because you're showing that it can be done. You now have a theory you didn't have before and many people suspect you can't have. So a lot of people who have high credences for a multiverse would be happy to switch to a theory that was compelling in its own right and wasn't a multiverse. And those credences can switch without any data being collected, just with a theoretical advance. So in that broad sense of evidence, like what are the pieces of data? The question you ask as a good Bayesian is, what is the likelihood of seeing this data if this theory is true versus that theory is true? Right?

1:20:23.3 SC: So right now, I would say the best theory we have for an explanatory account of the value of the cosmological constant, for example, the vacuum energy, is the multiverse. That's not to say that it's the only theory out there. It's not to say that it's necessarily true. It's not to say that all the problems have been solved. But of the theories that have been proposed to explain the value of the cosmological constant, the multiverse does the best job. So therefore, the existence of a non-zero, small positive value for the cosmological constant is evidence for the multiverse in that Bayesian sense. So to put it in even more starker terms, we discovered the acceleration of the universe in 1998. Before 1998, people who thought about the value of the cosmological constant already knew that it was much, much smaller than it should be from some effective field theory back-of-the-envelope calculation. And in the space of all possible theories, it was easier for them to predict a theory, to imagine a theory where it was exactly zero, than to imagine a theory where it would be a small, non-zero number. Because maybe there's some symmetry or some dynamics or whatever. It just seemed weird to make it a small, non-zero number. There was no reason for that. The one exception was the multiverse. The multiverse, as analyzed by Steven Weinberg and others, made a prediction that the vacuum energy should be observable. It should be non-zero, but it should be small. And they even made the right order of magnitude prediction. So when in 1998 we discovered the accelerating universe, any principled, respectable Bayesian should have increased their credence in the multiverse and decreased their credence in a single predictive theory for the value of the cosmological constant, because the non-zero positive number was likely under the multiverse scenario and unlikely under the others. Now, maybe your credence in the multiverse was so incredibly tiny that even though it increased, it still is incredibly tiny. That's okay. Bayes doesn't tell you how to arrange your priors. It only tells you how to update your priors when you collect new data. So in that sense, there absolutely is... In that broader sense, there absolutely is evidence for there being a cosmological multiverse in the sense that there are features of the universe we observe that are more likely under the multiverse hypothesis than under others.

1:22:57.7 SC: Darrell Cutting asks a priority question. "You recently discussed how the theory of Boltzmann brains is causing some concern in theoretical cosmological physics today. I do have difficulty in conceiving how such objects could be amassed materially from the increasingly diffuse atoms in the effectively infinite future. Dark energy is causing acceleration of the expansion of the universe. Isn't it possible that the reduction in matter density will be fast enough to outstrip any possible clumping of a complex random fluctuation, such as a macroscopic brain, which requires actual atoms whose protons have not yet been decayed, even given infinite time?"

1:23:32.6 SC: Well, that would be completely true if the world were classical. But the world is quantum mechanical, and that changes things. The point of the Boltzmann brain problem is it didn't become a problem until we discovered the cosmological constant in 1998, until we discovered the acceleration of the universe. And again, it still is only a problem if it truly is a cosmological constant. If it's gonna be a temporary dark energy that decays away, then the Boltzmann brain problem goes away, basically because the universe doesn't accelerate forever. But if it's a constant cosmological constant, the universe does accelerate forever. And in an accelerating universe like that, we have a horizon around us. That is to say, there are places far enough away that they're moving away from us apparently faster than the speed of light, and we will never see them. And Stephen Hawking and Gary Gibbons back in the 1970s pointed out that these cosmological horizons are very, very similar mathematically to black hole horizons. And just like black holes give off radiation, there is a sense in which, and there's details here I'm not gonna go into, but there's a sense in which these cosmological horizons give off radiation. And this radiation is at a very, very low temperature. So usually it would be unobservably small, but it's a temperature, and that means there are random fluctuations. And so the cosmological horizon, so the story goes, will occasionally give rise to fluctuations that give rise to Boltzmann brains. It's very rare, it's very unlikely, but you multiply the frequency of it happening by infinity, 'cause infinity is the number of years, and you find out that it actually happens a lot. So it's not the ambient protons and atoms that are surrounding us now that will be doing the assembling. It's these virtual particles popping into existence from the quantum vacuum because of the future horizon that we'll get in an accelerating universe.

1:25:28.3 SC: Thegreatdeceiver says, "I'm curious about your AMA recording routine, if you have one. Is it an early morning with a couple cups of strong coffee thing, or maybe just a relaxing weekday evening thing? Or my personal favorite, a Friday night with a few cocktails or glasses of fine wine thing? At three to four hours per episode, usually half in the bag, Sean answering our questions would be highly entertaining. Just saying."

1:25:51.3 SC: Yeah, I'm not gonna do that. I'm not gonna have cocktails or wine while doing the AMAs, in part because usually I do them on a Saturday. I'm doing this one on Monday mostly. Sometimes it takes more than a day, not because it takes literally more than a day, but because it takes different sessions on different days. But today it's Labor Day, which is a Monday, so I can just record it. And I start in the morning. If it's a three to four hour episode, it takes twice that to actually record. So I don't want to start drinking in the morning. [chuckle] And also clearly, as we mentioned in the intro, I already am making enough mistakes without having had anything to drink, so I don't think that's gonna improve anybody's success rate of answering questions like this.

1:26:37.9 SC: Andreas Nygaard says, "A recent paper in the Monthly Notices of the Royal Astronomical Society by Valentina Crespi and her collaborators argues that Sagittarius A*, which is the black hole at the center of our galaxy, could in principle be a horizonless self-gravitating core of fermionic dark matter rather than a supermassive black hole. They argue that such a compact core can reproduce the observed S-star orbits while the surrounding dark matter halo also fits the Milky Way's rotation curve. Previous work shows that it could even produce an Event Horizon Telescope image resembling a black hole shadow. How seriously should we take this kind of alternative? Is there a fundamental theoretical or observational reason to strongly favor a black hole over a compact fermionic object? Or is the nature of Sagittarius A* still more observationally open than we usually assume?"

1:27:28.9 SC: I think that I don't take it very seriously. I haven't read the paper. I don't know who the authors are. I have no reason to doubt that it's a completely plausible, conceivable theory. But it's a much more speculative theory than just saying that it's a black hole, right? The reason why most people would put much lower credence on something like this than the black hole is that black holes are made of gravity, and gravity is known to exist. And we've had Einstein's theory for over a century now. We have the predictions of how to make a black hole. It's easy. Adding some new, somewhat exotic fermionic dark matter particle is something that you can try to do, but why doesn't it just collapse to a black hole? The matter density in Sagittarius A* is sufficiently high that it's very close to being a black hole if it's not actually a black hole. There's enough stuff there in a sufficiently small region. So I would presume that you kind of have to work hard to avoid making it a black hole. And you don't really gain anything by making it something else rather than a black hole. Maybe your dark matter candidate gains things on other issues like the spiral galaxy rotation curve or something like that. I don't know 'cause I haven't read the paper. But this is the way science works. That's okay. People propose speculative scenarios. And the importance of that is not that someone reads it and goes, "Oh, that's probably right." It's that someone reads and goes, "Oh, that would have this other implication that I can chase down." And it might take a long time before people realize, "Oh, this idea actually does pay off somehow." Remember, Schwarzschild wrote down the Schwarzschild metric in 1917, and black holes weren't really even begun to be understood until the 1950s. So it can take a long time between when someone has an idea and when we begin to appreciate all of its implications. And there's a lot of ideas out there, a lot of papers appear, et cetera, et cetera. That's fine. And there's a process by which the paper could just die a peaceful death, no one pays attention to it, or people could realize, "Oh, actually, that's kind of promising for this or that reason. Let's follow it up. Let's look for an observation to test it, or let's look for a theoretical consequence." All of those things. It just takes time. That's just how science works. Sorry about that.

1:29:57.3 SC: Red Lynx says, "An acquaintance of mine told me that she is convinced that there is a ghost in her house." She says, "Things in the house have moved unexplainably and that she has even seen the ghost itself at one time. As someone not believing in ghosts, I didn't really know what to say in response. How do you talk to someone like her?"

1:30:15.5 SC: Well, it depends a lot on her. Like, is she asking questions? Is she saying, "I think there might be a ghost. Do you think that's plausible? How should I think about this possibility?" Or is she saying, "Oh, there's a ghost, I know it." In one case, you might try to make a case that in all the possible theoretical explanations of things moving around the house, ghosts should be pretty low in her list of credences, given other things we know about the universe. But if she's already convinced, then don't bother. I wouldn't really... I don't get much pleasure personally out of trying to beat someone out of their beliefs, even if they're irrational. I will talk to people who want to be talked to and want to learn something. Maybe I will learn something. Wouldn't that be weird? But I'm not trying to convince people of true things who don't want to be convinced of a true thing at that moment.

1:31:12.2 SC: Jason Bryant says, "The recent eclipse got me thinking about inference in design. If I found 10 stones perfectly aligned and equally spaced on a beach, I'd reasonably infer agency rather than coincidence. Yet the extraordinary sun-moon apparent size match doesn't seem to justify the same inference. What is the key epistemic difference between these cases? And when does an improbable coincidence legitimately become evidence for design rather than a post hoc pattern we happen to notice?"

1:31:40.9 SC: I actually think this is a super good question. I think if you asked me a couple years ago, I might not have been that interested. But the more I think about it, the more questions like this are super important because it goes to what we mean by fine-tuning and unnaturalness in nature. A classic example... People have talked about this stuff, of course. I think that one of the first people to talk about it must have been at Cambridge University because there was some example like, what if the stars in the night sky had arranged themselves to... They didn't move, but they were always out there, arranged in the figure of the words, "Cambridge is the best." That's not a small number or a fine-tuning in the conventional sense, but it would make us think that it was not just purely random, that there was some reason why that had happened. And so it's very interesting to think about what are the characteristics of some thing we observe in nature that makes us think it's designed rather than completely random. My suspicion is that it has something to do with algorithmic compressibility. That is to say, if you find... What was the example here? 10 stones perfectly aligned and equally spaced on a beach. Well, in the space of all possible arrangements of stones, that's a very, very tiny fraction of all the possible arrangements. But if you had some mechanism, whether it was design or something else, that was encapsulated by a very short, elegant algorithm, then it'd be very easy to come up with 10 perfectly aligned stones on the beach. Therefore, in the space of all possible arrangements of stones, something like that should increase your credence in the algorithm arranging them and decrease your credence that it's purely random. But again, I don't know anyone who's worked that out in a fully rigorous way. Probably someone has tried out there in the world of philosophical questions. But it is a good question. I think it has to do with how we coarse-grain all the ways the universe could be. Some of them look weirder to us than others, and the weird ones we try to find explanations for.

1:33:58.3 SC: Peter Bamber says, "In a recent AMA, you said emphatically that compatibilism means what it says, that the underlying laws of physics and free will are compatible with each other. You then went on to say that you thought that the term free will might not be helpful and that you tried to stop using it. Before a podcast with Sam Harris, you'd asked him whether the two of you could avoid the term altogether, instead talking about what happens and what doesn't happen. Substituting into your definition of compatibilism, we get that the underlying laws of physics are compatible with what happens and what doesn't happen. This seems self-evident and leads me to think that compatibilism is a useless concept. I'm trying to understand what you're getting at. Is free will fundamentally a meaningless term? Does it add anything to scientific descriptions of reality? Does it belong only in the realm of theology?"

1:34:42.1 SC: Well, when I say... When I suggest to anyone that rather than talking about free will, we should talk about what can happen, what does happen, what doesn't happen, I'm not saying that those two concepts are equivalent. I'm just saying that the overall conversation would be clearer if we didn't, number one, spend a lot of time arguing about the definition of free will, or number two, using different definitions than each other. I just want to be clear. There's no substantive claim whatsoever in that move. I'm just saying let's talk in ways that we each understand what we're trying to say. I think that's a good way of making progress. Free will is not an empty concept because if you were Laplace's demon, it wouldn't exist. If you're Laplace's demon, if you really knew everything about what did happen and what was going to happen, then there would be absolutely no need for free will, in my version of it anyway. I'm a compatibilist because I'm not Laplace's demon, and I never will be, and nor will you be. The best explanation I have for how to describe the world at the large-scale, macroscopic level is one in which I can't predict exactly what will happen. Instead, I have a theory of the probability of different things happening, and that theory involves human beings thinking about things and making decisions. And, by the way, as I always like to say, everyone else has the same theory. Everyone else talks about people thinking through things and making decisions. Some of them refuse to call that free will. Some allow it to be called free will. That's where I don't care. I don't care if you want to call it free will or not. What I mean is talking about human beings as agents making decisions is almost unavoidable, and we should just let ourselves do it.

1:36:32.9 SC: Anonymous says, "Intelligence is highly multidimensional. What are some dimensions of intelligence you don't think that AIs post-LLMs are likely to exceed humans at? Computer chips operate at a frequency 10 million times faster than neurons, and so once you develop algorithmic parity, I can't think of any practical dimension of intelligence where brains keep beating silicon."

1:36:53.5 SC: Yeah, neither can I, honestly. I've tried to be consistent in saying that, in principle, there's nothing that a human brain can do that an AI can't do. The only question is whether the modern versions of AI are anything like that, are even the right first step to getting there. I suspect that in order to really do better than humans at all aspects of intelligence, we're gonna have to look at some radically different architectures for AI. But there's no reason why that can't be done in silicon as well as biology, in my own particular point of view. It might just have to be different even in the silicon aspect of things.

1:37:32.7 SC: Tyler Briggs says, "A single high-end desktop today matches the double-precision compute of a shared university cluster from two decades ago. You've often described yourself as a pen-and-paper theorist, but across theoretical physics as a whole, has ubiquitous personal compute shifted how the disciplines explore and stress-test new ideas? Or does breakthrough remain bottlenecked by insight, not hardware?"

1:37:55.3 SC: Depends exactly on what part of theoretical physics you're talking about. In the kinds of theoretical physics I do, it's not the amount of computational power that matters. It's that computation isn't what I try to do. I try to come up with new concepts that are relevant to understanding the puzzles of physics. So again, as with the answer to the previous question, if there's a completely new kind of thing which is able to generate hypothetical concepts with some promise to them and then stress test them itself and then suggest them, that kind of creativity is not there in the current generation of LLMs, but there's no reason why it couldn't be there in principle. What we have now, of course, forgetting about AI and LLMs and all that stuff, is the ability to do wonderful numerical simulations, solve numerical solutions to differential equations and all those things. That's very useful. In my most recent paper, we did have numerical simulations... Numerical solutions, I should say, not really simulations, to very simple differential equations. So that's useful to do, but it's not gonna tell me what quantum gravity is. That's just not a calculation, that's a different kind of thing.

1:39:12.2 SC: I'm gonna group two questions together. One is from Kevin's Disobedience, who says, "I've heard you say many times that the world is one thing. And while I agree with the spirit of this claim, I sometimes wonder whether our notion of monism, which is the idea that the world is one thing, isn't yet another aesthetic hangover from Greek philosophy. Do you suppose we might make more progress by worrying less about unification and instead approach metaphysics from the other direction? That is, by starting with the assumption that the one emerges from the many? In other words, maybe by thinking about the emergent world as pluralistic from the start, the evolution laws take on a new dimension in relation to one another." And then Gustavo Chavez says, "I wonder if some of the existing physical theories are fundamental or if we're far from the ultimate fundamental theory, if such a thing exists. Is it conceivable that there is no fundamental theory of everything after all? That we'll forever come up only with new effective physical theories, never finding a really fundamental one? And if so, do you think it's likely?"

1:40:11.8 SC: So for the monism versus pluralism thing, I mean, there is an ongoing debate in philosophy of science. Should we look for some unifying theory of everything, or should we describe the world in a more patchwork sense by the laws applying to one area versus another area? I think this is a little silly. I think this is a little misguided as to what's going on. I'm more or less completely convinced that there is a theory of everything. I'm not necessarily convinced that the theory of everything is simple. There is something that the universe does. Any full description of what the universe does would be the theory of everything. We're nowhere close to having that. We have various attempts at finding regularities that we call the laws of physics. The regularities might be very difficult to distill down to simple rules. I think that's what people are actually trying to get at when they talk about the patchwork of rules. There's still a consistency underlying everything. I may be able to describe the table as a table made of wood, things like that, or as atoms fit together with certain properties and so forth, but they have to be consistent with each other. So the patchwork has to overlap, and there has to be one way to describe the whole world all at once.

1:41:31.4 SC: I actually think that monism versus pluralism in terms of what the world is, is again, a very misguided question to ask. If I have two oranges, do I have two things, one orange and another orange, or do I have one thing, one collection of two oranges? That's a silly question. Come on. That's just two ways of talking about the same thing. I think that the interesting thing to do is to say that the world can always be described as a single thing, but it can also be divided up. And the interesting thing to ask is, what are the right ways to divide it up? Is there a single way to divide it up? Is there a unique way to divide it up? Are there many ways to divide it up? Are they all equal to each other? Are some better than others? Those are interesting questions. But of course, the world is both one thing and many things, that seems perfectly obvious to me. As far as whether we can find the single most fundamental theory, I don't see why not. Again, I don't think that we're there yet. But why in the world would we think that we can't? Or why in the world would we think that we can? [chuckle] All we can do is try to do better, try to keep doing better until we come as close as we possibly can. I think that the history of physics has been an amazing story of unification and extension of simple ideas to cover more and more ground. There's no reason in my mind to think that's gonna stop anytime soon.

1:43:02.5 SC: Malcolm McGregor says, "You've mentioned in the past that you occasionally get messages from scientific cranks. Maybe you have a different name for them, but the kind of people who are outside your field but are convinced they have made some kind of major scientific breakthrough, like a theory of everything, but who are obviously wrong due to their limited understanding. Is there a list of common sets of characteristics that these people have or red flags that you can use to spot them?"

1:43:25.2 SC: Yeah, there are, as a matter of practical fact. I don't wanna say that there's some necessary connection here because you can always be wrong, right? You can get a really brilliant breakthrough by some cranky person, cranky-sounding person, and you can get people who sound perfectly respectable who turn out to be cranks. However, the dividing line is actually more clear than you would think. You might think there's a spectrum, there's a continuum, but really there's not so much. It's usually pretty obvious. Very often, the most important deep principle of being a crank or a crackpot is that people don't ask questions. People are not interested in learning things. People who are of this cast of mind think they have it all figured out. And then when you realize that they don't actually have any formal training in the field and they don't understand quantum field theory, et cetera, then it's pretty obvious that they are in the crackpot basket. People who are serious want a dialogue. They want to ask questions, they want to go back and forth, they've read your papers, they've thought about other things going on in the field, they're knowledgeable about what's happening. People who are crackpots just want to talk at you. They just want to tell you what's going on, and it's kind of tiresome.

1:44:41.7 SC: So the advent of the internet and email has made it much, much easier to spread your crackpot ideas and to email them to everyone. The advent of AI has made it much easier to come up with a crackpot theory that sounds good. And everyone can go on an LLM and get the LLM convinced that their crackpot idea is a good idea, right? I mean, they're in some sense sycophancy machines. They want to tell you that you're brilliant, and they might start out by telling you you made a mistake, but you can always convince them to change their minds. So it's very easy to be convinced that you're on the right track even though you don't know really what's going on. But again, usually it's just even simpler than that. They just don't understand at a very basic level what it would mean to have a theory of physics, right? They're like, "Well, what if gravity is really force? Or what if energy is really time?" There's no equations. There's just some concepts being batted about in some kind of random way, and it's not anything like a real theoretical physics paper. So it's very rare, actually, to get one of these emails that is hard to diagnose.

1:45:55.7 SC: Heather Constant says, "My favorite two podcasts are Mindscape and another one called Science Vs. So imagine my happiness when I was listening to the latter's episode on free will and you turned up. I thought the episode and your portion were great. How did your participation come about?"

1:46:12.6 SC: They invited me, and in a moment of weakness, I said yes. Look, I get a lot of invitations to appear on podcasts. I'm sure some people listening right now are podcasters who've invited me onto their podcast, and I almost always say no. And I feel bad about it because I do want to support other podcasts and things like that. I just don't have the time. And neither making podcasts nor appearing on other podcasts is really my day job, and I mentally justify making my own podcast by limiting the number of hours per week that I put into it so that I can focus on teaching and research, which are my actual jobs. And if I spent time appearing on other people's podcasts, that would get out of hand very, very quickly. So I'm sorry if I've ever said no to anyone out there, but I do try to occasionally appear on one or another podcasts, and ideally, if I've heard of it already and I know good things about it, that increases the chances that I will say yes. So they asked me and I said yes, and I thought it went pretty well.

1:47:15.7 SC: Now I'm gonna group two questions together again. Edward Sackinger says, "Although there are many branches in the wave function of the universe, we experience only one. Similarly, although time is a continuous variable, we experience only one instant, namely now. Are these two effects related?" And Emil Zevkov says, "I feel like Everettian QM makes an interesting claim about human consciousness, that it's the kind of thing that lives in a single branch or world rather than spanning across branches. Have there been any interesting discussions on the topic?"

1:47:48.9 SC: So to Edward's question, let me get the time thing out of the way first, because that's very easy to do. When you say we experience only one instant, the right question to ask is, how in the world could it have been different, right? I mean, by the word "we", by the word "you" at any one moment, you mean you at that moment. The you that exists at one moment is experiencing just one moment. It literally could not have been any other way. Except there is a slightly more sophisticated question to ask, which is, why is it possible for us to assign coherence to our instantiation at any one moment at all, if indeed we also have experiences over other moments? And that is related to the Everett question. Like, if you think that there are other branches of the wave function, why do we have coherence only within one branch, not across multiple ones? But again, the answers there are pretty basic. It has to do with what we are. We are some way of talking about a collection of particles and atoms and forces in the human brain and body that are interacting with each other. You think about these interactions and the answer becomes pretty clear. If you have two collections of atoms that are making up a human body, but they're on two different branches of the wave function, they do not interact with each other in any way. So there's just no temptation or reason to think of them as two parts of the same mind. They're completely independent. If one gets shot by a bullet, the other one doesn't, right? At least in principle, at least possibly.

1:49:27.2 SC: So it just makes sense that consciousness exists on one branch of the wave function at a time. It has nothing special to do with consciousness or with wave functions. It just has to do with how matter works and the fact that brains are big macroscopic objects interacting with each other. When I say interacting, the time scale over which different atoms in your brain or whatever interact is very fast, right? Milliseconds, maybe, something like that. So there is some constant change in who we are as we evolve over time. But as long as a couple milliseconds pass, what is happening in our brain is different than what happened in our brain a couple milliseconds ago. And there's no interaction of different parts of the brain across time, right? If I poke a part of the brain, someone else's brain, I mean, either literally poke or just introduce an electrical signal or something like that, it propagates outward at some propagation speed. It doesn't instantly affect something in the future and then get affected back by something in the future. It affects what is going on in the brain right there and then. So it's that interactive coherence that both attaches our awareness to different moments of time and attaches our awareness to different branches of the wave function of the universe. In both cases, it has nothing special to do with conscious experience. It's just a feature of the interaction of particles that make up some macroscopic kind of object.

1:50:59.8 SC: Michael Lacey says, "It seems to me that much of the debate over free will is semantics. Those who don't believe in free will are really talking about libertarian free will, which compatibilists don't believe in either. And when you say free will is emergent, I think you mean our sense of free will. That still leaves the philosophical debate, but I agree that most free will skeptics are de facto compatibilists simply because it's difficult to live our lives or have a conversation without invoking a sense of free will. So in the end, I don't see much difference between the two positions. What are your thoughts?"

1:51:29.8 SC: So I'm not gonna dive into the answer to this. I've talked about free will too much, and in fact, I will advertise we have yet another free will episode coming up. We had the episode with Christian List not long ago to give the compatibilist point of view. Now we're gonna do an episode with an anti-compatibilist who just doesn't believe in free will. I'm perfectly happy with the claim that the argument over free will is about compatibilism... Sorry, is about semantics, and that's why I'm very happy not to use the term. People force me to do it. I don't wanna do it myself. You do need to use some terms to communicate with other people, but you need to define those terms as precisely as you can. However, that doesn't mean it's a meaningless argument. Semantic things can matter because there are implications. You might think that whether or not a person has free will is important for morality. Can you judge them right or wrong whether or not they have free will? I mean, there's certainly an argument that if people don't have free will, then you can't judge their actions to be moral or immoral, right? Those arguments are important and they're not semantic. And so I do think that it's useful to come up with a definition of free will and then stick to it and stop arguing about it. But people don't want to do that. So instead, I try to use other words when I talk about the relevant issues.

1:52:49.7 SC: David JS says, "As a Brit, I'm frequently surprised by the news which emerges from the US, most recently by the decision of Donald Trump and RFK to limit or at least make more expensive and difficult the access to childhood vaccines. Why would any politician want to do this? Is it a religious thing or a more secular disapproval of science and the concept of public health?"

1:53:11.1 SC: I'm also frequently surprised by the news that emerges from the US. I don't think the US is that special. I think a lot of crazy things emerge from all sorts of countries. I live in the US, though, so I'm more aware of the crazy and stupid things that go on here. And yes, certainly limiting access to childhood vaccines is one of the craziest, stupidest things we can do. There's some kind of pathology in certain segments of our political classes and other elites that really, really is just resentful of being told what to do. It's very immature, right? It's very insecure. It's very childish. But there's this idea that I shouldn't be asked to pay any taxes, I shouldn't be asked to follow any rules or regulations, I shouldn't be asked to take any medicines. And it all kind of goes together in this stew of resentment, which sadly caught on with enough popularity to get people elected over multiple election cycles, and we are paying the price for it. It's not very popular. Like, now that we're almost two years into the second Trump administration, he has historically low levels of popularity. People don't like it, but people are just not very good at voting for things that they're gonna like. That's a very big problem in our current political system, is just how to get people to have some more clear idea of the relationship between what they want and who they're voting for, because right now it's not very good.

1:54:39.6 SC: I don't think it's religious. I don't think it's even a disapproval of science or the concept of public health. I don't think that's really what's at the heart of it. I really do think it's this ethos of rugged individualism taken to a completely pathological degree. You can't tell me what to do. That's RFK's thing. And so you end up inventing new rules and new weird things about eating sauerkraut or whatever, that's one of RFK's latest things, and then making that the thing that everyone is supposed to do. And then when all the experts say you're bizarrely misinformed, you take that as a badge of honor. You're like, "Oh, see the experts, I've triggered them. See how weird and weak they are." It's all very sad. I don't have any... What I've just said is very casual analysis based on personal experience, not based on careful looking at the data or anything like that. So I'm very happy to get my priors updated on this particular set of issues. But the thing is that you don't get a lot of clear-headed analysis of this stuff because the people who are against it, like myself, just get really emotionally against it because people are dying for no good reason, and the people who are for it come up with weird justifications and ways to ignore all the bad stuff happening, and overall the level of discourse gets to be very low very quickly.

1:56:04.8 SC: Dylan Samuel says, "In the last AMA, there were a few questions about whether new proof-solving tools such as Lean, i.e., proof-solving programming tools, could have any use in physics or the physical sciences. Your response was that physical models of the world require experimentation, which is a requirement that pure mathematics does not have to contend with. That totally makes sense. However, I'm curious if you think that these tools could be useful to prove theorems within a certain theoretical framework. Say you have a theory based on axioms you find from some experiments. It seems like you could use something like Lean to make further predictions within your theory and see if those predictions agree with experiment to confirm the robustness of your theory. Do you see this as being useful at all in theoretical research?"

1:56:44.8 SC: Well, it's a little bit useful. It's not very useful, and it's not the useful thing. Like, if you want to do that kind of thing, what physics needs is something different than what math needs. If you read physics papers, which I encourage you all to do, you find very, very few theorems being proven. That's just not how physics works. You do see equations being solved, you see models being proposed, and you see approximations being developed to understand what the predictions of those models are, et cetera, et cetera, et cetera. But that's not proving a theorem. Proving a theorem is something much more rigorous and entirely suited to being checked by a computer. Of course, if I solve an equation, I can check that equation solution on a computer, and using computers is very useful for that kind of thing. But I don't need a formal, logically rigorous theorem-proving checker like Lean or anything like that. So I think that, again, what would be useful in physics is absolutely something creative, something that suggests new models that might map onto physical reality. That would be very, very useful. And various ways of going from those models to what the predictions of those models are. But going from a model to a prediction is very different than going from axioms to theorems.

1:58:09.9 SC: Alexander Knuckle says, "In your recent episode with Chandra Sripada, your guest argued that LLMs exhibit similarities to the human mind: fast and slow system versus in-weights and in-context, and so on. My gut feeling is that LLMs of current design cannot become conscious in any meaningful way due to their very forward-feeding process during inference. Do you have an opinion or hunch whether becoming conscious is something we could identify in ML systems, either in their behavior or in their layout?"

1:58:43.4 SC: Not really. I really don't. I have a line on this and I'll just keep saying it. I don't know what consciousness is. I don't pretend to know what consciousness is. I just don't think it violates the laws of physics. And given that, I think that given that the laws of physics are the same for a human mind and for an artificial system, I see no obstacles in principle to artificial systems becoming conscious. However, I also think that people are very overly simple-minded about what goes on in the human brain. You can abstract away some of what goes on in the human brain as information processing, but there are other things that go on in the human brain also. And if you capture the information processing part of it without capturing all of the more biological parts of it, maybe you're not gonna get to consciousness. I don't know because I don't know what consciousness is. But again, I see no obstacles in principle. If people want to make that happen, they should pursue that.

1:59:42.6 SC: Timji Nitsos says, "You mentioned in a recent solo episode, 355, that one of your favorite papers was called Locality from the Spectrum, where it's argued that when there exists a way of subdividing Hilbert space to get locality, it's almost always unique. But you mentioned that there has been some pushback, which claims more assumptions are required to get this idea off the ground. If the pushback is correct, does that change the picture you painted in your Mad-Dog Everettianism paper?"

2:00:09.1 SC: Not in any substantial way, not yet. There's been various pushbacks. Not a lot of pushback, but a little bit. Some of it is more convincing or interesting than others. I think that the most convincing suggestion is that you need not only the Hamiltonian and the spectrum of the Hamiltonian, the energy eigenstates, but you also need a specific state to start with before you do anything. Different states might give you different answers. I'm willing to believe that. I don't think it changes a lot the spirit of what we were saying. The actual thing that I'm interested in these days is extending from the finite-dimensional case, finite Hilbert space dimension, which is what we thought we had some control over, to the infinite Hilbert space dimension case, which is harder, maybe more interesting, maybe more important. So that's what we're trying to do now. And in that case, you might need even more than just a state. That I don't know. That's what we gotta figure out.

2:01:08.2 SC: Krasimir Gorov says, "Humans today are exposed to way too much information. Would you share your thoughts or habits around information processing and synthesis? E.g., when approaching a new topic for a podcast, research, writing, teaching, or just learning for fun, how do you work your way up and down the layers of emergence and organize scattered pieces of information into mental models with structured relationships?"

2:01:31.3 SC: Yeah, this is one of those technique questions. People ask me these things and I get it. How do you do this? How do you work? How do you write? How do you do the AMA and whatever? But the reality for me personally is that it's just not that systematic. I don't have an algorithm for organizing information, for remembering things. It's very slapdash and scattershot, and I rely on people helping me out and various bits of info that I've written down and recorded in various places, and it's all just not that systematic at the end of the day. I agree with the premise of the question that there's so much information out there in the world that it's very, very hard. This goes back to our podcast conversation with Zeynep Tufekci, where she made the point, which has been made by others before, but she's the one who sort of highlighted it to me, the currency of, the thing that is scarce now is not information, but attention. What do you pay attention to? This is the point I try to make with crackpots with physics papers, right? Like, I'm sorry, I just don't have the time to pay attention to this. Maybe you're right, but the chances you're right are much smaller than the chance that something that Ed Witten writes is right. So I gotta prioritize my time. And the way that. And the way that you break through that, of course, is to have a track record of doing good things and people recognize like, "Oh, yeah, this person has something interesting to say, so even if what they say sounds crazy, I'm gonna pay some attention to them." So there's different heuristics, I would say, like, who is this person? Do I know them? And I think a big one... I'm sorry, I'm slipping away from your question, Krasimir, back to the sort of crackpot question. But one of the big things that I think is when you come across a paper or a claim that is dramatic, when you're hearing something that would be amazing if it were true and you want to know, "Should I pay attention to it? Should I spend my time?" 'cause we're exposed to too much information, "Should I put effort into coming to grips with this?" You can't just say, "Well, if it were true, it would be important." That's the case for all sorts of crazy things, right? That's far too weak a criterion to actually use.

2:03:45.5 SC: So one of the most important things is if there's some obvious objection to the claim being made that I can think of, I would like to see that objection addressed. I would like to be sure that the person suggesting this crazy idea is just as aware as I am of the possible objections to it. And they say like, "Okay, you might think this, you might be skeptical for this reason, here's why that possible objection is not relevant here." That's a really good sign that the person knows what they're talking about rather than just making things up. But anyway, back to Krasimir's question. Yeah, you get your favorite techniques, your favorite news sources. I also think that in the AI era, two things will happen. One is good, one is bad. The bad thing is that so much of the information that people can get, especially online, is just gonna be degraded, right? It's gonna be fake, it's gonna be AI-generated. It's already easily enough to fake things by human beings faking them. AI just makes it much easier and you can fake all sorts of evidence, right? Pictures and videos and whatever. So the amount of unreliable and outright false information is gonna go way up. The slightly good thing, the slight, tiny little silver lining here is that it will become increasingly important that people get their information from trusted sources. So rather than just seeing something on Facebook or Instagram, people who actually care about getting the truth will pick and choose sources who they recognize as reliable. And therefore there is still an absolutely crucial role being played by reliable sources of information. Maybe those sources of information can even make a living doing this. That would be wonderful. But anyway, none of this is really answering Krasimir's question. I don't have a good answer to it. I just wanted to agree with the spirit of the question itself. There's too much information going around. How to deal with it is a big challenge.

2:05:49.3 SC: Floris Kuick says, "You described the Wheeler-DeWitt equation as suggesting a fundamentally timeless universal quantum state, with time potentially emerging through correlations between a clock subsystem and the rest of the Hilbert space. You also emphasized that in Everettian quantum mechanics, branching isn't fundamental but emerges through decoherence. How should I picture those two kinds of emergence together?"

2:06:12.2 SC: I think that there's no special connection between them except to the extent that branching requires time evolution. Okay, so again, I'm gonna... I should do this more often, but I'm gonna try to explain what's going on here to people who are not already experts. In Everettian quantum mechanics or in Many-Worlds, you have a single thing called the quantum state. It's a monistic theory. But that quantum state evolves with time. And there are subsystems, and you can talk about the entanglement or lack of entanglement between subsystems. And branching happens when some big classical-looking system, or even a tiny system, becomes entangled with its environment. That's the process we call decoherence, and that's an irreversible process. Given the kind of universe we live in with a very strong arrow of time, a very obvious distinction between past and future, branching is only gonna happen toward the future, not toward the past. So you need that time evolution. It happens one way in time but not the other way. So you might wonder if it can possibly work in a timeless formulation of quantum mechanics. And that's what you can... That's what maybe you get from quantum gravity, the Wheeler-DeWitt equation, et cetera. Time disappears and you have to talk about an emergent version of time. But here's the thing. No matter what your formulation of quantum mechanics is, time is fundamental or time is emergent, time still exists. And time has to exist because we experience it. I see it, I feel it, it's there. If you actually said, "I have a theory where not only does time not exist, it can't even emerge, it can't possibly exist," then I don't think your theory is gonna do a good job fitting the data. So the way this actually plays out in attempts to solve the problem of time in quantum gravity is by having some effective Hamiltonian, some effective time evolution operator, which is generated in an emergent way, and people argue about that. But the point is that I'm finally getting to here, once you have the effective, emergent notion of time, then you have branching and decoherence with respect to that. So whenever I talk about the Hamiltonian and trying to derive physics from the Hamiltonian, et cetera, people sometimes say, "Well, what if you have the Wheeler-DeWitt equation? The Hamiltonian is just the number zero, effectively." And I say, "Well, there's gotta be some effective time evolution, and I'm gonna deal with that. Whatever I was talking about the Hamiltonian doing, apply that to the effective Hamiltonian and go on your day."

2:08:47.1 SC: Speaking of which, I don't know if this is a related question or not, but Neil Glue says, "Is my subjective first-person experience deterministic or stochastic?"

2:09:00.1 SC: I'm not sure...I guess what I want to say to this question is the adjectives of deterministic or stochastic do not apply to subjective first-person experiences. The ideas of determinism or stochasticity apply to time-series data in the very broadest of possible senses. Right? So you might literally have some machine outputting numbers, and that's time-series data. But also the history of the universe or the history of some physical subsystem of the universe are time-series data in the sense that at every moment there's something the universe or the system is doing. Deterministic just means all I need to know is what happened in the past and I can, with 100% confidence, predict the future. Stochastic means the best possible prediction I can do on the basis of what has happened in the past involves a probability distribution over possible future outcomes. So I don't think that subjective first-person experience is a time-series data. So I don't know that these adjectives apply. There is time-series data in the sense of the list of experiences you have over time or the list of subjective impressions you have of what's going on in the world. Da, da, da, da, da. But that, I would say, is 100% supervenient on the physical stuff going on in your brain. Right? I don't think there's any mystical, ineffable substance that is causing you to be conscious. I think it's just an emergent way of talking about the physical stuff of your brain. So if you want to know if that is deterministic or stochastic, the question is: is physics deterministic or stochastic? And then you get into one more slightly fussy question, do you mean in principle or in practice? This is because quantum mechanics says very clearly that in practice, physics is stochastic. When we have a particle that is pointed in the x-direction of its spin, it's spinning in the x-direction, and we measure its spin along the z-axis. Quantum mechanics says indisputably that you cannot predict what you're gonna get. There is a random number that comes in. Sometimes it will be spin up, sometimes it will be spin down. Now, there is an in-principle question that says, is there a broader picture outside of our observable universe in which I can attach a deterministic theory to what's going on? For example, Everett does that, but also Bohmian mechanics does that, other versions of quantum mechanics do that. Great. All in favor of that kind of thing. It makes absolutely no difference to your personal, subjective, agentic experience of the world. As far as a person is concerned, the world is not deterministic.

2:11:49.1 SC: Okay, grouping two questions together. Not sure where these questions came from. They're good questions. I'm not sure why we got two of them the same month. Maybe something's going on. But Adrian says, "As a contributor to many different kinds of intellectual property, TV, movies, books, scientific papers, podcasts, et cetera, what are your thoughts on how IP currently works and how should it work?" And Population Thinking says, "I assume you believe in some form of protection for intellectual property. My question is what your justification for IP protection is. Is IP justified by the moral interests of creators to control their work, a utilitarian argument based on the social value of innovation and creativity, or something else? Put another way, do you feel someone who plagiarizes your work harms you personally?"

2:12:39.8 SC: Yeah, these are all very good questions. I think I'm gonna drop the ball on Adrian's question because I don't know enough about it. To give the short answer to Population Thinking's question, I do believe that there should be protection for intellectual property of some sort. Adrian is asking what exactly should it be, how does the current system work? It's a mess, I think, the current system. But I haven't done the work to figure out what a good system would be like. I do think that there's a lot of financial interests here, and sometimes those get out of control. I do think that once a creator dies, or at least a certain number of years after the creator dies, things should go into the public domain rather expeditiously. I don't really have a lot of strong feelings for protecting the rights of descendants of people who have created things. I do have strong feelings about protecting the rights of people who have created things themselves. And I do think that probably right now we're extending copyright and other protections far long past when we should. What is the moral justification for doing that? Again, I have not thought that through very carefully, but I would lean toward the utilitarian side of things for this particular issue. I say it again and again, I don't really have, despite the feeling that I should have, a totally well-thought-out moral theory. I'm a constructivist about meta-ethics, but that just says, "Okay, we're allowed to construct ethics in concordance with our moral intuitions." But the important question is, "Okay, so what ethical system did you construct?" I don't have very strong feelings about it. I'm not a utilitarian myself, but I do think that there are places where utilitarian kind of arguments have a role, and I think this is one of them.

2:14:34.9 SC: I mean, I say about this podcast, people don't like having ads or having to sign up for Patreon or whatever. They just want the podcast to exist and be free and piped into their heads. But if that were the case, I probably wouldn't be making it. I do get fun out of making the podcast. I learn a lot, don't get me wrong. But there are priorities, right? And I could do more research and I could write more books if I weren't doing the podcast. And if there were no monetary incentive whatsoever, maybe I'd still be doing it, I don't know for sure, but the incentive to do it would be less, let's put it that way. And if what I did was just put out there and instantly rebranded and stolen by somebody else, I would probably be making less money. And that's nothing compared to people who really have bestselling books or big movies or whatever. I'm in favor of incentivizing people to do good creative things, and I think that protection for intellectual property rights does that to some extent. I do think, like I said, it goes too far. Like, I don't know the details about this, but my impression is the estate of Albert Einstein exerts enormous control over when you can use a picture of Albert Einstein and things like that, which is a little silly. He died over 50 years ago, and they're not him, so what right do they have to do that? Would be my take, but again, I haven't thought it through carefully. Happy to hear important arguments on either side.

2:15:59.9 SC: Eugene Brevdo says, "Is it possible that if Mindscape AMAs ran for another billion years, you would one day stop saying yes to most questions that start with 'Is it possible that'?"

2:16:12.2 SC: Yes, it is possible that. Mindscape AMAs are not gonna run for another billion years. I don't think that's actually possible. So let me just say two things here about this ongoing joke that we have here on Mindscape, that the answer to "Is it possible" questions is always yes. First, of course there are loopholes. Of course there are questions that are purely about logic. Is it possible that, according to the axioms of Euclidean geometry, a triangle on a plane Euclidean two-dimensional surface has interior angles that add up to 183 degrees? No, that is not possible. That's a mathematical thing, right? We cannot actually wriggle our way out of that given those assumptions that were buried into the question. The "Is it possible that" questions are always referring to science questions, because in science we have to keep an open mind. But there's a reason why I keep telling the joke in that way, it's always, you always say yes to "Is it possible that" questions, because I think that people don't really mean "Is it possible that X or Y is true" in the literal sense of it being possible. I think that people have an impression, which I don't think is a good impression to have, that if something's possible, then it's plausible, or even worse, that if it's possible, then it's worth spending time thinking about, right? "Is it possible that this theory is on the right track?" Sure, it's possible. That doesn't mean you should spend any time thinking about that particular theory. So part of the program, part of the ambition of constantly saying yes, it's possible, is just to nudge people toward being better Bayesian reasoners and to graduate from thinking, "Well, it's possible, therefore it's worthwhile, to really thinking about how much credence they put on these different things." There's very, very different degrees of possibility. Something having a 10 to the minus 100 chance of being true and something having a 10 to the minus 1 chance of being true are both possible, but they are not equally deserving of your attention.

2:18:14.6 SC: Max Cant says, "I was pleasantly surprised to hear on one of the last AMAs that your favorite time-wasting mobile games are the Kingdom Rush series, as these are also pretty much the only ones I play. I'm curious what difficulty setting you prefer, since my favorite part of these kinds of games is playing on the hardest difficulty and slowly figuring out the right placement and sequence of towers to beat each level. It turns a silly mobile game into a surprisingly satisfying problem-solving experience. I'm wondering if you do something similar or prefer to play on an easier difficulty where you can more or less just shut your brain off."

2:18:49.4 SC: No, actually, I also play on the hardest difficulty, and I think that there's an interesting balance here. If the game were just something where I won every time without trying very hard, that wouldn't be that interesting. But also if it's too hard, if I have to put too much brain power into thinking about how to do it and practicing and doing many, many, many times before I beat it, that's also not what I want out of my silly little iPad game. So I think that usually Kingdom Rush gets it right. There are some levels, et cetera, that are harder than they should be and it's just annoying. There are plenty of ones that are easy, but that's okay. To mix in some easy ones and some hard ones is just about right. You don't want to put brain power at all into it every single time. But I do think that... I think that the later editions of the game have actually become too complicated. There's too many moving parts, too many towers and heroes and whatever. And I'm like, I don't want to think about all that. That's not really why I'm there. I want to do a little bit of thinking. That's why it's hard for me to complain too legitimately, because I do want to do some thinking, just not too much.

2:19:58.1 SC: Mark Kumari says, "After recently reading Brian Greene's The Hidden Reality, I've been thinking more about the different physical theories that imply a multiverse. I know you have a very high credence in many-worlds and a relatively low credence in Max Tegmark's Level 4 multiverse, where all mathematical structures correspond to a physical reality. I'm curious where you stand on the various other physical theories that lead to the multiverse. Since I cannot ask all that here, I'll simply ask your credence on the multiverse implied by eternal inflation. I know it is lower than many-worlds and greater than Level 4, but can you provide a bit more color on exactly how serious we should take eternal inflation?"

2:20:35.7 SC: Yeah, for eternal inflation, I'm just gonna totally weasel out here and put my credence at about 0.5, at about 50%, which is the least informative answer that I can give. If I said it was 0.1, that would say I really think it's probably not there. If I said 0.9, I probably would think that it is there. 0.5 says I just don't know. And that's because there are really good arguments on both sides, existence and non-existence. On the pro-existence side, we have the fact that inflation is very useful for accounting for the initial conditions we see in our universe, and inflation tends to typically be eternal in many models. That's a speculative question. People don't necessarily agree on that, but that is a very plausible reading of the whole thing. And out of that, you can get a multiverse that has explanatory power for things like the value of the vacuum energy in our universe, et cetera. So it's kind of a nice picture that you get. On the minus side, number one, it absolutely invokes completely ill-understood regimes of physics, energies and times and things like that that we have no direct evidence of. So we should be not very tied to any specific theories. We should be kind of open-minded. And secondly, there's real problems with eternal inflation. There's the measure problem, there's the predictability problem, there's the entropy problem of the early universe. I think that... The fact that... If you really truly believe the naive picture of the multiverse, it's very hard to say what it predicts because everything happens an infinite number of times. Everything that can happen at all happens an infinite number of times. And you might want to say, "Okay, but some things happen twice as often as other things." That's your intuition, but that turns out to be very hard to make careful and rigorous. It's just very hard to say that certain things happen twice as often as other things, and therefore it's very hard to make predictions. And I think that people who believe in inflation have mostly swept that under the rug and ignored it, and I think they should pay closer attention to it. In fact, just saying this out loud makes me think I should include that in the colloquium I'm planning to give soon that I mentioned other things I think physicists should think about. So I should include the cosmological measure problem as one of them.

2:22:55.5 SC: Grace Monk says, "Regarding your time in New Mexico, red, green, or Christmas?"

2:23:00.6 SC: Usually red, but sometimes Christmas just to mix things up. These, of course, if you don't know, these refer to colors of chili that you put on your meal. All sorts of meals can be covered with various kinds of chili sauce. And they can either be red chili, green chili, or Christmas means half and half, red and green. Get it? Christmas. So I went for the Christmas for a long time, but I think I do enjoy the red usually more than the green.

2:23:24.5 SC: Tatiana says, "What can we say about the interior of a black hole by studying its event horizon, taking into account the holographic principle, in particular by doing practical observations from Earth?"

2:23:37.1 SC: Well, I think that the cheap and correct answer is nothing, [chuckle] in the sense that if you mean... I should have grouped this together with the question about is there evidence for the multiverse, 'cause it's a very spiritually similar question. You can't see in the interior of a black hole. You cannot get data from the interior of a black hole by any observation that you can do practically here on Earth. Open and shut case. What you can do, though, is have a comprehensive theory that predicts something about the interior of a black hole, and it also predicts something observable about the exterior of the black hole. And then if you get data about that, you can infer the existence of this unobservable thing inside. Again, it's exactly what the logic is for the many-worlds interpretation of quantum mechanics. You cannot observe these other worlds. Why do you think they're there? Well, because they are predicted very strongly on the basis of equations that are observable. The equations, the Schrödinger equation, entanglement, decoherence, these are all observable things, and they predict the existence of other worlds. So I think that the usual story about the interior of a black hole is we need to understand the theory, and then we should be able to predict what happens with good confidence. But that theory means not just general relativity, but also quantum gravity and other interesting quantum things that we don't have a great handle on quite yet.

2:25:05.2 SC: David Kudaverdian says, "Could you please explain what superdeterminism means? I can't quite understand whether it's a standalone interpretation of quantum mechanics, and I can't wrap my head around the articles I've read. They immediately jump into Bell's theorem and describe the superdeterminism as a loophole in it."

2:25:23.4 SC: Right. So superdeterminism is an attempt to restore locality to quantum mechanics. And the reason why they always jump right into Bell's theorem is because Bell's theorem is the result that says that there is apparently violation of locality in quantum measurement. That when I measure something here, if it's entangled with something far away, there's an instantaneous change that has to be nonlocal. That's why they jump right into Bell's theorem. But superdeterminism itself is easy to state. Various results like Bell's theorem tell you what can and cannot happen on the basis of doing certain measurements, right? In quantum mechanics, you do measurements, you get results. And determinism is just the idea that, well, there is some initial state of the universe, and if in principle you knew the initial state, like if you're an Everettian, you could run the wave function of the universe through the Schrödinger equation and get it to predict everything that's gonna happen. That's determinism. Superdeterminism says, okay, that determinism, plus the initial conditions of the universe are really delicately set up so that certain observations never get made. [chuckle] So the reason why they jump right to Bell's theorem is that an assumption to Bell's theorem, or an assumption to make Bell's theorem relevant, is that I can do any observation I want. I can imagine measuring the spin along the x-axis or the y-axis or the z-axis, or any angle in between, et cetera, et cetera, right? And that's certainly allowed. But superdeterminism comes along and says, "Ah, but you are part of the physical universe. Your choices as to what experiments to do are secretly encoded in the initial conditions of the universe." And they take advantage of this loophole to say, "And the initial conditions of the universe are such that you think the Bell inequalities are being violated like they are in conventional quantum mechanics, even though strictly speaking, everything's perfectly local, you just can't notice it because you're never gonna do that experiment. To most people, this seems entirely wacky. But I can't use the wackiness argument myself because to many people, many worlds seems entirely wacky. The better argument for me is that many worlds just seems very simple and robust and what instantly falls into your lap once you understand entanglement. Superdeterminism seems like the opposite of that. It seems like a whole bunch of special pleading and crazy building of superstructure to avoid a result that is really just staring you in the face.

2:28:00.0 SC: Rob Atkerson says, "Do you think large organizations like governments and corporations could qualify as an abstract form of life? Do you think the ideas, ideologies, or narratives could qualify as life? And can you imagine any other abstracts that might fit this definition?"

2:28:16.9 SC: Well, if you go back to the podcast we did long ago with Stuart Bartlett, he and his friend... Was it Michael Wong? Was Michael Wong his collaborator? I think maybe he was, put forward this idea that they called lyfe, which I don't like the pronunciation of, but it's spelled L-Y-F-E. I very much like the spirit behind what they were trying to do, which is to say, forget about life. Forget about defining what is and is not life. It's very much in the spirit of the free will discussion. Forget about that. Talk about what are the characteristics that living systems usually have. They process information, they reproduce, they have homeostatic responses to external stimuli. They have a whole list, like seven different things that living life as we know it includes. And then they ask the question, could you imagine systems that have some of those features but not others? And the answer is very often, yes, you could. So rather than saying, is an organization qualifying as life, I would rather say, what are the various characteristics that life has? Which one of those are shared by large organizations? Some of them are, some of them are not. So I think that's... And I don't want to go through the list, but you can imagine going through the list yourself. I think that's a more intellectually satisfying question to ask. The thing I would say, though, is there is again a question of time scales, just like we bumped into when talking about why consciousness only sees one moment of time or one branch of the wave function. The brain, the human brain, is small enough and the electrochemical signals inside are rapid enough that the brain can act more or less coherently. Right? Your brain can do something all at once, roughly speaking. It doesn't take that long for the brain to get its act together and to organize all the different parts of the brain toward a common goal. Something like a large organization or a government, I don't think has that property. The different parts talk to each other, but they don't talk to each other fast enough and they don't organize fast enough to respond to the various things that happen to them. So that is one reason why I would not be overly tempted to identify an organization, a corporation, a government, whatever, as a living organism. I don't think it quite has that interactive coherence that I want my organisms to have. And this is likewise why when people say, "Could the galaxy be alive?" Or whatever. No. It's just things are too far apart. It's a very, very basic scientific limiting factor.

2:31:05.5 SC: S. Sanders says, "What is your guess for the most recent modern discovery in fundamental physics that could have in principle been discovered by the ancient Greeks? For example, given the technology they had access to, could the Greeks have discovered the wave properties of light if they'd been looking in the right direction?"

2:31:23.7 SC: No, I don't think so. The real difficulty there, maybe there's something clever that I can't think of very much, but they didn't have the ability... Sorry, let me back up. We had trouble discovering the wave properties of light in the modern world, right? Newton and Huygens completely disagreed about it. It really wasn't until the 19th century that we nailed that one down. And it was the double-slit experiment by Thomas Young. And the double-slit experiment in its classical form, forget about the quantum mechanical version, just the classical form, send a wave through two slits, they'll have an interference pattern on the other side. It only works if your wave is relatively coherent, right? It has the same amplitude and... Not the same amplitude, but the same wavelength and things like that. Setting up the double-slit experiment is harder than it looks. You could have made arguments like Newton and Huygens did. The ancient Greeks could have made similar arguments. They didn't have quite the technology. I don't know if they had the kind of glass that you could use to make a prism like Newton did, et cetera. But maybe they could have made arguments, but those arguments would not have been at all definitive. I think if you want to imagine what the Greeks could have done earlier, relativity is maybe a better guess, because they could have certainly invented the principle of relativity. Galileo did, right? And he was thinking about dropping cannonballs from moving ships in the harbor and stuff like that. And the Greeks knew all about that.

2:32:51.1 SC: So the Greeks absolutely could have invented the principle of relativity, conservation of momentum, things like that. Conservation of momentum didn't really come along until the Islamic Golden Age, like circa the year 1000. But the Greeks could have invented it; they just didn't. And if you have conservation of momentum, that is to say, in the absence of friction and air resistance and things like that, particles would move, or objects would move at a constant velocity, Newton's first law, then that plus relativity could get you the idea of relativity. It wouldn't give you special relativity because you don't know about the speed of light. You don't know about Maxwell's equations. You don't know there's any reason to think that the speed of light is constant in all frames. I'm not sure that they could have reasoned their way into thinking that. Could they have done something like the Michelson-Morley experiment? Probably not. The technology was not that great for the ancient Greeks, I gotta say. So I think that most of what we think of as "modern physics" really requires more technology than the Greeks had. They absolutely could have done classical mechanics; they could have done Newtonian mechanics. They just didn't have any reason to because they didn't have any experience experimentally with things moving at constant velocities, et cetera. Galileo showed them how you could do that with very little technology. So they could have done it, but they didn't really quite have the motivation for that.

2:34:15.7 SC: Michael Bailey says, "Why is entropy not considered a force? In my head, almost everything emerges from entropy, including the actual forces and gravity. Why is this the wrong way to think about it? Is this what came first, the chicken or the egg type of scenario?"

2:34:29.6 SC: Well, sadly, this is a theme of this month's podcast, but it depends on what you mean by force, right? It depends on the definition. It used to be, back in the day, forces were defined operationally as things that gave rise to accelerations, right? F=ma. Why is something accelerating? Well, because there's a force acting on it. Now, in modern quantum field theory, of course, this has slightly changed its definition because we had gravity and we had electromagnetism, and those were the two best understood forces, and those are both described by gauge theories in quantum mechanics or quantum field theory, or even classically, for that matter. And so when it came along that we invented the strong and weak nuclear forces and those were also defined by gauge theories, we started saying there are four forces. Because what we really meant is there are four gauge theories, and gravity is not the same kind of gauge theory as the other three, but it's close enough to lump it into that bucket. That's not exactly the sort of Newtonian definition of force. It's really a much more narrow conception. If you go back to that sort of old-school Newtonian definition, then there's absolutely other forces of nature that are very relevant. When I am placing my coffee cup on the table in front of me, the thing that holds the coffee cup up is not electromagnetism, much less gravity or the nuclear forces. It's the degeneracy pressure. It's the exclusion principle from the atomic orbitals in the matter that is making up the coffee cup and the table. That's not listed among one of the four forces, but it's absolutely important for thinking about forces in our everyday lives.

2:36:06.0 SC: The reason I'm saying all this is just to point out that there's a little bit of "it's up to you" when you say, like, "Does this count as a force or does that count as a force?" There is absolutely an idea called entropic forces, and they are, what is the right way to say it? They're distinguished from mechanical forces. I think I talk a little bit about this in the Complexity and Emergence book that will be coming out at the probably the beginning of 2027. A mechanical force is when you literally have some system that has a potential energy that depends on its size or location, right? Like a spring has a mechanical force, and if you stretch it, it'll want to stretch back. If you compress it, it'll want to push back. An entropic force doesn't have any mechanical pushing or pulling, but it's a system that has a favorite size based on its entropy. Like if instead of a spring, if you have a very, very low-mass chain, you have little links in a chain, but they're very low-mass. They could just lay on the floor and they would exert no force whatsoever. But then you heat up the chain, so now it has a temperature, okay? And now the links in the chain are bouncing around and you can calculate the entropy of any given configuration of the links in the chain. And if you try to squeeze the chain to too small a region, it pushes back because there's more entropy if it's extended. But if you pull it, it also pushes back, because if you pull the chain to be completely straight, that has fewer accessible states and therefore lower entropy. So that is the idea of an entropic force, and it's different than a mechanical force. But I don't think that we have much good reason to think that all forces are entropic. As far as we know, there really are mechanical forces out there in the world. Since we don't know the theory of everything, that might change. But as of our current best understandings of things like gravity, electromagnetism, et cetera, these are mechanical forces, not entropic ones.

2:38:08.1 SC: Ryan Patrick says, "I recently finished reading your book Quanta and Fields. If observations are what cause the wave function to collapse, then assuming the early universe began as only quantum fields, what caused those early quantum fields to collapse into particles with definitive states and properties?"

2:38:25.3 SC: Well, if you're really interested in collapse of the wave function and all the philosophical issues of the measurement problem, et cetera, Quanta and Fields was not the book to read. It is an important book to read, but you should also read Something Deeply Hidden to really get into what happens when a measurement occurs, et cetera. It has nothing to do with measurements. It has nothing to do with observations, with consciousness, with awareness, with intelligence, with agency or anything like that. It's all just physics. It's all ultimately just the Schrödinger equation and what becomes entangled with what. Now, in the early universe, it is some work to decide what counts as the system and what counts as the environment. If you're looking at Schrödinger's cat, it's pretty straightforward. The cat's the system, all the particles in the air and the light in the photons are the environment, and there's that clear-cut distinction so you can say when decoherence happens. There's an ongoing attempt to figure out what is the best way to divide the early universe into system and environment. I even wrote a paper about that with Kim Boddy and Jason Pollack. And how it affects eternal inflation. But ultimately it doesn't matter that much. We want to know what we're going to see at the end of the day, including in the microwave background or something like that. And the way the quantum mechanics works is when you attach words like wave function collapse or decoherence or whatever, doesn't matter as long as you're just solving the Schrödinger equation overall. You can wait until the very last second to say what is being seen by this observer in the universe, and you will always get the same answer even if you sort of used the vocabulary of decoherence at earlier times. So, roughly speaking, it's the rest of the universe. Any given one particle interacts with all the other fields in the universe, and it becomes useful to talk about them when they're becoming entangled with the rest of the universe. But the details also matter, and those are a little bit slippery in the early universe period.

2:40:30.9 SC: Bill McDonald asks about Jason Arday and says, "Should academic standards apply equally to everyone regardless of race?"

2:40:39.5 SC: Sure, they should. Yeah, that's an easy question. For those of you who don't know, Jason Arday was a Cambridge professor of education, maybe, or sociology. I don't know the details. I didn't follow the case very carefully. But he was accused of plagiarism, and he was pretty obviously guilty of plagiarism. And he was a Black guy. And so the accusations weren't just that he was a plagiarist, but that he was hired for his race rather than his scholarship, et cetera. And the right-wing sides of the British press hounded him to the point where he resigned from his position, and eventually he... Well, he died. It is thought that he committed suicide, but I'm not sure how clear that is. So it's kind of a tragic story in some ways. So on the one hand, it's very easy to say academic standards should apply equally to everyone regardless of race. It's also very easy to say that all sorts of standards are not applied equally to everyone regardless of race. There are cases, and I'm not going to go through them, you can look them up, there are cases here in the United States of prominent White academics who are very clearly committing research fraud and yet still have their jobs and their cushy gigs and things like that. So I don't think that the injustice is that somehow a Black guy got a job. I think that we don't apply standards of academic rectitude equally to people of different races, and that's kind of too bad.

2:42:16.9 SC: Anonymous says, "I'm starting my second year of a PhD in EHEP." I presume that's Experimental High Energy Physics. "In between my master's and PhD, I worked in industry for four years. I left industry because fundamental physics is my primary passion, and I've always wanted to pursue a career in academia HEP research. I recently received several job offers from recruiters, one of which is at a national research lab. It's very physics research focused, and if I were to leave academia, I could see it becoming a lifelong career that I'm content with. It also pays well, 175k per year. I want to finish my PhD and pursue a career studying fundamental physics, but as a broke grad student, this is very tempting. Additionally, I'm told that many people must spend six years in postdoc programs before they have a serious shot at a professorship, and many end up leaving academia anyway. I would really appreciate some perspective and advice. What would you say if one of your students came to you with this question, and did you ever consider leaving academia?"

2:43:15.5 SC: Yeah, I definitely considered leaving academia at various points in my career when it was not clear that I would have a successful job. One's employability as an academic goes up and down depending on what papers one has written recently, and that can be not a straight line upward or downward. So, happily, I was always in a position where, for the most part, not only did I have job offers, but I had more than one. So I could be pretty happy in taking the jobs that I was offered. Again, there's no way for me to answer this question for your particular situation. It depends a lot on things that I don't know about your career, what your research has been, what kind of things you're studying, et cetera. I'll say a bunch of things that are true, and you can take them into consideration as you go. Like I said before, most people who get PhDs, even from the very best places, are not able to become tenured professors. And you can just kind of obviously see that by the numbers game, right? In a physics department, many professors are graduating a student every year, but they don't retire every year. So what happens to those students? There's just not enough room for them. Philosophy departments try much harder to make sure that a large fraction of graduating PhD students do get jobs as professional philosophers, and as a result, we accept far, far fewer students into the graduate programs. But the reason why it's not a moral catastrophe for the physics departments to accept so many students is precisely because of situations like yours, where even if you don't get a job as a physics professor, you get some other job, typically that pays better than being a physics professor.

2:44:58.4 SC: And so I would say a couple things. It's hard to get a job. Don't be surprised if you don't as a tenured professor. Number two, it depends a lot on what kind of research you're doing. Is it the kind that departments are looking for in new professors, or is it something that is particular to you and you think is fascinating, but the rest of the world is not that interested in? What do you think your letters of recommendation are going to look like? Do you think that the people who are senior to you and going to be writing for you are really enthusiastic about your candidacy and will push you hard, or are they going along with it and trying to support you, but not actually really thinking that you're going to be a star someday? Also very important to think is how happy would you be in this other job? Maybe if it's not the absolute favorite job that you want, you could still make a good career out of it. And if you're making that kind of money to start, it sounds like you could be making a lot of money down the road. But on the other hand, on the fourth hand or whatever it is, if that job is there now, maybe a similar job is going to be there in a few years once you've gotten your PhD, right? And there's absolutely a case to be made for, if you do want to become a physics professor or professor of any kind, giving it your best shot and then seeing what happens. The great thing about a physics PhD is it does prepare you to do a whole bunch of different things, so you won't be out on the street at the end of it. So what I say to my own students is I can give you my feelings for probabilities, but the choice is ultimately yours. You have to do whatever it is your passion tells you to do. Like my very first graduate student who I graduated didn't even try to get academic jobs. He went right into consulting. Other people have tried for many years and struggled, and it's been hard. I've had a lot of luck recently with my students getting faculty jobs, but it comes and goes. The numbers are small and it's unpredictable, and academia as a whole, especially here in the United States, is not in a great position right now. So take all those things into consideration. All I can do is wish you luck.

2:47:07.7 SC: David asks a priority question. He says, "When using Bayes' theorem for updating your beliefs on a particular subject matter, whether it's are we a Boltzmann brain, is the many-worlds interpretation the correct interpretation of quantum mechanics, is God real? Et cetera, how do you frame those internally using Bayes' theorem? Do you have a specific probability that you handicap the subject matter in question and move the percentages in either direction every time you get a new piece of information to update your priors? Or is it more along the lines of, 'I strongly believe this is the correct interpretation,' and then you just adjust the descriptor as new information comes along?"

2:47:41.9 SC: Well, I think that you're supposed to think in terms of percentages, and ideally you would. It's actually notoriously difficult. Like, it's kind of not a simple calculation to be good about Bayes' theorem. What's relatively simple is to say this piece of information causes me to increase my credences in some things, decrease them in others. But the mathematical complication comes in because you have to make sure that all of your credences at the end of the day still add up to one, 100%, right? I mean, it's just a little bit of extra arithmetic. It's not terrible, but it's something that is tedious and we generally don't do it. And I certainly don't do it. So there's sort of a rolling update, I would say, as you're looking at more information and more facts about the world and thinking about how they all fit together, some credences go up, some go down. When I say things like being a good Bayesian, what I mean is think again about what we said before about string theory and the cosmological constant. It is too easy for people to say something like, "I think string theory is on the right track. I don't know whether string theory is compatible with a positive cosmological constant or not. Therefore, when we discovered the accelerating universe, it did not change my credences in string theory." I think that's kind of a mistake. I think that maybe it's permissible to have a positive cosmological constant in string theory, but I think in 1997, before we found the accelerating universe, it's not what you would have predicted. So I think it's okay and I think it's better to say, "My credences went down a little bit." Even if they were 99.99, they wouldn't have gone down that much. So maybe now they're 99.98, right? It doesn't affect your life in any way. It doesn't kick you out of the field or whatever. But you have to be honest. I think that Bayes' theorem is less a calculational necessity on a day-to-day basis and more a reality check and a way of keeping you honest about how information should change your beliefs about the world.

2:49:44.4 SC: Balachandran says, "I am struck by the number of interests I seem to share with you, a well-crafted cocktail, mechanical watches, reading a book with a cat purring softly on your lap. I have a hypothesis that there's a set of aesthetic preferences that correlate highly with interest in science and philosophy. This makes a testable prediction. Imagine designing your ideal car. I predict that it will have a dark forest green exterior with a supple dark saddle brown interior. How close was I?"

2:50:11.5 SC: I would say you're medium close. I'm a little skeptical of your hypothesis overall because I know a lot of people that have interests in both science and philosophy, indeed in almost exactly the same kinds of science and philosophy that I do, but who have very different aesthetic interests and preferences than I do. So I'm not sure the data back you up. I mean, given that we share interests in cocktails and watches and cats and books, maybe that extends to other aesthetic preferences, but I don't think that the science and philosophy thing is quite the right direction to point on. The car question is difficult because I don't really have an imaginary ideal car. I think I could imagine myself being happy in all sorts of different cars, to be perfectly honest. I did for a while, when we were living in LA, have a Jaguar XK8 convertible that I loved to pieces. It kind of fell apart there at the end, but it had over 100,000 miles. It wasn't like it fell apart very early. A Jaguar is not the kind of car that you're gonna drive for 500,000 miles. So it was at its natural lifespan point when I traded it in. But it was lovely. And Jaguar, of course, famously... The classic Jaguars often came in this Jaguar Racing Green color. So the idea of a green exterior convertible car is very appealing to me for that reason more than the fact that I like green in general. But green in general is good. Brown interior, I'm not so sure. I might go with the black interior. So that's a slight data point that should decrease your credence in your hypothesis just a little bit.

2:51:53.7 SC: Bran Muffin says, "To what extent can we think of the entire large-scale structure of the universe, including great walls and galaxy superclusters billions of light-years away, as it all exists now? If every moment is relative to the observer and there's no God's-eye view from outside the universe and we can only see past light-cone slices, does it even make sense to think that what I would be able to see with if I had a God's-eye view of the entire universe?"

2:52:20.0 SC: It does absolutely make sense to think of that. It doesn't make sense to assign some preference to one way of slicing the universe into moments of time. There's a slight preference 'cause cosmology does give you a rest frame of the universe, right? The frame at which you are at rest... Sorry, the frame with respect to which you're at rest to the cosmic microwave background and galaxies and things like that. But you can slice the universe into slices of constant time in different ways according to the rules of general relativity. So if you could imagine the God's-eye view, it would be a four-dimensional view. You'd be imagining the universe throughout all of space and all of time. You can't see it, right? You can't immediately, as a real non-God person, imagine experiencing that. But in science, all the time we take the observations that we do have and we construct pictures based on them. So there's nothing wrong with constructing a picture of what the universe looks like right now, even though we can't see it right now. We need to recognize that our picture might not be accurate, but that's also always true. That's not anything remarkably special about the universe or general relativity.

2:53:37.9 SC: David Harper says, "How do you document your research progress for personal usage? Do you create slides as you go, or write LaTeX documents, or write in notebooks by hand?"

2:53:47.8 SC: Yeah, I should have grouped this one with the thing about how I record the podcast 'cause, as usual, it's a mess. Things have changed for me since the invention of the Apple Pencil, okay? So there's a program... I experimented with different programs, I eventually landed on Notability is the one I use. And so Notability on my iPad is what I use for a lot of work in different ways. I use it to write my lecture notes for my classes. I use it to read PDF papers 'cause I can mark them up and take notes on them, as well as... You know, I have other documents. When I'm teaching a class, I have the photo roster of people in the class and things like that. So it's been super-duper useful. Taking notes of a good seminar or something like that. And then I can also file them, right? I can... You know, you can have the papers and my own notes in the same filing system and I can look at them and find them, which is kind of difficult. But that's just personal. For my papers that I write, most of my papers are written with my students and my collaborators, and we use a hodgepodge of different things. We are usually writing our LaTeX documents using Overleaf, which is the standard way of sharing LaTeX documents, which is the scientific typesetting system. I've also... I've recently, it took me a while since coming to Johns Hopkins, but I finally built up a good-sized research group. And so we're moving into the modern era where we use Slack to communicate like real businesses do. And we have a Zotero, which is a way of organizing papers, PDF files, and web pages that they come from. And again, you can organize them by topic and things like that. So hopefully this will supercharge our research progress in different ways. We'll see how that goes. But, again, for the most part, it's kind of haphazard and I have all sorts of different ways. I certainly don't create slides as I go. I try to create written documents when I have ideas that are half-baked. But again, sometimes that happens more effectively than others.

2:56:03.8 SC: Robert Ruxandrescu says, "Back in the 2000s, I started watching the NBA and playing NBA Live and fell in love with Vince Carter's game. He used to dunk on people left and right and have 30-foot game-winners and simply dominate basketball games in a way that I've seen no player do. Ever since then, I started training to dunk and still do more than 20 years later, and I saw over a thousand of his games. I'm wondering if there's a similar story for you with Dr. J. Did you like Dr. J for similar reasons? Did you follow his career? Can you relate to my story?"

2:56:34.9 SC: Yeah, I can relate to that story. I do think that one of the great things about basketball in particular, but also just sports in general, is that you root for a team, but you get to know the players, right? You get to know something about the individuals, especially in this day and age. Like, you have no idea what it was like in the '70s and '80s without the internet. It was just much harder to get information on what the team was doing, what the players were like, and things like that. Now they all have their Instagram accounts, which are super boring. I don't follow any athletes on their Instagram accounts. But yeah, I mean, Dr. J was purchased by the Philadelphia 76ers in 1976, I think, when the American Basketball Association folded into the NBA. And the New Jersey Nets, who I think were the New York Nets at the time and are now the Brooklyn Nets, had to raise money to join the NBA, and they did so by selling Dr. J to the Philadelphia 76ers. So I was around 10 years old at the time, just getting interested in sports, so that was one of the first news stories in sports that I noticed. And I liked basketball right away. It was fast-paced and fun and athletic, and the Sixers were good. If I had started a few years earlier, the Sixers had one of the worst, I think the worst seasons in the history of the NBA. So maybe my story would have been very different. But Dr. J was great for several reasons. I mean, he was a great ambassador for the game. This was universally recognized. He sort of personally kept the ABA afloat for several years, and it was a huge coup when the NBA finally got him. He was well-spoken and had relationships with the fans that were really, really good. He was not a jerk like some players are. But also he was a winner. As soon as he got to the Sixers, they went to the NBA Finals. And I think he was with the Sixers for what, 10 years between getting there and retiring, and they went to the finals four times and they won once, which is a pretty good track record overall for anyone. I like to point to 1980, when he took the Sixers to the finals and literally no one else on the team was an All-Star except for Dr. J. I think that his ability to impact winning is a little bit underappreciated by the modern basketball fans.

2:58:58.3 SC: It was a different time in terms of statistics, so he doesn't have the gaudy numbers that a lot of modern players have. And there were no three-point shots, but he did a little bit of everything. He played defense really well. He got blocks and steals and rebounds and assists in addition to scoring. So that was very... I liked that aspect of him. But then, of course, the final thing is it was beautiful. He had the most aesthetically pleasing basketball game of anyone I've ever seen until the present day. He would dunk, he would run, and he would jump very high. And he was one of those... Very much like Vince Carter, he was one of those players where it looked like he would jump and then instead of coming down, he would start accelerating and keep going up again. And you know that's an optical illusion. I know the laws of physics don't really allow for that, but it looked like that. And he would not only have the dunks, he would have all of these twisty moves and finger rolls and acrobatic plays that just made it very pleasurable. You knew that at least once a game, Dr. J would bust out something that just made the whole audience gasp. The Sixers got better attendance on the road than they did at home because everyone wanted to see Dr. J. And by the end of the game, the other team's fans would be rooting for the Sixers because the Sixers had Dr. J. And it was just a very good side to be on as a young basketball fan getting to know the game.

3:00:25.1 SC: Calvin Firth says, "I think I mostly understand and am on board with your moral justification for eating animals, the fact that they can't imagine and be sad about a future where they are dead. But I want to try to push against it a bit. Imagine if someone went out and painlessly murdered a bunch of kittens. These kittens aren't anyone's pet, and no other beings will be adversely affected by their deaths. Based on your moral reasoning, would the action of killing the kittens not be morally reprehensible? The more fundamental question I'm getting at is, is there any special inherent moral value to life?"

3:00:57.2 SC: So in my view, in my personal picture here, there's not any... The way to think about it is not by saying things like there's any special inherent moral value to life. And I think that your thought experiment actually indicates this very clearly because there are people who painlessly murder bunches of kittens, namely people who work at animal shelters where there's abandoned cats and dogs and they can't find homes for them. And those people are doing a morally good thing, right? They're saving these animals from really kind of miserable lives. And it's a shame that it has to happen. It would be better to imagine a hypothetical world where every cat found a good home. And I've done my part to help find kittens good homes, but there's too many kittens out there. And that's very different than if someone just murdered kittens, even if they did it painlessly, just because it gave them a thrill, right? Just because they kind of liked murdering kittens. That, I would judge as morally wrong in my own personal view, because that's a different thing. You're not doing it for some greater good. You're doing it for a not very good part of yourself. And so I think that it matters why you do it. And I think that all these are very complicated and I don't have strong feelings about any of them. Well, I have strong feelings, but I don't have strong convictions about my feelings. I'm willing to hear arguments on different sides. But we're all gonna die. Death happens to everybody. And so I don't think that death is the moral wrong. The reason why you force someone to die or you bring about the death of a person or an animal matters a lot. So that's why I say from the start, the right way to think about it is not, "Oh, there is special inherent moral value to life that we have to preserve." I think that's not dissimilar from the right answer, but it is not the best possible way to think about it.

3:03:02.2 SC: Melo Honek says, "In the concept of a quantum cyclical universe, does each iteration consist of the same rules and constants?"

3:03:14.1 SC: Well, different people have different models. So the recent model that I proposed with Sakshi Dulani and Nadia Diachenko is different than everybody else's model in a very specific way. Like most people who have cyclic models of cosmology, they're not periodic models. So they go through a cycle, but each cycle is a little bit different. And it depends on the model whether the rules are the same and the constants are the same. I think that in the space of actual models that actual people have actually written down, usually the rules and the constants are the same in a cyclical universe cosmology. I can imagine examples where they wouldn't be. I can absolutely imagine theories where that wouldn't be the case. But you have to ask yourself, why are you inventing this? Like, what is your motivation for trying to construct such a model? We had a very clear motivation for constructing our model, which is we wanted to know whether or not you could have a quantum cosmology based on a vector evolving. Unitarily, smoothly, in Hilbert space, in a finite-dimensional Hilbert space, without running into a Boltzmann brain problem. And we found a very specific, delicately chosen loophole that allowed us to do that. So we had a question we wanted the answer to, and we found it. I think that, roughly speaking, cyclical models are often thought of as alternatives to multiverse kind of models. And therefore, people who like cyclical models generally are not trying to make the rules and constants different from cycle to cycle.

3:04:48.1 SC: John Eastman says, "In your Dark Matter, Dark Energy Great Courses lecture four, you mentioned that it is possible to assume all length scales are shrinking rather than space expanding, although you say making such an assumption is silly. I was wondering if one could test the hypothesis with a tabletop experiment comprising a pair of charged masses on a frictionless track such that their gravitational attraction is balanced by their electrical repulsion."

3:05:14.1 SC: No, one cannot test the hypothesis because it is not a hypothesis. There's no science in saying that you can assume all length scales are shrinking rather than space expanding. Those are simply two different descriptions of exactly the same physical theory. It's like saying taking two numbers and adding them together using base 10 versus using base 2. You're gonna get the same numerical answer for a quantity. It might be expressed typographically differently, but the number you get by doing addition is the same no matter what base you use. And that's exactly the same thing if you use coordinates where space is expanding versus things shrinking. So there's no possible experimental test you can do because they're not actually two different hypotheses.

3:06:02.3 SC: Julio Cattaneo says, "Do we have any theory on how dark matter interacts with a black hole? I'm guessing that even if dark matter falls into a black hole or is around a black hole, we don't know that it is there."

3:06:14.3 SC: Well, the theory... Yes, we do have a theory. The theory is that dark matter interacts with a black hole exactly the same way that ordinary matter does. Dark matter... Well, let's go backwards. One of the great things about gravity is that everything interacts with it in exactly the same way. In particular, different particles might move on different trajectories if they have different speeds. But any two particles, if they're starting with the same initial conditions, they can be made of different stuff, but they will move in exactly the same way. That's the principle of equivalence. So dark matter particles move in exactly the same way as ordinary matter particles do. Now, once you go from one particle to many, the story changes by a lot. The thing about black holes are they're very small, right? The black hole at the center of our galaxy might be big to most of us by our everyday standards. It might be a light-year across, but compared to the size of the galaxy, that's really, really tiny. And even if it's a million times the mass of the sun, the galaxy is 100 billion times the mass of the sun. So it's tiny even by mass standards. So what that means is that even if there's a lot of dark matter in the galaxy, very little of it ever comes close to the black hole, much less falls in. So therefore, there's not a lot of relevance to black holes from the dark matter. The dark matter basically mostly is far away from the black hole and stays that way.

3:07:45.9 SC: With ordinary matter, there's a huge difference, which is that ordinary matter feels the force of electromagnetism, and therefore two atoms or two molecules or even just two particles can bump into each other and lose energy by emitting a photon, the particle of electromagnetism. And that is a source of dissipation. That means losing energy. And when you lose energy, if you're a particle in a gravitational potential, you fall toward the center. So the thing is that all that dissipation that ordinary matter has allows the ordinary matter to move in the direction of the black hole and to fall into it and to increase its size. As far as we can tell, and there's actually good constraints on this, dark matter doesn't have any dissipation. Dark matter doesn't interact with electromagnetism by the nature that it's dark, okay? Therefore, it doesn't lose energy, it doesn't fall into black holes. Now, I did once write a paper with some collaborators called Dark Matter and Dark Radiation, where we imagined a theory where dark matter has its own version of electromagnetism. And that might change the story a little bit, but only a little bit, because to make the story work, the dark matter needed to be very heavy and heavy enough that it wouldn't actually bump into other dark matter that much and lose energy. So we think that most of what goes into making a black hole is ordinary matter, just because it's easier to squeeze ordinary matter into small regions, whereas dark matter stays large and puffy.

3:09:20.3 SC: Sean Bentley says, "My 14-year-old would like to know why increased motion through space causes decreased motion through time. I could get as far as explaining that it's a feature of spacetime geometry per special relativity. But beyond that, I thought I'd consult an expert. Is there anything more fundamental to say, or is this a brute fact, or do we simply not know?"

3:09:39.7 SC: Well, there's an underlying brute fact, but there's different ways of stating the brute fact. And this is a general feature of brute facts. When you say that some chain of explanation bottoms out in brute facts, it's often the case that you're allowed to choose different brute facts as your explanations. It's exactly like saying a given axiomatic system in geometry or logic or number theory, you can replace some axioms for others and get exactly the same system overall. So for something like spacetime... Think of it this way. Let's just start with the idea that spacetime has a geometry. And let's ignore general relativity and complications of curvature of geometry. Let's imagine the geometry is flat. Let's make our lives easy. In that case, every dimension of spacetime comes with two options, either the spacetime interval along it is positive or it's negative, or it's zero. But usually we choose things so that the zero direction is some combination of positive and negative, and the fundamental directions like X, Y, Z, and T all have a positive sign associated with them or a negative sign. So the brute fact is the distribution of positive and negative signs in the spacetime interval that is appropriate to spacetime. In Euclidean geometry, all the spacetime intervals are positive, and therefore the shortest distance between any two points is a straight line, because by deviating in any direction, you add more and more to your spacetime interval. It so happens the brute fact is that we live in a spacetime where what we call time has a negative number in front of it for the spacetime interval, and all the other spatial directions have a positive number. Now, in fact, that's a convention. You could have done it the other way. You could have given the positive number to time and the negative number to the spatial directions. But they have to be opposite. That's the important thing. And because they're the opposite, you have an option of moving in space or moving in time. And if you move a little bit in space but come back to where you started, you will always experience less time. The deviation in space adds the opposite amount of spacetime interval than the deviation in time. And that's the brute fact that you have to deal with. And you're always allowed to ask, could it have been different? Are there reasons why? People have tried to have more than one dimension of time, have negative spacetime intervals with more than one dimension. It generally doesn't work. Generally bad things happen and most people are not encouraged. But there's always some people who are really into the idea and they push it, and so they're thinking about things like that. But no reason to get too excited about it just right now.

3:12:36.4 SC: Miran Mizrahi says, "In your recent podcast with Vitor Cardoso, he said, 'Whereas black holes, because gravitational collapse acts so efficiently, it just cleans them of any detail or of any other detail. They're really just simple curled up vacuum.' Could you elaborate on this? Where does the matter of the collapsing star go? Is it all the energy converted to gravitational energy and pressure, which then drives the curvature so that light can no longer escape?"

3:13:06.5 SC: I encourage people to check out the books I've written about things like this, such as Space, Time and Motion, or even, I think, back in From Eternity to Here, I talked about this thing. The thing is, the matter doesn't go anywhere in some very real sense up until the point where it hits the singularity. When a star collapses to make a black hole, you can think of at any moment of time, if you think about four-dimensional spacetime and then slicing it into different moments of constant time, that slice will always hit the star. The star will get smaller and smaller, but it will never hit zero size at any one moment of time in the usual way of slicing spacetime into moments of time. Now, if you fall into the black hole and you say, "I want to go hit the star," you might not be able to do that. In fact, very, very rapidly, you will reach a point where you're too late and you're gonna hit the singularity, which is in your future before you hit the star. So there's no way to measure how big the star is in some very real sense, but maybe you can see the light emitted by it or something like that. But nothing magical happens inside the black hole. I think that when someone like Vitor says gravitational collapse acts so efficiently, it just, it's nothing but simple curled up vacuum, there's two things going on. One is when we first study black hole spacetimes in our general relativity class, it's too complicated to think about a real star collapsing. We just think about an eternal black hole that was always there and has no matter associated with it. This is the eternal Schwarzschild vacuum solution. That's perfectly legitimate. It's not representing the real world. In the real world, black holes come about because of stars collapsing.

3:14:55.3 SC: The eternal Schwarzschild solution has all sorts of weird properties. It has not only a black hole, but it has a white hole in the past out of which things can come but nothing can go in. And it also has a wormhole connecting a whole other copy of the universe. One asymptotically flat spacetime is connected to another one by the Schwarzschild wormhole, which was invented by Einstein, Podolsky, and Rosen in their EPR paper... Sorry, not Einstein, Podolsky, and Rosen, Einstein and Rosen without Podolsky in their ER paper, the Einstein-Rosen bridge, which is where this wormhole idea came from. But the Einstein-Rosen bridge and the white hole are not thought to really exist in the world, but we study them because there's a simple mathematical setup to look at first. And the other thing going on is that even if you do know that in the real world black holes are made by collapsing matter, the exterior of the black hole is the same in the version where it was made by collapsing matter as it was in the simple mathematical idealization without any matter. So you can start thinking of it in exactly the same way. That was why it's worth your time to study the simple Schwarzschild black hole without any matter in it. The point is that the Schwarzschild solution lets you describe a black hole without any matter creating it, but we think that real-world black holes are created by matter, but the end result looks exactly the same either way. So I think that's what's going on in that description.

3:16:32.9 SC: Bo Parrizo says, "In a recent podcast, the guest used the term 'fractal-like' in the same way that I'm used to hearing you use 'scale-free.' Is there anything interesting to say about the difference?"

3:16:44.6 SC: Probably not. They're very closely related. I would say that fractals are scale-free in some sense. That's one way of saying what fractal is supposed to mean. I think that fractal has a more narrow definition. A fractal is a kind of a geometric figure, right? It is something that has a shape or something like that. The coast of Norway is fractal at some resolution, but there's all these curves you can construct mathematically that are infinitely fractal all the way down to zero size. And so they're scale-free in the sense that if you take a fractal and you zoom in on it, the statistical properties of the little subset that you're looking at are the same as the statistical properties of the fractal as a whole. That's what it means to be scale-free. But if I say that the distribution of, I don't know, city populations follows a power law, power laws are scale-free. There's no preferred size, there's just a relative size. If you're twice as big, then you appear half as often or something like that. So it's a more general notion. So fractals are a subset of scale-free behavior, I guess I would say. And I don't want people necessarily thinking of some pointy, sharp geometric figure when I say that some relationship is a power law and therefore scale-free.

3:18:09.7 SC: Anonymous asks a priority question. "Some people have suggested that time may not really exist." Oh no, what kind of people would do that?" After reading From Eternity to Here and listening to Mindscape, it is relatively easy to understand how the arrow of time may be an artifact and time passing may be an illusion created by our brains. But the fact remains, it takes different times for things to happen. Have you any idea how people who suggest time may not exist explain that other than time does exist?"

3:18:39.5 SC: Yeah, again, I should have grouped this with the earlier one, but the point is time has to emerge. The people who say time doesn't exist should, if they're intellectually consistent, say that tables and chairs don't exist either. And there are people who say that, and those people deserve your pity, not your contempt. It's just hard to go through life not believing in tables and chairs or the existence of time, okay? You can think that time is not fundamental, but chairs are not fundamental either. That's okay, they still exist, right? Again, just like the ability of human beings to make choices, whatever you want to call it, the ability to travel into the future by letting time pass is something that is indubitably real. It definitely exists. So I think that people who say that time doesn't really exist are literally just thinking or using words in sloppy ways, let's put it that way. Just tell them to show up at some time and see how well they do. If they're able to show up at the right time, they must be fibbing to you when they say that time doesn't exist.

3:19:42.6 SC: Lishan Aklog says, "In the last AMA, you made the perhaps unintentionally provocative statement that what we call the electron neutrino and what we call the electron are basically the same, with their differences arising from symmetry breaking. This felt like one of the 'Professor Sean lifts the fog' conceptual breakthroughs I've experienced decades after graduating with a physics degree. I'd appreciate some elaboration. Is this related to the question of whether related particles such as electrons and positrons are not fundamentally separate entities, but different excitations of a common underlying field?"

3:20:22.0 SC: Yeah, it's related to that, but it's also a little bit different than that. So, I mean, electrons and neutrinos are not the same story as electrons and positrons, okay? But in both cases, you have different components of some kind of underlying field. The difference is with the electrons and the neutrinos, if you think about before the symmetry is broken, so you think about before electroweak symmetry breaking, before the Higgs boson got a non-zero value in empty space. The Higgs boson in the sufficiently early universe, the Higgs field had an expectation value of zero. So it was sitting there, not doing anything. And in that case, there is a two-dimensional field which you could call the electron/neutrino field. And it's like... It's literally, think of a plane, right, with two directions on it. And there's no difference between going in one direction and going in the other direction. After the symmetry is broken, what the Higgs field does is basically, it makes a valley. So if you're... If you have some physics background, imagine that your plane is given coordinates X and Y. And there's no intrinsic difference between X and Y. You could rotate your coordinates. It just doesn't matter. There's no such thing as which is the electron, which is the neutrino. It was just two directions somewhere in the plane. But then someone adds a potential energy to the plane. And there is a coordinate system, X and Y, such that the potential energy is Y squared, which means that if I move in the X direction, there's no potential at all. But suddenly, if I move in the Y direction, now it costs me potential energy. And that X direction, ex post facto, is what we call the neutrino. It's nearly massless, no potential. And the Y direction is what we call the electron. It costs energy to move in that direction, which we interpret as the mass of the electron. And the previous rotational symmetry that we had in the plane has been broken by the existence of that potential energy. And this is exactly, this is not even an analogy. This is just literally what the Higgs expectation value does. It treats one direction differently than the other one. And there was no inherent choice as to which direction was going to be treated differently. In fact, it's a gauge symmetry. So there's literally no fact of the matter about which direction was going to be treated differently. It's just a coordinate difference Whereas electrons and positrons, those are antiparticles, and those were always antiparticles, and those were always different. They're, again, part of the same Dirac field, if you want to think about it that way, but they were always pretty... The fact that they have opposite charges was always true, let's put it that way.

3:23:11.5 SC: Armchair epitomologist. I don't think this is a real word. It's not epidemiologist or epistemologist. Maybe the S is missing. It's epistemologist without an S, epitomologist. "Taking the simplistic model that human civilization's technological advancement is linear, something like the Kardashev scale, what sort of general credences do you put on our civilization progressing technologically versus regressing in the next thousand years, 10k years, 100k years, et cetera? I know it's not a fully specified question and obviously impossible to know, but I'm curious of your general degree of belief in our civilization to achieve these sci-fi levels of technological advancement before we're wiped out or sent back to the Stone Age by ourselves or some external factors, or maybe we just get stuck at around our current stage." I think this is fun stuff to think about. I think that we don't have any right to be especially confident in our thoughts about this. And back when we were talking about how do you detect the crackpots, acting super confident about questions like this is a kind of a signal of crackpottery right there, because we don't know. It's like acting like you know what happened at the Big Bang. Like, you might have theories and you should take them seriously, but you shouldn't be too wedded to them. I think two things are simultaneously true that make it very, very difficult to answer this question. One thing is that the advance of technology has given human beings powers that they didn't have before. You know, back in the days of 100,000 years ago, there weren't that many human beings on the earth compared to how many there are now, but the ability of any one human being to do damage to the entire rest of humanity was negligible, right? Like, how many other people could you hurt before they stopped you? Not that many. The ability of technology has given individual human beings an outsized potential influence on the rest of the world. We obviously now as a species, we can change the whole environment of the planet, right? Like, we're doing it. It's very noticeable.

3:25:22.8 SC: But also between nuclear weapons and biological weapons and various internet-based technological things, there's a whole bunch of things that human beings can do that have an outsized impact on the world in which we live. So that raises the risk level of what humanity can do to itself in an unprecedented way. That's one thought that makes it difficult to predict the future. The other side has the fact that complex systems tend to be resilient. Like an individual complex system, like a human being, is not that difficult to kill. Chop off a head, usually you kill it. But complex systems as a whole that are distributed in all sorts of different ways, very, very difficult to completely wipe out. In fact, it's hard. Despite the fact that we have nuclear weapons and all this stuff, it's actually not easy to come up with a scenario where human beings, even if we tried, could literally kill off all the human beings on Earth. We can imagine creating diseases that kill off a lot of people, or having a nuclear war that kills an enormous amount of people, like really tremendous tragedies. But the chances that some human beings would survive are still quite high even in those relatively extreme situations. And so both of these are very fuzzy, right? On the one hand, we're saying we can have a huge impact on the planet. On the other, we're saying we're pretty plucky. And neither one of those is anywhere close to specific enough to make predictions for what's gonna happen the next thousand years, 10,000 years, et cetera.

3:27:03.8 SC: I will also say that human beings are terrible at timescales. We're good at imagining things. We're very bad at imagining how long something will take. So I think that the fact that we can build rocket ships and things like that now has caused some people to think about going to Mars or going to the moon or whatever, colonizing the solar system. It's pretty easy to write science fiction novels about this. It's really hard to do it in practice. We had the podcast with Zach and Kelly Wienersmith where they explained some of the obstacles to that in very clear terms. And the thing is, the people who tend to be most interested in these questions also tend to be most invested in making it happen, so they're not always the most realistic. So in 10,000 years from now, I would not be at all surprised if there were plenty of human beings living off of the Earth somewhere. 100 years from now, I would be very surprised if there were a lot of human beings living off the Earth. There'll be a few. There are a few right now, if you count living in the space station. But timescales are hard. And think about what you would have said about the future a thousand years ago and how wrong you would have been. Well, we'll probably be just as wrong if we try to predict what the future is like a thousand years from now. So I'm optimistic about the ability of humanity to adapt. I am pessimistic about our ability to say what that would actually mean in terms of what things will be like even a thousand years from now, much less 100,000.

3:28:41.4 SC: Jeroen says, "Can you shed light on the relation between symmetries and the laws of physics? Is it just an aspect of the math, or are there philosophical reasons why a broken symmetry must mean there is a particle in the universe?"

3:28:54.6 SC: So I got confused. I thought I was following you there, Jeroen, in the first half of the question, symmetries and the laws of physics. But why a broken symmetry must mean there is a particle in the universe, I'm not quite sure what that is pointing at. I'm gonna guess that what you mean is the Higgs phenomenon, which we were just talking about a second ago. The Higgs mechanism is a very specific kind of symmetry breaking. It's not the general idea of symmetry breaking, which is quite general and powerful, nor is it the general idea of symmetries, full stop. The Higgs field is... The Higgs phenomenon, I should say, the Higgs mechanism is a particular way to break a symmetry because a field takes on a non-zero value in empty space. And this goes back to Brout and Englert and Higgs and others. Philip Anderson sort of had a premonition of it beforehand. You can break a symmetry by a field having a value. It's exactly like this. Imagine you have a circle and you say, well, all the... Or a sphere is even better. Imagine you have a two-dimensional sphere, the surface of a ball, and it's perfectly smooth and there's no distinguishing features on it. It's perfectly symmetric. But you pick randomly a point to call the North Pole and the South Pole that you're gonna rotate it around. You have broken the symmetry, by picking out those preferred points on the sphere. They weren't picked out as preferred ahead of time. So you need something to break the symmetry, and the Higgs field is what breaks it in quantum field theory, in the Standard Model of particle physics.

3:30:33.8 SC: And so it's not that breaking a symmetry means there must be a particle. It's that this particular symmetry breaking mechanism involves a quantum field taking on a non-zero value. And whenever you have a field, you have a particle, because particles are just vibrations in fields. Now, there's a deeper question that I thought you were getting into in the first part of your question, why are the laws of physics as we know them full of symmetries? That's maybe even too deep for me to really give a sensible answer to. I certainly think it's possible that the symmetries that we see in the laws of physics are kind of an artifact. They're kind of a reflection of how we as human beings find it easiest to describe what is going on. Just like, again, with the sphere, maybe we have different directions of space and they're all equal in some way, and so we invent a mathematical representation of this where there's a representation that acts like a rotation and gives us a different view, but it's the same underlying stuff, so we call it a symmetry. Okay. So whether or not the laws of physics really love symmetry very much, or whether or not we human beings love symmetry very much because it helps us describe the laws of physics, I think is a good question to which I don't have a perfect answer right now.

3:31:54.2 SC: Carmaria says, "I'm sad that LeBron James didn't join the Warriors. I take comfort that we have Steph Curry, and one could argue that in basketball, philosophy, and physics, and other endeavors, one's ability to change the game may be more important than stats. Do you agree that Steph Curry is more of a game changer than LeBron? And do you think that the greatest strides in physics and philosophy are made by game changers or those who have great stats, e.g., they publish a lot?"

3:32:22.6 SC: Well, Carmaria, I think that the analogy is a little bit strained, actually. I don't think that the changing the game versus publishing a lot is really quite the right distinction to draw. Look, I've said this before. I'm not a believer that you can rank basketball players or athletes or scientists, for that matter, in some objective criteria of excellence once and for all. I think that you can look at what they did. You could look at the championships they've won, the statistics they've accumulated, the ways in which they've helped their teams and things like that. I think that what you're getting at is LeBron James, as amazing as he is, the things he's done are kind of things that other people have done in the past. He's just doing them better. Whereas Steph Curry really did something that no one else had ever done before in his particular way of playing the game and changing the game. Just being able to shoot so quickly, being able to move around the court so quickly that he could get open to shoot even though he's not especially tall and everyone knew that he was gonna shoot. And to make these wild shots from a long way away at absurdly high efficiency was a different kind of excellence than someone like LeBron James has. I mean, LeBron in many ways is somewhere in between Michael Jordan and Magic Johnson, right? He has the scoring chops, but he also has the passing chops. He's a tall point guard kind of player like Magic was, but a very efficient scoring machine like Michael Jordan was. But those are predecessors, right? It's harder to find an obvious predecessor for Steph Curry. I think they're both important. I don't think that greater strides are made by one or the other. I think at the end of the day, it's who wins the game. At the end of the day, it's what discoveries were made by what scientist and how they went about doing it. The greatness or the lack of greatness is less interesting to me than what was actually accomplished.

3:34:34.4 SC: Marie Raskue says, "I saw you on the Lives Well Lived podcast earlier in August and it was quite an unusual one with those questions on morality of the many worlds, among other things. I was wondering if we should even ask such questions, like the morality of many worlds interpretation, or any theory of physics for that matter. How did you feel the talk went? Was it time well spent?"

3:34:55.7 SC: Yes, you might remember that we had Peter Singer here on the Mindscape podcast. And Peter has his own podcast called Lives Well Lived and he invited me on, so I appeared on that. And I think it went pretty well from my perspective. It was a weird mixture because Peter Singer doesn't care about the philosophy of physics. He cares about ethics and morality more than anything else. And so he kind of wanted to talk about that, but he had me on, so he also talked about physics. He asked me some questions about many worlds. And he has a... I'm sorry, I should look this up. I'm not gonna look it up right now because I'm in the middle of doing the AMA. But he has a co-podcaster, I'm sorry, I forgot her name, but she did as much talking and we had a three-way conversation. It was great. What I guess I was happiest about was that he didn't point out any obvious mistakes to me in my own moral philosophy, even though we completely disagree. I think that when we got to talking about moral philosophy, even though we had different takes on it, I still, coming out of it, agreed with my own reasoning more than his, right? Sometimes when you talk to someone really smart who you know disagrees with you, you have this crisis of confidence moment where you start thinking like, "Oh my God, what if they're right? What if their arguments... I haven't thought about that one. They clearly know this better than I do." So I didn't have that with Peter Singer. That could be a flaw in me. Maybe the better arguments were there and I've just lost the ability to hear them, so entrenched am I in my worldview. But I don't have that impression coming out. But I do think that I was happy with how I did defend my own views, and hopefully some people got something out of the conversation, which I thought was indeed time well spent.

3:36:41.3 SC: Jeremy Thornton says, "My interpretation of your description of emergence makes me worry for the long-term mental health of Laplace's demon. Am I right to do so? My take is that they would either have no necessity for considering such constructions as trees, cats, and humans useful, so the universe would seem entirely mundane, or they would find in any arbitrary collection of matter equally useful and become befuddled by the variation. Should we set up a GoFundMe page in anticipation of their future needs?"

3:37:12.6 SC: Well, the good news is Laplace's demon doesn't care about those things, so their mental health is not in any danger. But your intuition is correct. Laplace's demon... There's different versions of the thought experiment, but you can certainly imagine a completely self-consistent and self-contained version of Laplace's demon that never knows the notions of trees or cats or humans or entropy or temperature or classical physics or anything like that. They just solve the wave function of the universe once and for all and that's it. They have an exact description of the entire world and they don't need to think about any of the emergent levels. That's completely allowed. But Laplace's demon doesn't exist, so don't set up any GoFundMe pages. If you do see any out there, they're probably just scams.

3:38:04.5 SC: Alex Dubrow says, "In your recent conversation with Chandra Sripada, you said his arguments shifted your credences about whether LLMs have rediscovered some of the ways humans think. I'm curious where your credences are now. Given your views on emergence and levels of description, to what extent do you think that today's most capable LLMs exhibit intelligence or cognition in something like the same sense that humans do? At the emergent level where we actually recognize and evaluate intelligence, do you now think differences in substrate and lower-level mechanism become less important than the higher-level regularities that the two systems share? You mentioned convergent evolution. Might the similarities arise because prediction has shaped both systems, humans through direct interaction with the world as well as culture and language, and LLMs indirectly through the enormous cultural record humans have produced from that interaction?"

3:38:56.8 SC: Well, the thing at the end about prediction shaping both systems is exactly Chandra's point. That's what he wants to hypothesize is actually happening here. But I want to separate this issue from the issue of silicon versus biology, okay? That may or may not become important to questions of consciousness, and we'll get to that in the very next question. But cognition was what Chandra was thinking about. And cognition is much simpler in some sense. It's just reasoning, just thinking, just adding numbers together, making decisions, things like that. It's not conscious experience or any such thing.

3:39:34.4 SC: So the question with LLMs always was the following. They're not built, they're not architected or designed to mimic the actual structures in the human brain, okay? They don't have a cerebellum and frontal cortex and all that stuff. Their architecture is inspired by the idea of neurons and how they talk to each other, but they're not wired together in some hierarchical way relative to what the human brain is doing. Instead, they are trained to give outputs based on inputs that resemble what human beings would do. And so the open question always is, did LLMs manage to accurately mimic what human beings would do by rediscovering the way human beings do them, or by discovering a new way to do really similar things? And there's good arguments on both sides, right? Some arguments for the similarities come from just the similarities of the inputs and the outputs. Some arguments for the differences come from the fact that when LLMs do make mistakes, there's sort of a characteristic kind of mistake that they make that human beings wouldn't make, that indicates that they're reasoning differently in some ways. And so I think it's an interesting open question. And I've often said this, and people said, "Oh no, they clearly have rediscovered human ways of thinking," but all of their evidence was just... They sound human at the end of the day. That's what it always came down to, or they're really smart. But it was the kind of smartness that was under question, not the existence of their smartness. And I thought that Chandra gave a good job of at least making an argument that we have evidence that the way in which they're thinking is closely related to that of humans. Now, others strongly disagree with this, right? This is not a consensus view by any stretch. There is no consensus view about this, but I was glad to see some real evidence put forward. I think that it's just the beginning of us understanding this. I don't think that we're close to figuring out once and for all the similarities or differences between humans and LLMs.

3:41:44.9 SC: As Chandra readily admits, you need a lot more input, you need a lot more tokens to train your LLM to give those human-like responses. And that's kind of obvious why. It's namely because human brains have structure in them that LLMs are not born with. So that's an important difference. So I don't think it's... All of that discussion is completely aside from the question of silicon versus biology or anything like that. But it's very, very interesting to imagine that the way that we human beings go about cogitating, as optimized by natural selection and the ability, the desire to make accurate predictions, turns out via convergent evolution to be similar to the way that LLMs go about reasoning. I don't have strong feelings one way or the other, but I'm absolutely open to the evidence pushing us in one direction or the other.

3:42:40.0 SC: Then Josh says that, "Chandra said you used to lean functionalist about consciousness, but you might have drifted away from it. Where do you actually stand now? Did his mechanistic case pull you back at all, or are you still leaning towards something like Anil Seth's biological naturalism?"

3:42:56.6 SC: So I'm on neither side. So number one, Chandra was not talking about consciousness. We were not talking about consciousness in that podcast. We were talking about cognition, which is something different. And number two, Anil Seth and also Ned Block, we recently had on the podcast, have made cases against computational functionalism as a way of thinking about consciousness without giving up on physicalism about consciousness. So again, I'm not pretending to be an expert on consciousness. I really promise you I am not. And part of my non-expertness was that I never quite appreciated that you could be a physicalist without being a computational functionalist. But there is a leap involved, and I think that Anil and Ned are both completely correct about this, that just because at the end of the day there's physical stuff moving around in different ways, obeying the laws of physics, that's not enough to make you instantly leap to thinking that whatever consciousness is, it's a matter of functional processing of information. There might be more to it than that, even if the more is still 100% physical, not weird, spooky, ineffable, spiritual stuff, right? So I think that's all. I think that's the only change that has happened in my conception of consciousness is that I have a greater appreciation for the difference conceptually in saying that you're a physicalist and saying you're a computational functionalist. So I'm open to the possibility that there are things going on which don't map on in any direct way to computation, but are going on in the human brain or human biology, just in the metabolic processes, the entropic gradients, whatever is going on in the brain that doesn't come across in a computational description, but is nevertheless important for consciousness. I'm open to that possibility. I'm very open to lots of things when it comes to consciousness as long as they're compatible with the laws of physics.

3:45:04.1 SC: Brett Slog says, "What do you make of the growing movement to oppose the construction of data centers? There seems to be some elements of a moral panic, particularly around the idea they consume large amounts of water, mixed with some semi-legitimate NIMBY-esque objections about noise and aesthetics, all clouded by general anxiety about the future of AI."

3:45:24.9 SC: Yeah, I think that your reading is pretty similar to my own. I don't have strong opinions here one way or the other. It reminds me of a question we got a couple months ago in AMA about my feeling about the lab leak hypothesis for the origin of COVID-19. And I said, look, I don't care in some ways. That's not the case with data centers. For data centers, I care. For COVID, I didn't care. I don't care if it's natural origins or synthetic origins. What I care about is being honest and scientific about getting the answer. And it became very difficult to do that because people have so many non-epistemic values and interests that come into the discussion of these issues. And so I think the same thing is true with data centers. There's, on the one hand, there's an argument to be made that data centers are useful. On the other hand, there are absolutely legitimate arguments that they cause harm if they're done badly, right? That they cause pollution and that they're a waste of money and they run rampant over people's rights and lands and all these things. And it's just very, very difficult to have that conversation because the pro-data center side is entirely biased by money. They want to make money. They're going to tell you the data centers are good because if they build one, they'll make billions of dollars. And so it's kind of hard to take them at face value when they make those arguments. But the anti-data center side is also contaminated by semi-Luddite objections to the very ideas of technology and capitalism or whatever. And I think you need a nuanced discussion, and it's just hard to find nuanced discussions. So I don't dig very much into that myself. I don't get my hackles up about data centers pro or con.

3:47:16.5 SC: The one thing, I guess, that really does kind of rub me the wrong way is the gigantic imbalance of power between rich people and non-rich people in the country, right? If I had to pick a side, I would be, given that I'm not very up on the actual issues here, I would be predisposed to pick the side of the people who are not billionaires and trillionaires because their voices are harder to get heard, right? And so I would think that if you just listened to the voices out there, you're going to get a biased view in a very particular direction. But that's no way to make difficult decisions about public policy. So I think that it's weird to see, for example, the Democratic Party, which in many ways you might think would be anti-data center. And why would you think that? Well, number one, they want to preserve the environment. Number two, they want to oppose Republicans who are very pro-data center. And number three, data centers are really unpopular right now. So you might think, wow, this is a real obvious winner for Democratic politicians. And then you look at the actual positions of Democratic politicians, and many of them are very pro-data center. And you go, oh, my goodness, what's happening? Guess what's happening. They're getting lots of money in their campaign contributions from the companies who want to build data centers. So that huge amount of money, they just distort the discussion, the public dialogue around these things by quite a bit. And that does absolutely rub me the wrong way as much as any specific issue about using up water or whatever for the data centers.

3:48:59.8 SC: Okay, the final question of this month's AMA comes from David Sotalongo, who says, "I'm curious what is the latest on your Physics of Democracy book and whether you'd be up to publish the course syllabus or reading list for the Physics of Democracy class that you taught?"

3:49:16.2 SC: Yeah, I'm writing it, getting there. Progress is being made. I have a deadline. I'm rushing to meet it. It is quite a lot of work these days going into the Physics of Democracy book, but that means it will come out, that it'll be a success that way. You always wish you had more time, but sometimes it's good to have a deadline. It focuses your mind. I actually, precisely because I'm right in the throes of writing it, I would rather not say too much about it because it changes, right? This is what it's like to write things. You write things... I've written so many words that I've then deleted, right? Or I've moved chapters around, I've changed my opinions about important issues. I've written things pushing one point of view and then realized, no, actually I don't even agree with that point of view. I gotta change it. So I don't want to talk too much about that. When the book is all done, you'll know. And I think so far, so good. I'm having fun writing the book. I'm learning a lot. I'm liking how it's turning out. I need to restrain myself about getting into the technical weeds a little bit. There's no equations in this book. Even when I talk about power laws, I even use words like lognormal, but I don't define what it means. And that's just... It really hurts me. It really stabs me in the heart to not explain to you what a lognormal distribution looks like. But I feel I have to say it. Maybe my editor will take it out, but at the moment, that's where I am. I'm just trying to say things in a fun, sprightly, page-turning way that gets a lot of real physics and a lot of real political science across, even occasionally economics and computer science and philosophy, in ways that are compelling to everybody but still more or less accurate. I can already see that there's gonna be criticisms of the book no matter what I do. I had one discussion, I won't go into details, but as soon as you say physics of democracy, people are like, "Oh yeah, that's cool." And then they say, "You better include issue X, which is my hobby horse, and you need to talk about it in your book." Everyone's gonna have an issue X that they are very sad that I didn't talk about. And what can you do? I can encourage them to write their books. That's the best I can do. But it will come out. A year from now, maybe it'll be out. Who knows? We'll have to see.

3:51:34.7 SC: So thanks as always for listening this far in, for supporting the Mindscape podcast. The support means a lot to me. I will talk to you next time. Take care.

[music]