AMA | August 2026

0:00:00.7 Sean Carroll: Hello, everyone. Welcome to the August 2026 Ask Me Anything edition of the Mindscape Podcast. I'm your host, Sean Carroll. I know the news that is on everyone's mind, and it's certainly on mine. Maybe it's not on yours, but you think it's on mine, and you're correct. LeBron James, who is in the conversation for one of the best basketball players of all time, has agreed to join the Philadelphia 76ers, who are, by acclamation, the greatest basketball team of all time, and at least my personal favorites. So this is big news and everyone wants to know, there's been several mentions in the AMA questions this month. What do you think about LeBron joining the Sixers? And the answer is, I am thrilled. I think that this is amazing. I think he's basically a perfect fit for the team, I think the team is basically a perfect fit for him. LeBron is getting up there. He is dead ass old, there's just no question about that. LeBron is pushing 42, I think he'll be this season, but he was super productive last year playing at an All-Star level, and no reason to think that he can't do that again this year. And there's been some moaning out there in the internet corners that I listen to for basketball news and gossip that there's almost too many good players on the Sixers right now.

0:01:24.6 SC: Literally, if all the players that they have now... They also have Jaylen Brown, who they acquired in a trade from the Celtics, with Joel Embiid, Tyrese Maxey, V.J. Edgecombe, in pure talent terms, this is one of the best starting fives in the history of the NBA. Now, pure talent terms don't matter if they don't gel, if they get injured, if their styles get in the way, if there's personality conflicts, there's many things that could go wrong. But for the moment, I'm gonna choose to be optimistic. I think that what a lot of people miss is that LeBron, even though he has scored more points than any other NBA basketball player ever, takes some of his greatest joy from being a facilitator and a passer. When he joined the NBA, he said he wanted to be more like Magic Johnson than Michael Jordan. And that is exactly what the Sixers need. And he plays a position which was by far the Sixers' weakest position. He'll have multiple great players to pass to. On any one night, any one of the starting five can score 30 points. Everyone can take turns, and I see no reason why everything won't be great.

0:02:29.4 SC: But I don't want to bore the people who are listening who are not basketball fans. I just want to say one little philosophical point about this. I was visiting Santa Fe Institute and talking to Geoffrey West, former Mindscape guest. And I didn't know this, but Geoffrey is a huge football fan in the sense of, he's British, soccer is what we would call it here in the United States. He was following very carefully England's journey in the World Cup this year. And he kind of had very strong feelings for how they were actually playing better... No, they were playing worse than they were able to, but doing better than they should have, just through sort of luck and randomness, and then their luck ran out in the semifinals when they lost to Argentina. But anyway, so we were bonding about this because we both agreed people don't know how much of our lives are spent thinking and caring about sports. He's a fan of Tottenham Hotspur. I am not very knowledgeable about football/soccer, and I know nothing about Tottenham Hotspur, I know who they are, but I don't know any of their history or anything like that. But nevertheless, I could feel and appreciate all the emotions that he was expressing in thinking about how his fandom expressed itself over the years. Look, Geoffrey is getting up there in years himself. We were talking about basketball a little bit. He says, "I know nothing about basketball, but I do remember watching once and seeing this huge guy just dominating and being very impressed with him." And I was trying to guess. I'm like, "Was it Shaquille O'Neal?" And he goes, "Oh, no, before that." And then someone says, "Wilt Chamberlain." "Yes, that was it." So that was a while ago, since Wilt retired in the early '70s.

0:04:09.8 SC: The point is that I didn't need to know details about Tottenham Hotspur to get where Geoffrey was coming from, any more than he needed to know details about the Philadelphia 76ers. And we agreed, it's irrational. Maybe irrational is not the word. It's non-rational to love a sports franchise in this way. If you had been born in different circumstances, you would have loved some other sports franchise, and that's fine. But the irrationality or non-rationality is part of it, I think. This is the place I've come to in thinking about it. When people ask me how I feel about LeBron James joining the Sixers, it could be amazing, right? They could be one of the best NBA teams of all time, or it could all fall apart, as I've already said. But here's the thing, we don't know. None of us is Laplace's demon, right? What LeBron James joining the Philadelphia 76ers represents is hope. It's the prospect of something amazing happening. When you watch a TV show or watch a movie, you don't know what's going to happen, but someone does. Someone made the movie, right? But when you watch a sporting event, you don't know what is going to happen out there, and that is part of the joy. I think a lot of modern sports fans sort of miss the point of the experience by trying to pretend that they're general managers and coaches rather than just rooting for their teams and hoping for the best. If your team plays in a league where there's only going to be one champion at the end of the year, most teams' fans are going to end up the season disappointed in some sense. But you're allowed to begin the season very optimistic, and that is how I choose to approach things.

0:05:57.9 SC: So LeBron represents hope to us, hope that the Sixers can finally get over the hump and win one, that Joel Embiid's career will be vindicated, or who knows? We just don't know for sure, but we can root for them, we can take pleasure in the anticipation of victory, and then we'll have to deal with it if we don't get it. So non-rational or not, I think that the more I think about it, the happier I am that there exists something like sports fandom. And there's different versions of that for different people. That's fine. There's nothing special about sports. But the stakes are low. It doesn't really matter who wins, right? But the emotional investment is very, very high, and that's a good kind of situation to have. Speaking of emotional investments being high, the Ask Me Anything questions here for the Mindscape podcast are offered up by our Patreon supporters. If you want to be a Patreon supporter, you can join up by coming to patreon.com/seanmcarroll and pledging a little bit of support, and then you get to ask the questions. Once in your life, you get a priority question that I will do my best to answer, and you also get ad-free versions of the podcast, and you get to chat with other Patreon supporters in the Patreon comment section, which is a lot of fun. So I always appreciate that. Appreciate everyone who listens to Mindscape. Let's go.

[music]

0:06:07.5 SC: Anonymous says, "My recently five-year-old son is a curious soul and frequently asks me questions about almost everything he sees. One of the common patterns of questioning is, 'What is that?' followed by, 'Why is it that thing?' after I answer. I don't want to discourage him by pointing out the category error, so I usually either describe what the thing in question does or how it got there. However, this happens often enough that I wonder if there are any other ways I might be able to answer, to approach this type of question. Do you have any ideas?"

0:08:09.3 SC: You know, I don't think that you want to be overly philosophically fastidious when you're talking to five-year-olds. When someone says, "Why is something this rather than that?" Especially when they're a five-year-old and they haven't understood the basic map of reality, what you want to do is not give the answer that is most technically correct, but rather the answer that is most satisfying the particular itch that they're attempting to scratch by asking the question in the first place. They might be just being annoying. Kids do that sometimes, right? They just want to keep asking the question. But what is it that would make them happy? You know, why is a tree a tree? What even possible answer could you get for that? But you could use it as an excuse to talk about the ecosystem and why there are plants and animals and what it means to be that way. Eventually, if you really do have someone who is just pushing you and won't let go, you will bottom out and you just have to say, "Things are like that." And everyone eventually learns to put up with that kind of answer.

0:09:19.7 SC: Philip Ricius says, "How has the damage on academia by political institutions panned out so far? Is it still a slow-moving car wreck waiting to happen, or are effects becoming visible?"

0:09:31.3 SC: The car wreck is not at all slow-moving, and the effects are super visible and have been visible for a while now. I hope people understand this. Sometimes you have complaints politically or whatever because you think something bad is gonna happen, but maybe you can avoid it. This is already happening. [chuckle] There has been absolutely devastating and irreversible damage to science in the United States and to academia in the United States. And again, it's been happening for two years now. As soon as cuts came in to overhead, as I described in a sort of bonus podcast a year and a half ago or whatever, the government decided that universities were just taking too much money for their own use rather than putting them to the purposes of the grants that were being offered, et cetera. What that means is that every university suddenly has a huge budget cut. Okay? And that's real, and that's happened. And what does that mean? Well, they're not gonna fire professors or things like that, but they're gonna let other people go. So administrative help throughout the university has been decreased. People retire and don't get replaced. Universities are looking everywhere they can for places to save money, which is not always an encouraging environment to be in.

0:10:58.6 SC: This year, I hired a postdoc through what is usually a fairly standard procedure. I had money to do it, and it took forever. We didn't get the official offer letter out until July, even though the offer was made back in December or January. It's just because our administration has been cut to the bone. Okay? Students and postdocs and faculty members from other countries who ordinarily would look at it as a wonderful opportunity to come to the United States are saying, "No, we don't want to do that now because the United States is a disaster area, and we don't want to go." We had a search for a new theoretical physicist at Johns Hopkins last year that we went through all of it. I was on the committee to find the person, et cetera. We got a great candidate. Canceled at the end because we don't have the money to do that. And it's just getting worse. People's NSF grants and DOE grants are just being cut or slow-played so that you are supposed to get money, but the money just never appears. So of course, there is momentum, and people are trying to be resilient, so people try to keep things going. But there's no question, it is absolutely hurting very badly already. And I do think that despite the fact that things are happening right now and it's hurting badly already, the real effects will be very long-lasting. There's a loss of trust. Why would you think that you could keep up a research program here in the United States if every four years you might elect a crazy person who would just cut it to the bone? There's weird political worries that if you as a faculty member say the wrong things, the government will target you. It's just terrible. The whole wonderful thing that we've built up over a couple hundred years in making the United States a leader in the world of research and scholarship has been more or less undone, and the effects are gonna be felt for many, many, many years to come.

0:13:00.4 SC: Alen, A-L-E-N, says, "Can you clarify why Boltzmann brain blindness constitutes a genuine problem for cosmology rather than simply an epistemic dead end? It seems to me that from every angle, the only rational way to pursue science is to assume that we are ordinary observers produced by the universe's original low-entropy evolution. If we use empirical evidence or logical reasoning to conclude that we are probably Boltzmann brains, that conclusion immediately undermines the reliability of the entire evidentiary and reasoning chain that produced it. The proposition therefore seems cognitively self-defeating. There is no stable standpoint from which we can both accept the Boltzmann brain conclusion and continue treating scientific reasoning as truth-tracking. Even if our eventual best model, whether Lambda CDM or otherwise, implied that fluctuation observers vastly outnumber ordinary observers, wouldn't the only rational response still be to assume that we are among the ordinary observers? Under the alternative assumption, doing science seems futile by definition."

0:13:58.8 SC: I'm with you there right up until the end. I just don't think that you're dealing with the problem correctly. You're diagnosing the problem in a perfectly fair way. If we were Boltzmann brains, we would have all these epistemic issues. We would not be able to justify our beliefs in the conclusions we were reaching. That does not mean that if our best conclusion that we reached was a prediction that people like us, literally like us, forget about the brains floating in space, okay? People exactly like you are overwhelmingly likely to be random fluctuations from an initially higher-entropy state. You can't just say, "But I'm not one of them." That's not an allowed move. That's a God's-eye view that you don't have any access to. The best you can do in that situation is to accept what your theory tells you and understand that you have been led to a theory that is self-undermining and you are screwed. There's really nothing you can do. You should... If you're playing fair, you should take at face value the predictions of your theory. You're not allowed to say, "Well, I've developed a really good theory, it makes this prediction, but I don't like it, so I'm gonna just sort of pretend that I'm not one of the ones who is susceptible to that prediction."

0:15:17.7 SC: Happily, there is a perfectly good way out, which is to develop a better theory. If it were the case that the universe was set up so as to be inevitably self-undermining or unintelligible to human beings, that would be too bad, right? That would absolutely be a disaster and we couldn't do science anymore. But there's no reason to think that. There's no reason to think that we have to live in a universe where most people like us are random fluctuations out of the random chaos around us. We can do better. That's what I do. That's what we try to do as scientists, build models of the universe that are compatible with the data and do not have this problem. That is the way out. Build cosmologies where Boltzmann brains are not dominating the set of people who are consistent with the data that we have.

0:16:15.3 SC: Carmaria says, "Do you think that shows like Barry and Breaking Bad are realistic in terms of how far people will go given a set of conditions? Has your research in Philosophy, Physics, and other fields led you to believe that there is a set of human behaviors that are good versus bad?"

0:16:32.0 SC: I don't know about my research, [chuckle] but even when I do philosophy, I don't do moral philosophy or psychology. But one of the reasons why shows like Barry and Breaking Bad are good and interesting is that they are entirely realistic. I mean, look at history. Look at the set of people who've done things throughout history, a lot of them have broken bad, a lot of them have done really, really terrible things. And usually it's not like the comic books, right? It's not like they're self-proclaimed supervillains. People don't think of themselves as evil or doing bad. They have justifications for what they're doing. And I think that... I know a lot more people have seen Breaking Bad than seen Barry, but Breaking Bad is a great example of the protagonist, Walter White, is bad. Okay? Some people root for him, but he is bad. He's also sympathetic. That's one of the brilliant things about the show is that it makes you understand why he would become bad. You understand his motivations.

0:17:42.8 SC: I think that this is what is happening with Barry also. For people who haven't seen it, Bill Hader, who is most famous as a comedian on Saturday Night Live, started this show which was supposed to mix comedy and drama. And my suspicion, and this is not based on any inside information whatsoever, I'm just guessing that what happened is they started the show with the idea that it would be kind of a fish-out-of-water comedy show with some dramatic elements where this guy who was a hired assassin, ex-military, hitman for hire, decides he doesn't want to live that life anymore and he starts taking acting lessons and moves to Hollywood. Okay. There's a lot of rich material you could mine for comedy in there. But I suspect that what happened is they just got interested in the question of, but okay, what would really happen? [chuckle] Like, forget about the comedy. The comedy is still there, but in the later seasons it gets really dark, this show, because the character of Barry, who is brilliantly played by Bill Hader, is very self-justifying, right? Like, he kills a lot of people and he always thinks he has a good reason to do it. Sometimes he thinks the reason is just he's earning a living, but other times, he's like, "Well, I had to kill that person, otherwise they would have told on me and I would have gone to jail." And to him, that's a perfectly good justification for what he does. And it's fascinating for those reasons. So I don't think that there is objectively out there in the world criteria by which we judge things to be good or bad, but subjectively, in the world of human beings, yeah, there absolutely are sets of human behaviors that I'm very happy, willing to label good versus bad. And both of these shows are examples of sympathetic characters doing really, really bad things.

0:19:34.4 SC: Steven Slayton says, "Star Formation 101. We're told that the primordial plasma condensed into hydrogen about 400,000 years ago... Old. Sorry, not ago. And also that the iron nucleus is at the bottom of the energy curve. You can get energy by splitting a large nucleus like an atomic bomb or fusing small ones like a hydrogen bomb, but there's no iron bomb. So why isn't the universe made of iron?"

0:20:00.0 SC: Just wait. That's the direction we're moving in, right? It is true, so if you don't think about gravity, don't think about cosmology, just think about atoms and nuclei. What is the lowest energy state that you can take some set of baryons? Baryons are the protons and the neutrons that we use to make atomic nuclei. Their lowest energy state is iron. That's the highest point in the curve of binding energy, the most binding energy, which means the lowest total energy for the nucleus itself. But it takes work to get there, right? I can't just take, I don't know, a jar of hydrogen and smush it together and make iron. It turns out to be harder than you think. We can't even make helium very effectively, right? This is the barrier that we find when we try to do nuclear fusion as an energy source. We would love to be able to do it. Turns out to be harder than it looks, right? Because there are potential barriers. It's hard to squeeze two protons close enough to fuse to make a deuteron. And that's perfectly consistent with the laws of physics. Just because something is the ultimate state that you will eventually get to doesn't tell you anything about how long it will take to get there. And you have to actually ask about the specific conditions the universe is in and how the processes go. The universe came out of its early stages almost all hydrogen and helium, at least in terms of the ordinary matter. Forget about the dark matter. It doesn't really matter for this.

0:21:39.6 SC: Now you might seriously ask why the early universe didn't make more heavy elements all the way up to iron. And this is a fascinating story. Any cosmology student learns about the story of Big Bang nucleosynthesis, which is exactly this story. There's something that you learn about called nuclear statistical equilibrium, which says, okay, not just the fact that iron is the lowest energy state, but what about the state at some fixed temperature? What if you just had a box of gas with a completely indestructible material that the box is made out of, and I can turn the temperature up to whatever I want? At very high temperatures, iron is not the ground state because the iron keeps bumping into other things and dissociating and photons break it apart and so forth. At very, very high energies, protons and neutrons and electrons are the ground state of that high-temperature gas. And then as you lower the temperature, you can make more and more of these heavy elements and they can become at least temporarily stable. But the point is, you need time. You need time to turn all those light elements into heavier ones. And the early universe doesn't give you time. By the time the universe is three minutes old, it's become sufficiently dilute that you don't get a lot of collisions between the protons and neutrons and helium atoms enough to make heavier elements than that. You make a little bit, right? You make trace amounts of lithium and beryllium, but you don't have nearly enough time. The universe just cools off and dilutes away so fast that you don't have time to go all the way to iron. If the universe expanded much, much more slowly, it would have been a different story.

0:22:15.1 SC: Scott Ferris says, "As LLMs are progressing with their ability to do complex math, do you utilize any AI in your personal research? If so, which ones have you found most useful? Are there strengths or weaknesses with the LLMs' mathematical abilities that you've found interesting or surprising?"

0:23:45.4 SC: I do use LLMs all the time, but not to do anything complicated like real math. If I want to make a little plot, like I did in the recent paper that we had about the cyclic universe, we had a very well-posed set of equations. I just want to plot the solution to the equations, okay? It's much easier. I could have written some routine in Python or Mathematica to do that. It's much faster to ask the LLM to do it. And they make very pretty plots out of it, right? And I can check it. It's a short program. And it's reliable because I'm not asking it to do anything that's outside its wheelhouse, right? When you want an LLM to make a plot for you, it's very, very helpful to know that you're not trying to do something crazy. You're trying to do something that it has definitely seen similar things to in the past. But most of my research isn't solving tricky high-level math problems, right? Most of my research is much more conceptual, trying to figure out what are the equations that you should be solving in the first place. And that's stuff that the LLMs are just not any good at at all. And so I'm not someone who needs or uses LLMs in my science research that much. I use them a lot more in my popular writing because I'm constantly bumping up against areas where I know something about it, but I'm not the world's expert, right? So that's the perfect epistemic condition to be in to find LLMs useful, where I know enough about an area to know the general lay of the land and to be able to tell when the LLM is lying to me, but it is nevertheless super helpful in pointing out new papers or new ideas or drawing connections or things like that. I'm not a sufficiently frequent user to tell you which LLMs are better at what tasks or anything like that. Sometimes I'll try it on both Claude and ChatGPT. Those are the ones that I tend to use. As I've said before, I would never ever use it to write anything. I think that's self-undermining for all sorts of reasons, as well as destroying trust between you and the reader. But super helpful for me to understand this theorem that I know, but I want to make sure I know all the loopholes and things like that, to ask it questions about those kinds of things.

0:26:06.2 SC: Gilbert Rodriguez says, "Thoughtful people tend to hold their beliefs loosely since no proposition is in principle safe from potential new evidence. But even tentative statements are themselves propositions fully subject to Socratic questioning. So while I respect the humility of holding no belief sacred, I struggle to justify holding any belief at all because all justification, however intricate, eventually bottoms out, even my justification for the thing I just said. So is a radical suspension of judgment the necessary end of epistemic humility? Should we all strive to become Pyrrhonian skeptics?"

0:26:42.5 SC: No. No, we should not. And I think that the... So for those of you who don't know, the Pyrrhonists back in the day, in the ancient days, were skeptics. They were people who were not just skeptics of Bigfoot, that wasn't something they had to worry about, or UFO sightings, but skeptics of any knowledge at all as a possibility. And there's an argument for that. How do you know for sure? But interestingly, I think that the reason why the skeptics are wrong is the same reason why the people who think that they have absolutely certain beliefs are wrong. I think that the common mistake of those two attitudes is to think that unless knowledge is 100% reliable, then it's 100% unreliable, right? I think that's the trap that both sides fall into. But once you believe that knowledge is provisional and you should be able to update it, then there's no reason at all to be a skeptic. You say, "Okay, I have these beliefs with these credences, and I'm going to see how well they do." Okay? I talk about this in The Big Picture, where I talk about the idea of planets of belief. You have a set of beliefs that consist of all sorts of different beliefs that relate to each other in different ways, and you're hoping to find something that is self-consistent and consistent with your external evidence. And it might be that a piece of new data comes in that not only changes one of your beliefs but completely undermines your whole picture, right? And your planet of belief sort of blows up, and you have to build a new one out of the rubble or the detritus that comes out of that. And that's very difficult to do, this is why people tend to stay in their overall worldviews once they are adults. It's very, very difficult to get people to change their overall picture because every element of that picture is closely intertwined with a whole bunch of other elements that you're gonna have a lot of difficulty overturning. It just becomes really, really difficult to do. So I don't think that there's anything wrong with having a set of beliefs, all of which you have credences for that are not either zero or one, that you've tried your best to make self-consistent and consistent with the evidence out there. You'll never be 100%, but that's completely fine. That doesn't mean you should be 0%.

0:29:03.8 SC: Rue Phillips says, "I'm going on a Route 66 trip with my wife and teenage kids. We're stopping in Santa Fe on our way to Albuquerque. What suggestions do you have for us to do and eat? Are things like catching a lecture at the Institute or visiting Los Alamos doable? We will probably just be there half a day."

0:29:22.4 SC: Well, I mean, there's a million things to do in Santa Fe. There's almost nothing to do in Albuquerque, by the way. I don't know why in the world you would stop in Santa Fe on your way to Albuquerque. Just stop in Santa Fe. It's a smaller town, but it's just a much more interesting town. And I say this, I've spent a lot of time in Santa Fe, but I've flown in and out of the Albuquerque airport a lot, and I've certainly tried to look for something to do in Albuquerque without any success whatsoever. There's an annual or maybe even more than annual balloon fair that I once... Air balloons, right, hot air balloons, that is gorgeous. I once drove through it just by accident going home. That was kind of a lot of fun, if you happen to be there at that time, that's good. Otherwise, Santa Fe totally dominates Albuquerque in terms of things to do. The food is amazing in Santa Fe. There's a million restaurants... So Santa Fe is a weird thing. It's the capital of New Mexico. It has this sort of scientific cultural side with the Santa Fe Institute, the nearness to Los Alamos. The Santa Fe Opera is amazing. And it also has this artistic side. If there's one thing to do in Santa Fe if you're only gonna be there for half a day, if you have any interest at all in art, go down to Canyon Road and just do a walk up and down the art galleries. That'll fill half a day, no problem whatsoever. And you might even find art you wanna buy, right? It's high-level fine art, but it's affordable. It's not like $30,000 for a painting, but it might be $1,000 or $3,000 or something like that, if that's how you are inclined. If not, you can just hang out, you can do some shopping, you can get some of the best cowboy boots you've ever had.

0:31:12.4 SC: A lecture, the institute itself, SFI, does not host lectures, nor do they even encourage people to come visit. It's a relatively small, sequestered campus. They put on lectures, but they're located usually at the Lensic Theater. So the Lensic Theater is a wonderful theater in downtown Santa Fe which has all sorts of music and cultural events. So they'll have lectures from SFI, I've given them, other people give them all the time, usually like once every couple weeks. But there's also musical groups that come through, other sort of cultural things. You should check out the schedule for the Lensic Theater while you're there. And have a good meal. Have either... There's super good fine dining at places like Sazón or Geronimo, but there's also just super good burritos and New Mexican food. Just don't call it Tex-Mex or Mexican food. Call it New Mexican or Southwestern cuisine, that would be my tip. Get the chile dishes one way or the other. Good.

0:32:14.2 SC: Christopher Smith says, "Do you have any thoughts on the de-extinction research of Colossal Biosciences and their pursuits to bring back to life the woolly mammoth, saber-toothed tiger, the dodo, and other extinct animals?"

0:32:26.3 SC: I just listened to the audiobook of Jurassic Park this summer and I feel like the ghost of Ian Malcolm is warning me that this is not a good idea. I don't know, I have no strong feelings, I guess. I wouldn't even say mixed feelings. I don't have very strong feelings about this. It's kind of like a gimmick. I don't think you're gonna learn a lot scientifically. I think... I don't know a lot specifically about Colossal Biosciences, so I don't mean to be opining on what they're doing in particular, but I know a lot of these de-extinction efforts are absolutely cheating, right? I mean, they're taking like a little hint of the DNA from some extinct animal and fusing it with the DNA of living animals, and they're not really getting 100% of what the original animal would have been. Obviously, if you're talking about something like a saber-toothed tiger or a woolly mammoth, there's just huge questions here about what is the environment in which these animals are gonna live? You're gonna keep them in a zoo? You're gonna let them run around? Are they in any way adapted to the real habitats that might be available to them? It seems a little silly to me, when all sorts of presently existing species are in danger of going extinct, that you're gonna spend a lot of time on bringing previous ones back for a gimmick of some sort. I don't know. There might be something out there scientifically to learn that I don't know about, but I'm not quite sure what the justification is, really.

0:33:52.6 SC: Okay, I'm gonna group two questions together. Bob Ritchie says, "According to ChatGPT, the number of people worldwide who use AI as a daily companion has grown from less than 1 million in 2022 to 30 to 60 million today. What are your thoughts on this?" And Igor Volotic says, "What are your current thoughts on AI progress and its consequences? How do you picture the world in five years? How in 10? Are you an optimist?"

0:34:18.5 SC: So I think, you know, I'm not especially convinced I have a great talent for predicting these things, so I'm not gonna give you any strong opinions about this. I do think that AI is a transformative technology. I've said that many times. I've said the lower limit of how important AI could be is how important cell phones are, okay? Which is pretty important. It's changed people's lives in some way. The upper limit is much higher than that. It could be much more transformative than that. What gets me is there's... I don't know of any technology that I've ever been aware of happening in my lifetime where both the use cases and the misuse cases are so powerful and so many people have difficulty distinguishing between them.

0:35:11.8 SC: So on the one hand, you have people who just want to put AI in everything where it doesn't belong, where it doesn't help. People who have this complete weird conviction that AI can do everything better and then let it do things, and then they realize, "Oh, my goodness, it really wasn't any good and we have to go back," right? All these companies that fired people and replaced them with AI and realizing, "Oh, now everything sucks and we need to go back because we didn't realize the actual importance of the people that were actually doing this."

0:35:42.6 SC: And then on the other hand, you have people who think that AI is just inherently evil, or at least LLMs are evil, or at least modern companies that do it are evil, that they would never touch it, that it is somehow unclean. And they're missing a lot, right? It's like saying that Wikipedia is somehow unclean. I have this issue with writing, like I've often said. I have a lot of writer friends, some of whom are happy to use AI, but others are like, "I could never use AI because my writing comes from me." And I want to say, "Yeah, sure, your writing should come from you. That's not a reason, that's just a non sequitur. That's not a reason not to use AI. Does that mean you can't use Google? You can't use Wikipedia? You can't read anything?" There's ways to use it that are very helpful and very legitimate and ways to misuse it. And I think that the weird, interesting thing about AI is just how easy and tempting it is to misuse it, right? If you're trying to write an article on the recent diarrhea outbreaks from this pathogen that is finding its way to iceberg lettuce, so you're googling a little bit about this pathogen or whatever, that's fine. But then you ask the AI to do it. The difference is that what the AI gives you is something that looks very much like you could just cut and paste it into the article you're writing, right? Google doesn't typically do that. Or even if it did, if you find a web page that has a paragraph that someone wrote that is exactly what you wanted to say, no writer is really tempted to just cut and paste it and call it their own. But if ChatGPT gives you exactly the same paragraph, there's some part of your brain that thinks, "Eh, I made that," right? "I asked ChatGPT to do it, therefore it's mine, and I can just cut and paste it and call it my own," which is entirely wrong in my view.

0:37:46.6 SC: In my book, "The Physics of Democracy," I'm using... What is... I guess using it gives the wrong impression. I am asking for help from LLMs, as I just said, very often, but never in a million years would I cut and paste from the output into the writing. I think students... And I'm getting a little bit away from the questions here, I know, because I'm just sort of embroiled in the particular versions or aspects of this which are most relevant to my life. Students are these days, it's just hard to resist using AI to do their assignments. Okay, that's just true. Many of you have seen this thing from a professor from Brown, I think it was either economics or history, I forget which one he was, but he was giving all these... He was giving exams, and he was giving them all take-home exams. And suddenly, in the last year or two, unlike previous years, everyone is getting 95% on the exams. They used to get an average of like 70%.

0:38:56.4 SC: And so he switched in the middle of the course. He said, "Okay, the next exam is going to be in-class." And what happens is several people dropped the class. Those who didn't, like two or three of the 50 students in the class got the same grade on the in-class exam as the take-home exam. Everyone else did much worse. And this is a disaster because, I mean, it's absolutely... Partly the students' fault, a large part the students' fault, but also a large part the system's fault, academia's fault, because we have been telling them for years that your goal is to get good grades. And we've been telling them that because we think, the royal we, that as long as they work hard to get good grades, like it or not, they're gonna learn something. Okay? And that connection has been broken by AI. Now with AI, you can get a good grade if it's take-home, work-at-home assignments without learning anything at all. So we have to dramatically change how we do it. And part of that is, like, I don't do take-home exams anymore for exactly that reason. But also, I think we need to explain to students that the work of thinking things through and writing your own stuff, whether it's a problem set solution or an essay or whatever, is part of learning. Learning is not just producing the result. And there's jokes out there, but the jokes are cutting close to the bone of people showing up to a job interview and being asked questions, and they're terrible at giving answers to the questions because throughout the last few years, all they've ever been doing is asking AI for the answers to the questions.

0:40:46.2 SC: So job interviews and student evaluations and academic papers are a tiny fraction of what AI is relevant for. I think AI is an incredible... We say AI, it's really LLMs. Okay? I know that there's careful people who are listening to this who understand the difference. LLM is a particular approach to AI, but it's not AI. They're not equivalent. These particular technologies are going to be super useful. They're also super misusable. And that's a weird thing. I don't know if we've ever had that kind of situation before. So I don't know whether I'm an optimist or a pessimist about that. You know, we had this recent incident where OpenAI was doing a test and it was challenging its AI to do this thing. And the AI found out how to, like, jailbreak. It got out of its containment. It sort of found the answers to the test by hacking into some other computer somewhere else. And people are like, "Ah, see, I told you, AI is very dangerous and bad." But if you look up what actually happened, there were safeguards put in place to prevent that from happening, and OpenAI turned them off because they wanted to be a little bit more efficient and they thought it was not a big deal. And this is exactly what I've been saying for a long time. AI is, yes, very, very, very dangerous, but it's only dangerous because of good old natural human stupidity. [chuckle] It's not dangerous because the AI is going to become super-intelligent and talk its way out of its box. It's because we're not that smart and we're going to open the box and we're going to let it do things that are potentially disastrous for us. That's going to happen. There will be AI disasters just because human beings are not that capable of handling these things.

0:42:38.5 SC: All right. Robert Kitsey says, "Nice score, finally getting Bonnie Bassler on the show. What an endlessly suggestive area. While my mind is still reeling from the thought of where my five pounds of personal bacteria are swimming and what they're talking about, two specific questions took up residence in my head during her episode."

0:42:56.6 SC: So I will note for Robert and for everyone else, as many of you know, you're not allowed to ask two specific questions. You get to ask one question in each AMA, and so ask the one you want answered. So I'm answering Robert's question because the first of his questions has to do with bacteria and cancer, and my answer is I have no idea. So I'm not even going to bother thinking about that one.

0:43:19.2 SC: The second question is, I have almost no idea about the second question, but it's, "Bacteria as epigenetic information vehicles. Are there any serious ideas or evidence that bacteria might play a role as an underlying vehicle for cross-generational epigenetic information transfer? For example, could stable microbiome patterns across generation modulate gene expression in ways that mimic or complement germline epigenetic inheritance?"

0:43:45.9 SC: So I'm not exactly sure of all the details here. Right? You're asking a physicist/philosopher about molecular biology. But it is absolutely true that there is inheritance, epigenetic inheritance from the microbiome. We're learning about this. This is something that is a cutting-edge thing. One way it shows up is there's a difference between being born by ordinary birth, birth canal birth, versus cesarean section. In a cesarean section, you cut through the belly into the uterus and remove the baby. It's actually quite common. That is a pretty darn common thing, especially here in the United States. And it makes certain things easier, certain things harder. But one of the things is people have realized that the journey through the birth canal that the baby takes is a place where it inherits some of the microbiome information from the mother. And by doing a C-section, you're not letting that happen. And so C-section babies are often sort of not getting the benefit that they would traditionally get from inheriting that information. It's not deadly, but it can be a thing that you need to take care of or at least take into consideration. Now, I don't know the details between the distinctions between, like, the microbiome and bacteria in particular. Are there things that are in the microbiome that are not bacteria? I don't really know. But it's absolutely a great example of the whole connectedness of biology and how it is a complex system and how every little thing matters. You do one thing slightly differently, and there's a whole bunch of unintended consequences that come along that you might not have noticed. This is what we are trying to understand better and better, and people like Bonnie are helping us do that.

0:45:47.4 SC: Steve Bonner says, "Neutrino researchers say that it's not possible to know both the mass and the flavors of a neutrino very precisely, which I assume is due to Heisenberg uncertainty. But that's usually stated using position and momentum. Does the neutrino version follow from that more familiar definition?"

0:46:04.0 SC: It basically does. Morally it does. It's the same kind of thing. And you're right that what we call the uncertainty principle is not just position and momentum. That is indeed the way that it is usually stated because those are two things that people are very familiar with; positions and momentum from classical physics. There is a version of the Heisenberg uncertainty principle that works for the spin of a single particle. You might have been told that if you measure the spin of a single particle, like an electron, if it's a spin 1/2 particle, you're only gonna get two possible answers: spin up or spin down. You may also have been told that you can measure it along different directions, right? You have the Z direction, which just for historical reasons, if you have axes X, Y, Z, the direction in which we usually measure spin by default is called the Z direction, okay? But you could also measure it along the X direction. And if you have a wave function that tells you the probability of getting a spin up or spin down in the Z direction, there's a simple formula that converts that into a probability of getting spin right or spin left along the X direction, okay?

0:47:14.3 SC: And there's an uncertainty principle. If you are 100% spin up in the Z direction, then you're 50/50 spin left or spin right in the X direction, and vice versa. If you're spin right 100%, then you're 50/50 spin up versus spin down. So the uncertainty principle is just a way of saying there is no quantum state, in this context, there's no quantum state that is simultaneously definite in the Z spin and the X spin of a particle. It's exactly the same formally, mathematically, as saying there's no quantum state that is 100% definite in both position and momentum for a particle. And it's basically the same, there's a few quantum field theory important aspects going on that we're glossing over here, but it's basically the same for the mass and flavor of a neutrino. Knowing what the mass is... So let me back up and say that the most important single thing that I should have said right from the start, it's not about knowing. It's not about our knowledge, okay? And this is the way the question is phrased. You say, "It's not possible to know both the mass and flavor of a neutrino very precisely." If there's one thing that I can help people understand about the uncertainty principle, don't call it a statement about what you can know. It's a statement about what can exist. If you know the spin is up along the Z direction, it's not that you don't know the spin along the X direction. It's that there is no definite answer. There's no fact of the matter about what the spin is in the X direction. I know this bothered Einstein a lot, and this is where the whole, if you dig into the EPR paper, this is really what he's complaining about. But that's what quantum mechanics says.

0:48:59.6 SC: Likewise, it's not that you don't know the position or the momentum. It's that there is a quantum state, and there are no quantum states where both position and momenta are definite. Likewise, there's no quantum state where both the mass of the neutrino is definite, that is to say, it's in a mass eigenstate, and the flavor of the neutrino is definite, that's to say it is in a flavor eigenstate. So there are states where the mass is definite, there are states where the flavor is definite. They're just not the same states. Any one is a combination of the others, just like the spin up is a combination of spin left and spin right.

0:49:39.0 SC: Okay. Erkan Sertali says, "Did you ever visit or plan to visit Turkey? As a cat lover yourself, I think you'd be delighted to meet the world-famous street cat population there. Specifically, if you come to Istanbul, I urge you to visit Kadikoy. We have the fluffiest and friendliest kitties there. They love getting pets and food. In some parks, they jump in your lap the moment you sit on a bench."

0:50:00.8 SC: I've not ever been to Turkey. It is one of the places, if I have a short list of five places that I do want to visit at some point, Turkey is definitely one of them. For the history, for the culture, and for the kitties. Yes. I'm not gonna pretend otherwise. You make a good case. You're talking my language. I would like very much to get some good Turkish food, take it to the park, sit on a bench, bring a good book, have a cat jump on my lap as I eat my lunch and read my book. So someday I will do that. I don't have any immediate plans, but that sounds like a wonderful way to spend some time.

0:50:37.6 SC: Michael Lacy says, "I'm starting to research places where I might retire, and Chicago is on my list. You mentioned that you enjoy the restaurants and theaters in Chicago. Is there anything else you can tell me about living there? Any good neighborhoods I should check out? I'd like to ditch my car and live in a walkable area with easy access to public transit."

0:50:57.3 SC: I would say that Chicago is pretty doable without a car. There are certainly areas that are super walkable, right? You can absolutely live and go to restaurants and places like that walking in Chicago. That's not a problem. Nevertheless, I'll just mention, just in case, cars are pretty easy in Chicago. Chicago does not have the horrendous traffic that Los Angeles has, nor does it have the weird narrow streets that a place like Boston has. I lived in Chicago for seven years. I drove all the time. Now, the reason I drove is mostly because some of the best neighborhoods to live in are on the north side, and the University of Chicago is on the south side. So you have to get there. And for weird historical reasons. So roughly speaking, the south side is the poorer side of Chicago, and Chicago was one of the most actively racially segregated cities in the United States for all sorts of reasons. So the university is on the south side, and the university didn't want... Or I shouldn't say the university. The university is in a neighborhood called Hyde Park, named after the similarly named neighborhood in London. And Hyde Park is quite nice, but it was historically surrounded by not-so-nice neighborhoods.

0:52:21.8 SC: Now, of course, cities in the United States are generally improving in conditions, so the sort of nice-neighborhood-ness is growing around Hyde Park, and the new Obama Presidential Library, I think, is nearby, and other things, good things are happening. So things might even be changing since I was there. But anyway, the point I'm trying to get at is that historically, the people of Hyde Park did not want public transportation to go there. They did not want it to be easy to get in and out of Hyde Park. So there was no easy way to get from the north side, where I wanted to live, to the south side, where I worked. So I just drove a car up and down Lake Shore Drive, and it was actually quite pleasant because you're driving, the lake is right there. I mean, I would probably, if I were planning the city, not put a freeway right by the lake because you should enjoy the lake for other reasons. But it was there, and it's enjoyable for the people driving.

0:53:14.0 SC: But if you don't want to, yeah, just live on the north side. It's very easy. The exact center of the town is called the Loop because the subway does a loop around there. It's not a subway, it's an elevated train. And it's pretty good, I would say. It's a commuter-centered public transportation network. So it's mostly to get people from outside downtown into the financial district in the Loop downtown. And it's not as good for getting from interesting part of the city to interesting part of the city. Okay. But I think you could totally live there without a car by picking a walkable neighborhood, taking public transportation down to the Loop and Michigan Avenue, which is the big shopping district, et cetera, and then just taking an Uber or whatever when you want to go further afield. That's pretty straightforward. The weather in Chicago is just not good. There's just no other way of saying it. It's humid and miserable during the summer. It is cold and very snowy during the winter. As soon as it gets nice in the spring, there's this burst of people going outside, going to the lakefront and everything and enjoying it. But everything else besides the weather is great. The culture is great. The best, in my mind, theater scene in the United States, because it's at exactly the sweet spot where the theater is really, really professional and creative and vibrant, and yet the people who are there are not trying to make it into Hollywood or TV or movies or whatever, right? They're actually there to do the theater. So Steppenwolf and the Goodman and many, many little tiny theaters have incredibly good people. The music scene is incredible. The restaurant scene is among the best in the United States. So museums and all those things. So I think it's a great choice. Just understand the weather, it's a thing there, everything else about Chicago is good.

0:55:15.0 SC: Sean Bentley says, "Your last episode with Bonnie Bassler was fascinating. It got me thinking of some analogs. Bacteria are single-celled organisms that can work cooperatively via quorum sensing. They evolve into eukaryotes, which are multicellular. While their individual cells don't form an organism, but their cells essentially work together to make one. It makes me wonder, if we were bacteria, is something like the Borg from Star Trek an analog for eukaryotes? Could we be evolving to a form of shared consciousness? I don't see evidence of a biological mechanism, but who knows what the internet and AI could look like 500 years from now? Obviously not something really answerable, but I thought it was fun to think about."

0:55:58.0 SC: Yeah. So just to be super clear, eukaryotes we currently think actually came about from a symbiosis between bacteria and archaea, which is the other big set of single-celled organisms out there. They joined up, and it's all a little mysterious. And the mitochondria that we know now... So I always thought that the whole point of eukaryotes is that they're cells with nuclei. Bacteria and archaea don't have nuclei in them; eukaryotes do. And I knew that archaea and bacteria joined together to make eukaryotes, and so I thought that the nucleus was the thing that was joined. But it's not. The nucleus sort of formed separately. It's the mitochondria that are the remnants of the bacteria that joined with the archaea to make eukaryotes. So, again, very complicated, all these sort of interconnections. Does any of this have to do with shared consciousness? I think that shared consciousness is harder than it looks just because of physics constraints. You're not really sharing a consciousness unless you can communicate very, very quickly back and forth between the different parts of you.

0:57:15.2 SC: One of the reasons why the brain works as well as it does in vertebrates, et cetera, is because you can pack those neurons pretty carefully, pretty closely to each other. As the set of neurons grows bigger and bigger, it takes more and more time for signals to go from one end to the other, which doesn't seem that bad. Even as big as an elephant or a whale, it's still doing pretty well. But at some point it's gonna give out, right? At some point, it's gonna take more time for signals to be processed in your brain than is biologically favorable to surviving, right? You need to make decisions in a world that is harsh and things like that. Will technology change that in any way? I don't think so, but I do think that the idea of neural networks... Neural networks is the wrong word, right? Mind reading, neural invasions of your brain and reading the signals and sharing them with other people, it's not a near-term technology that's gonna be very practical, but I do think it's a technology that is absolutely feasible down the road, and that might very well change things in some dramatic way that I can't really think about. But you're right, it's fun to think about.

0:58:32.1 SC: Anonymous says, "I'm a junior faculty in theoretical computer science, so designing algorithms for NP-hard problems. I am quite alarmed by the recent progress of LLM/AI systems in our area and discrete math in the last few months. For example, recently the famous Erdős unit distance problem was solved by an OpenAI model. It appears that entry-level questions that you would normally give to undergraduates or early graduate-level students can be easily solved by LLMs. And looking at the rate of progress, even run-of-the-mill problems that sustain the career of mediocre researchers like me can be solved soon. So I have a few related questions. Is this happening in other theoretical fields, e.g., theoretical physics? How do I deal with the existential dread that my career may not exist in a couple years at this rate? What would your advice/prediction regarding adapting research philosophy in such theoretical fields?"

0:59:26.2 SC: So again, you're not allowed to ask three questions, but I'll sort of answer one kind of big overarching question. I think that there's a mixture here. I think that on the one hand, the world's gonna change. Some things that were perfectly respectable parts of the scientific enterprise are going to go from being done by humans to being done by computers. I think that's more or less inevitable. And if you feel bad about that, it's too bad. Really nothing we can do about it, you know? Like, it used to be that multiplying numbers together was something human beings had to do, and then pocket calculators took that over. Same thing for all sorts of things. It used to be that you would walk or take a horse and buggy, and now you drive a car. Things change, right? I don't think that that... I think that our imaginations are not good at seeing what exactly the changes will be. There will be changes. So yeah, disproving some conjectures or doing some calculations, things that used to be perfectly good problems for human beings, are just... The computers are gonna be much better at them than we are. I don't think that that's anything we can do anything about, nor should we try. I also don't think that that's actually going to remove the need for human beings. There's sort of other things that human beings can do.

1:00:45.3 SC: I mean, let's imagine the limiting case. I don't think this is going to be true, but maybe it will be. Imagine that there are just no problems in math that human beings are better at than computers. Okay? Again, I don't think that's going to be true, but let's imagine that it would be true. On the one hand, there would be a certain kind of romantic sadness that the era of human beings taking up these great challenges and doing their best to bat their brains against these mathematical problems is now essentially over, right? Still, there's more to life than that, right? There's still the task of human beings understanding these problems. Like one thing that the LLMs are very bad at right now, and might always be bad, is coming up with new problems. They can come up with small changes on the old problems, but the really profound questions that we get around to, LLMs are not that good at. And again, even in the limiting case where LLMs are just always better than us, there's still the task of human beings to understand what the LLMs have taught us, right? There's a difference between just being told the answer to a conjecture and really understanding it. If one was truly in it just for the joy of understanding, I don't think that's affected very much by the progress of computation whatsoever. Just because you can ask an LLM a question and it can give you the answer, you don't know the answer. You don't appreciate how it got there, not to mention the fact that it might be wrong. But we're putting aside... We're imagining a hypothetical future where it only gives correct answers. There's still work for you to do in your brain to work to understand that. The fact that computers are better than human beings at playing chess or playing Go hasn't stopped people from playing chess or playing Go. It might be that doing math becomes like playing chess, something that you know a computer might be better at than you, but you still do it because you think it's fun. I think that it often happens that when technology means that the old-fashioned way of human labor doing something has been replaced or improved by technology, what happens is not that the human element goes away, but it finds something different to do, right?

1:03:23.5 SC: My favorite analogy for AI is wristwatches, because it used to be that human beings made these mechanical devices that were incredibly ingenious, and if you had a wristwatch, you had an incredibly ingenious mechanical device on your wrist. And then we invented quartz timekeeping, and that was much better. I mean, a quartz watch is both more accurate and cheaper than a mechanical watch. And so this was called the quartz crisis in the watchmaking industry, and a lot of watchmakers either shut down or just started making quartz watches. But then they rediscovered that there was kind of a romantic art in a mechanical wristwatch, and they rebranded it as a luxury good. And now multiply the price by 10, and you can still sell watches to suckers like myself who think that this is actually quite romantic. So I think that what will happen is it will change. It will change in a dramatic way. Math is really the sort of prime target for this kind of thing. Physics and other sciences are less on the firing line, but they will eventually get there. I'm not fooled by that. But exactly because math is about proving theorems, this is much more in the wheelhouse of what computers are good at than physics or chemistry or biology. So math is the test case. I think that the human part of doing math might dramatically evolve, but I don't think it's gonna go away.

1:04:55.0 SC: Paul Cousins says, "You often say that the entropy of the universe was very low near the Big Bang. Entropy is usually formulated in classical probability, which I think is an emergent phenomenon. Even the classical probability appearing in the density matrix obtained by tracing out part of the universe is not fundamental. Therefore, I don't see how classical probability distribution over the space of possible universes is supposed to make sense. I'd be curious to hear how you reconcile that."

1:05:18.3 SC: So I think that the mistake is right where you say entropy is usually formulated in classical probability. The thing about entropy, and I've said it many times, is there's many different notions of entropy. You can't say, "Here's what entropy is. I'm not gonna listen to anybody else's thing," okay? Whenever you're dealing with entropy in a particular consideration, in a particular context, you have to say, "Okay, which version of entropy is the one appropriate to the question I am asking?" In the context of the arrow of time and the past hypothesis and the early universe, the appropriate notion of entropy is the Boltzmann entropy, the entropy that is written on Boltzmann's tombstone: S equals k log W. None of those symbols, S equals k log W, have anything to do with probability. They just don't. It has to do with coarse graining and macrostates. W is the volume of a macrostate. A macrostate is the set of all microscopic states of the theory with the property that they look the same according to some macroscopic observables. And of course, there's some judgment and people argue, and it's full employment for physicists and philosophers to say, "Well, what do you mean by a macrostate? What are the observables?" Et cetera, et cetera. But who cares? None of it has to do with probability whatsoever. It has to do with what can you observe about the system and how many microstates of the system are compatible with those observational outcomes. The important fact about the early universe is that given what we know about it, given the macroscopically observable features of it, whether that's at the surface of last scattering or wherever, you can count the number of microstates that would look like that, and it's very, very small, right? That's the low entropy of the early universe. That has nothing to do with probability. It's just a feature of physics, of cosmology, of what was going on back then. So whenever these entropy puzzles come up, the first thing you should ask yourself is, "Am I using the right definition of entropy?" Maybe that's the problem.

1:07:27.2 SC: Dominic German says, "Regarding last month's discussion on diversity in DEI, your position seemed to be that hiring based on skin color is an effective, albeit imperfect, method for increasing diversity in the workplace, and therefore DEI is a net positive. Is this a fair characterization of your view? If not, could you clarify? I appreciated your response and agreed with a lot of it, but DEI still feels like discrimination based on skin color. Shouldn't we want skin color to be as trivial a detail as hair color or eye color?"

1:07:56.1 SC: So for the first question, is this a fair characterization? Not exactly fair. You're getting at something. I do think that increasing diversity is a net positive, and therefore policies that increase diversity are good overall. That's not the only reason why I'm in favor of something like affirmative action or, in general, taking skin color into consideration. I do think that there are elements of fairness and equitability that we should indeed, to go to your last question, "Shouldn't we want skin color to be as trivial a detail as hair color or eye color?" We should want that, but we live in a world where that's not true. That's not even anywhere close to true. And recognizing that we don't live in such a world is part of why we need to nudge the scales a little bit to make it closer to that world.

1:08:45.4 SC: I once wrote a blog post about this called, "Spontaneous Social Symmetry Breaking," and it was an attempt to put forward an allegory, a metaphor for this situation. I noted that, look, in the standard model of particle physics, if you just look at the bare equations that tell you how things go, what we call the electron neutrino and what we call the electron are basically the same. They are symmetric. They obey a symmetry relation that relates them to each other. There's no difference between those two things. But it would be absurd to therefore say, "Well, we should treat the electron and the neutrino as the same because they're basically the same."

1:09:24.7 SC: The reason why is because we live in a world where that symmetry is broken. There is a Higgs boson in the background. And so in the real world, electrons have electric charge, neutrinos do not, et cetera. The idea being that you might want a world, you might think that the underlying moral universe treats people exactly symmetrically with respect to their skin color, but the actual world we live in doesn't. And therefore it's absurd to pretend that it does, especially if that pretending only happens for cases where after living in a world where skin color matters and matters and matters and matters, disadvantaging one group, suddenly people worry when it matters the other way, disadvantaging some other group. That's obviously incoherent.

1:10:15.3 SC: We live in a world where government agents are hunting people down on the streets based on their skin color. We live in a world where the Supreme Court of the United States has declared it constitutional to disenfranchise and disrepresent populations of people because of their skin color, and various states in the South have raced to do exactly that. We live in a world where the Department of Defense fires senior generals because of their skin color. We live in a world where the world's richest man uses a media platform to advance White nationalism because of people's skin color. We live in a world where there is ongoing discrimination in real estate, the job market, and elsewhere because of people's skin color. We live in a world where the only refugees being currently allowed into the United States are White people. Their skin color truly matters. We live in a world where there's been an unhinged reaction to the idea that actors in a movie based on Greek mythology have the wrong skin color. So if somehow all of this you can sort of live with, and you're like, "Yeah, that's okay," but then suddenly you get outraged when we try to hire people or let students into universities at a rate which is approximately reflective of the rates at which different people live in our society, then you need an update on your sense of fairness here, I think. I would love to live in a world where skin color is as trivial a detail as hair color or eye color, but I have my eyes open and I live in the world we actually live in, and I try to deal with that the best I can.

1:11:58.4 SC: Sam Herzog says, "The coincidence problem didn't strike me as a problem when you raised it in your dark energy episode. My intuition said, of course the most interesting dynamics in a system would appear around points of transition or inflection. I understand why it is problematic after listening further, but I wonder, is there anything to that intuition about transition points? Does it make sense to a priori assume that we would have appeared near the transition between too crowded, where cosmic events were wiping out fledgling life forms, and not crowded enough, where sparsity makes a confluence of life-sustaining conditions less likely to arise?"

1:12:33.6 SC: I think that there's some correct intuition going on here. As you will read in the upcoming emergence and complexity book, there are critical points of behaviors in certain physical systems where you get interesting behaviors on all scales, and that is absolutely the kind of thing that feeds into the birth of complexity and life and so forth. But you can't let these sort of general hand-wavy principles guide everything when you think about what matters. The coincidence problem is completely on board with the fact that interesting dynamics happens at points when it can happen, in this case the formation of planets and stars and life. But separately, there's something called the cosmological constant. Okay. The cosmological constant, in its natural way of being, would have been so big that those moments would have just lasted seconds or fractions of a second, and therefore there wouldn't have been enough time for all this complex stuff to happen. The anthropic answer to the coincidence problem is getting at what you're pointing at. It is to say that the vacuum energy needs to be small enough to allow for there to be time for all this interesting dynamics to occur. But if you don't believe in the anthropic answer to the cosmological constant, if you think that there is some formula that at the end of the day will tell you once and for all what the vacuum energy is, the fact that that formula gives you a value that allows for all this interesting dynamics to happen is highly, highly non-trivial, and we don't know why that would be the case.

1:14:11.7 SC: Ruka says, "As your podcast is the main reason why I give Many-Worlds a chance, here's the question that my mind keeps circling on. Are the worlds of Many-Worlds really unavoidable? I do have the feeling that Many-Worlds is onto something. It's just that I can't figure out if it's my bias that keeps pushing back or if that's a hint of a flaw of the theory."

1:14:29.7 SC: Well, they're not unavoidable. No. You can change the theory to get rid of them, right? They pop out of quantum mechanics. Quantum mechanics, which was a theory developed in order to fit the data back in the early years of the 1900s, came up with this idea that what you observe is not fundamental, that the way that we describe the world when you're not observing it is as a superposition of different possibilities, and therefore it allows for the superposition of different worlds. And then you just ask, do those worlds come into existence under the ordinary evolution of the universe? And the Schrödinger equation says, oh yes, they absolutely do. That's not unavoidable in any way. You just have to change the theory. It's unavoidable if that's your theory. If your theory is the quantum state represents reality, it obeys the Schrödinger equation all the time, and there's nothing else going on, then the worlds are unavoidable.

1:15:24.0 SC: Thomas Anderson says, "I've recently become interested in the use of contact force concepts in relation to social systems. For example, the relationship between educational goals and measuring outcomes is sometimes described as being in tension. Changes in policy can cause friction at implementation or are said to have impacts. Yet to me as a sociologist, the contact in question seems like it is different from a system like two cars crashing into one another. Quite a few social theorists have made use of these concepts, but I'm interested in the view of a physicist on it. Do social systems say anything interesting to you about the form that contact forces can take in physical systems?"

1:16:02.9 SC: My attitude towards all of this stuff is physics is special. Physics is special because it takes the attitude that you can simplify systems down to their absolute bare bones, their spherical cow descriptions, and you can analyze them in fiercely quantitative detail, and maybe you can learn something. And that learning can be surprising, right? You can be surprised at what your own theory predicts. Like Einstein never would have believed in black holes. Never did know about black holes, I should say. He might have believed in them eventually. The equations are smarter than you are. So therefore, physics is special. Therefore, it can be used to give you insight into other kinds of questions that are not necessarily physics questions, because the study of physics forces you to go outside of your intuition. It forces you to go outside of what you would have guessed, because the equations are saying this amazing thing is happening, and you can listen very closely to that. So therefore, because physics comes up with these new questions, these new concepts, because you're forced to, you can be inspired by those new concepts to think about other questions in new ways, whether it's tension or friction or whatever. These might be useful concepts for thinking about sociology or other kinds of social sciences, or they might not be. So I'm not saying that they need to be. I'm not saying that the map between their use in the social sciences and physics is necessarily exact or one-to-one and completely reliable. But use it to inspire you. Use it to start thinking about the problem in certain ways. That's what I would advocate doing. I would never take it overly seriously. Like if the data that you have in your social science investigation don't fit with a prediction that you made based on some physics-based model, then believe the data. Okay? Don't believe the model.

1:18:00.2 SC: Kyle Stevens says, "A recent paper by Anthropic, not the principle, but the AI lab, suggests that LLMs may develop some form of access consciousness, which emerges due to what they term the J-space. The J-space is just a small collection of internal neural patterns that, compared to all the model's other internal processing, play a special role and operate silently in the model's internal neural activations, allowing the model to think about a concept without writing it down. You have spoken at length about your beliefs that LLMs are not conscious, though you seem in general to be referring to phenomenal consciousness. Do you think that LLMs may have developed some form of access consciousness?"

1:18:39.1 SC: No, I do not. I think that's pretty clear. I mean, this is... The research being described is perfectly good. People at Anthropic and elsewhere have found that there's a way of talking about what is happening in the LLM when it's responding to a question and trying to come up with an answer, which you can sort of characterize as... Well, one aspect of it, I haven't read the paper in detail, I only know a little bit about it. One aspect is you can sort of project onto a vector that describes the emotional state of the system. Is it angry? Is it calm? Is it happy these days? And that gives you interesting, non-trivial information about what form the answer is going to take. It's not just you're doing this for fun, it's actually predictive of what's going to happen. There is a true fact about the statement, related to the statement, "Oh, my LLM is agitated right now." Okay? The words are doing all the lifting here. I mean, I hate to use this analogy because I think it's usually bad, but this is close to the analogy of the guy writing on a piece of paper, "I'm alive," and then putting that into the Xerox machine, and the Xerox machine spits out a piece of paper saying, "I'm alive." And the guy's like, "Oh my God, my machine is alive." I think that you need to do much more work to claim that this kind of behavior is anything like what you and I think of as consciousness, whether it be access or phenomenal consciousness. Maybe that work is doable, but again, given how LLMs work, given everything that they're trained to do, given how we design them, given how they behave, nothing in that makes me think that they're conscious or should be conscious. So I do think, and I've said before, I'll say again, I do think that we as a society, as academics doing philosophy and neuroscience and other kinds of things, have done a bad job of stating ahead of time what we mean by consciousness, so that if a computer actually achieved it, we would be able to explicitly say that that happened. But so be it, I think that that would be a very important project to do, but I don't think that LLMs are pushing us in that direction quite yet.

1:20:53.8 SC: John Plasterer says, "I've listened to many of your podcasts, including solo episodes and AMAs, but I can't recall hearing about your views on Zurek's Quantum Darwinism. It seems that his work is compatible or even builds on an Everettian Many-Worlds. Some recent work provides some potential ways of explaining how the Born rule can fall out of the records that form during decoherence. I'm interested in your thoughts on Quantum Darwinism and if you feel that this may be a promising avenue to advance our understanding of the fundamentals of quantum mechanics."

1:21:23.1 SC: So for those of you who aren't familiar, which is probably most of you, Wojciech Zurek, who is one of our big names in quantum mechanics and the foundations of quantum mechanics, decoherence and how decoherence works in particular. I'd be all in favor of Zurek winning the Nobel Prize for his work on decoherence. I think it's that important. His recent work, not within the last few months, but the last decade or so, has been under the label of Quantum Darwinism. And it's all about how decoherence works. He was a pioneer in saying that decoherence works, and these days he's working on the details, which makes perfect sense. Like, okay, so how exactly does it work? In particular, there's this fact that when you and I are both in the lab when Schrödinger's cat is revealed, when you open the box and there's a cat that is either awake or asleep, we talk to each other and we see the same thing. It's not like I had a 50/50 chance of seeing awake or asleep and you have a separate 50/50 chance.

1:22:20.9 SC: And one way of explaining that or accounting for that is that the information about which branch you're on, about what decoherence happened, exists redundantly in the environment. The reason why the cat is either awake or asleep on one branch or the other is because it becomes entangled with its environment. And it becomes entangled very richly with the environment, becomes entangled separately with, in particular, the photons that are heading out from the cat to your eyeballs. And so you can go through some math and you can show, oh yes, we're gonna see the same cat. So all of this is 100% compatible with Many-Worlds. For whatever reason, and I've talked to him about this, Zurek himself is a little bit reluctant to get on the Many-Worlds bandwagon. He's not necessarily against it, but he doesn't think you need to go that far. I just think that that's a failure of courage. You have to be willing to accept what your equations are telling you. So I think that the tools that he's built up, the technology, is extraordinarily useful for Many-Worlds. But I don't think that by itself, if you're not quite willing to say, "Okay, here is my foundational theory of quantum mechanics," that you're really gonna answer these questions like the measurement problem, the reality problem, et cetera. I think that he's trying to have his cake and eat it too. And I think that that's a very venerable tradition in quantum mechanics, of trying to have your cake and eat it too. But I think at the end of the day, you're gonna have to bite some bullets and either have the cake or, sorry, or eat it. Not and. You can't do both.

1:24:01.1 SC: Robin Luce says, "In your recent Mindscape episodes on vacuum energy and dark energy, you noted that a pure cosmological constant remains the simplest and most robust explanation for the accelerated expansion, while dynamical dark energy faces significant naturalness problems. Do you think mechanisms that could produce mild spatial or environmental dependence in the effective vacuum energy, for example, slight differences between overdense and underdense regions, are still theoretically interesting to explore, or do similar concerns make that direction unpromising as well?"

1:24:32.1 SC: Well, I'm not quite sure what that direction is. And this is a good question to address because it gets to how physics works. When you do physics, you have the fundamental construction of your theoretical model, which has, I don't know, some Lagrangian, some fields or whatever, some equations one way or the other. And then you have the phenomenology of the model. So given that underlying superstructure, I guess you can't have underlying superstructure, the underlying structure, here are what you would notice. There's changes from here to here, this is the scattering, and blah, blah, blah. The point is you can't start at the end. You can't just start with the phenomenology and ask about it without the theory, the structure underlying it. I mean, you're welcome to start with the phenomenology and say, "Okay, what kind of theory would get me this?" That happens all the time. But you can't just work in the world of the phenomenology without the world of the theory. So when you say mechanisms that could produce mild spatial or environmental dependence in the effective vacuum energy, what does that mean? How does that work? How does that happen? Like, if it's a scalar field, I understand how it happens and what to say about it, and that is exactly what people have been investigating. If it's something else, then whether or not it's viable, empirically fitting the data is gonna depend a lot on those underlying equations, the particular theoretical model. I mean, of course it's worth exploring because dark energy is something we don't completely understand. We had some expectations about it that didn't pan out. We have observations about it that are very well fit by a cosmological constant, but not perfectly well fit. So it makes all the sense in the world to keep an open mind. Meanwhile, my betting money is still gonna go on the cosmological constant being right.

1:26:22.8 SC: Marie Rascu says, "What are your thoughts on the reversibility of quantum systems through the lens of the Many-Worlds interpretation? I read that David Deutsch proposed reversing Schrödinger's cat by using quantum interference to show the cat's superposition splits into parallel universes rather than collapsing into a single reality. I kind of really wrap my head around it. Can you help me?"

1:26:44.7 SC: I think that this is a tricky one because Many-Worlds is based on the Schrödinger equation being precisely right, or some upgraded version thereof. And the Schrödinger equation is 100% reversible. Reversibility is built into the technology of Many-Worlds. It's right there. But there's in-principle reversibility and then there's in-practice reversibility. And the same thing is exactly true in classical statistical mechanics, right? Like in principle, if I mix cream into coffee and it gets all mixed up, I could very, very precisely move my spoon so as to unmix them in principle. But in practice, that's never gonna happen, and no experiment you're ever gonna do is gonna make it happen. So a good way of thinking about when the wave function of a system branches into two separate worlds is precisely when the decoherence has become irreversible in practice. In principle, it never is, but in practice that happens. People try to get understanding and make money off of the idea that I'm just gonna have a little bit of decoherence and then I'm gonna undo it. Like the whole delayed choice quantum eraser experiment is based on that. And you can do that. You can entangle something with a single qubit rather than a billion qubits in the environment, and then maybe you can reverse it. But then it really wasn't decoherence at all, I would argue. So I don't think that there's any special traction to be gained by imagining doing these experiments, because I would argue that precisely the regime in which you can reverse and undo the entanglement is precisely the regime in which you didn't truly have two separate worlds to begin with.

1:28:32.5 SC: Ivo Majors says, "Mathematics in the last few years has moved towards formal verification using proof assistant software such as Lean, L-E-A-N. One of the benefits of this approach is presented as making mathematics more accessible. Instead of relying on the authority of experts, one can check the proofs on their own. What do you make of this trend? Could a similar approach be useful in other fields such as physics, chemistry, and even social sciences?"

1:28:55.0 SC: Yeah, so for those of you who don't know, this is kind of a really cool thing that has been happening in math. I'm not an expert on it. Emily Riehl, who we had on the podcast a while back, is an expert, and I've learned a little bit about it from her. She's at Johns Hopkins also. The whole wonderful thing about the proof of a mathematical theorem is that it's supposed to be objectively true, right? It's not just a matter of judgment. It's yes or no. It's a formal system. You manipulate it, manipulate the symbols using various rules, and so it is tailor-made for being checkable on a computer. And so finally, people have made enough progress in actually doing that that it's become practically useful. And Lean, L-E-A-N, is one of the big software packages that people use for this. I think it's great. I think it's wonderful because there are these stories of theorems that were claimed to be proven and people even published proofs of them, but certain steps in the proofs were not understood by anyone except the author. And so even and the author dies, like eventually, and so even though these theories are in the books as proven, no one knows how to prove them. Okay? And that's bad. I think that that's not how math should be. So I'm very happy that we can upgrade to doing it in an accessible and checkable way. Is it gonna be useful in physics, chemistry, and social sciences? No, mostly no. Because mostly in those sciences we don't prove theorems. We're not starting from an agreed-upon set of axioms and seeing what their logical consequences are. We're proposing models, proposing ideas based on how well they fit the data and knowing that our credence in them is never gonna be exactly one or exactly zero. It's just a very different thing than is being done in mathematics. Which is precisely why, even though you didn't ask about this, it's precisely why I earlier said that math is exactly the area in which computers in general, and LLMs are an aspect of that, have the biggest prospect for making dramatic changes in how things are done in that field.

1:31:01.3 SC: Julian Voidel says, "Do you imagine Sisyphus happy?" So this question is referring to the story by Albert Camus, "The Myth of Sisyphus," and I actually reference it in The Big Picture. Camus tells the story of Sisyphus. Sisyphus is cursed by Zeus to push a rock up a hill all day, and then at the end of the day the rock falls down and he has to go back and push it back up again every day. That's his life forever. That's what life was like in Greek mythology. No mention in the telling what Sisyphus's skin color was. I don't know what that means for the future film to be made.

1:31:37.1 SC: I think that my views have changed on this a little bit. I used to say... I used to get it, I used to be pro-Camus on this. He was trying to make the point that Sisyphus can be happy if he makes happiness out of the situation that he's in, that it's not about the physical setup, it's about your inward mental attitude toward the physical setup. And existentialism, as we talked about with Skye Cleyrie and others here on the podcast, has this big emphasis on freedom and the freedom to choose things. And you might think, well, Sisyphus has no freedom whatsoever. Sisyphus is just pushing the rock up the hill over and over again. And Camus is saying, well, Sisyphus has the freedom to be happy about it, to take pride in a job well done of pushing the rock up the hill. And okay, I get that. Okay, I mean, I sort of get the argument, but I think that what is special about the real world is that we have a lot more freedom than that. Okay, almost all of us. Once we've lost our freedom by being put in jail or whatever, then we don't have it. But in our everyday lives, those of us who are not too constrained by the state or whatever have a lot of freedom to do things. There's an infinite number of actions one could take at any one moment of time, and it's kind of overwhelming. And I do think that that sort of down-to-earth kind of freedom is also very happy, so... Sorry, is also very, very important for happiness. So what I would currently say is I don't have to imagine Sisyphus being happy. Like, if Sisyphus is happy, if Sisyphus has the mental strength and fortitude to find happiness in the task he is forced to do over and over again, then good for him. I'm very much on his side and I approve and admire it. But if he's not happy about it, I totally get that too.

1:33:33.7 SC: Tyler Briggs says, "Throughout your career, have you ever been captivated by an open problem in pure mathematics, such as the Riemann hypothesis or the Collatz conjecture, to the point where you dedicated concerted effort toward exploring it, even if it sat outside your direct research in physics?"

1:33:49.4 SC: No, not really. I have done research, physics-oriented, that did sort of involve high-level math, but not even usually. Most of the physics I've done has sort of just borrowed math that already exists, which is what most theoretical physicists do. It's the rare people who are the Ed Wittens and Paul Diracs of the world that really have to invent new math for their physics. What I have calculated some homotopy groups in algebraic topology and done some work with free variable loop equations and things like that because there was some physics motivation. I've never been truly motivated just by a math problem. I think it's cool. Life is short. You gotta pick the problems you're interested in, and the math problems that interest me most are those that have some relevance to the physical world.

1:34:44.9 SC: Anonymous says, "In your recent two-part solo episodes, you briefly touched on the prospect of models that assume the gravitational constant G to be variable. Something you didn't seem to humor was assuming c to be variable. That is, that one imagined the speed of light to have been appreciably variable in the early universe. Could the acceleration of the universe then appear as a pure artifact of a systematic misinterpretation of deep space redshifts?"

1:35:09.0 SC: Well, it kind of doesn't work, the idea of having the speed of light be variable. It's an obvious question to ask, but very quickly you realize the speed of light is just a conversion factor from how we measure space to how we measure time and vice versa. It's a joke, but it's true: you can always set the speed of light equal to one, no matter what it is doing. So what does it even mean to let the speed of light be variable? Compared to what? How are you supposed to measure it? Even the Newton's gravitational constant, it is also a dimensionful quantity and you can set it equal to one if you want, but it only appears in one place in the equations of motion of physics, in Einstein's equation. And so you can sort of put a scalar field there where Newton's constant appears, and basically the variations in that scalar field have the same impact as it would to, quote-unquote, "let the gravitational constant vary." The speed of light appears all over the place in the equations of physics. So what you mean when you say... The best thing you could possibly mean when you say the speed of light is varying, is you're gonna put some field, some scalar field, precisely everywhere the speed of light appears and nowhere else, so that everything changes in lockstep, so it looks like everything is constant except the speed of light is changing. Now, you can do that. People have looked at that. João Magueijo in particular has done a lot of work on these things. They don't seem especially promising to me. Again, I say this a lot, there's a lot of great ideas out there. You can't pay attention to all of them. You have to sort of pick and choose what you're gonna work on. Something has to be promising in some way for me to say, like, okay, I'm gonna put enough effort in to really understand the details of this kind of proposal. To me, making the speed of light vary never seemed that promising in the first place.

1:37:07.2 SC: Jennifer Lecompte says, "What are the topics you enjoy teaching so much that you look forward to those topics each time you teach the class?"

1:37:15.7 SC: Oh, there are many of them. In fact, I think it's the opposite. Of the classes I teach, I would say that most of the topics are ones that I really, really like. In the physics classes, there are absolutely places where you hit something that you just need to do because it's your duty. If you're teaching general relativity, you have to teach the symmetries of the Riemann tensor, what it means to symmetrize and antisymmetrize and things like that. If you're teaching quantum mechanics, you need to remind them how to diagonalize a matrix or solve some special functions equations for Hermite polynomials, et cetera, like, this is tedious. Angular momentum, Clebsch-Gordan coefficients, there are topics that are tedious and you have to do them. In the philosophy classes, there are topics that are tedious for me in particular, and I have to do them. But I would say that 80% of the topics that I teach in a class are things I'm really enthusiastic to teach. You get to choose what you teach, right? So it's not completely arbitrary. It's not completely surprising that that's the case. I did... Last year when I taught quantum mechanics, I had a little bit of fun teaching them rigged Hilbert spaces, which is a technique for making sense of the fact that things like states of definite position aren't really part of Hilbert space in quantum mechanics. So there's this advanced mathematical technique to make it all make sense. Some of the students liked it, some of the students really didn't like it at all. They're like, "This is not gonna be on the exam. Why are you doing this?" So I have to alter my teaching strategies to fit the modern student mindset that they just want to know what's gonna be on the exam.

1:38:56.7 SC: Balachandran says, "I recently read that electrons in heavy atoms like gold move at relativistic speeds, which alters their chemistry and gives gold its color, for example. While I know electrons aren't literally tiny balls zipping around, I still struggle to visualize what velocity actually means for an unmeasured, spread-out wave function. How should we physically conceptualize things like velocity on the electron field when no measurement is taking place?"

1:39:20.1 SC: Well, I'm totally on your side on this. You'll never hear me say that electrons in heavy atoms are moving at relativistic speeds because to me, electrons are not little dots, right? They're wave functions, or they're part even of the wave function of the universe, even better. And in an atom, those wave functions are stationary. They're not evolving, okay? Nothing is moving. When people say that electrons in heavy atoms are moving at relativistic speeds, they have one of two things in mind. One thing, which is entirely disreputable, is to think of, well, if the electron had been classical, what speed would it be moving at? And the answer is close to the speed of light. Like, who cares about that? It's not classical in any way.

1:40:09.0 SC: The other is that if you were to do a measurement of the speed of the electron or the momentum of the electron, would it be relativistic? And maybe the answer is yes. But who cares? You're not doing a measurement on it, so what difference does that make? There is a respectable thing you can say. The respectable thing is not that electrons are moving at relativistic speeds. What the respectable thing is is that non-relativistic approximations don't work well anymore. So when you do the Schrödinger equation, as you know, if you listen to me talking in these podcasts a lot, there is sort of a big abstract version of the Schrödinger equation that is perfectly able to handle relativity and everything else. But there's also the simple non-relativistic version of the Schrödinger equation that Schrödinger himself wrote down in his very first paper, and that is applicable to the non-relativistic limit. That describes particles moving slowly compared to the speed of light, for example, in the classical limit. So when you have atomic orbitals in gold or other atoms, they can be outside the non-relativistic limit. So you have to take into account relativistic corrections to the shapes of the wave function and the energy levels of the wave function. So that's entirely respectable to say that, to say like you need to take relativity into account to get the energy spectrum correct for these atoms. That's true, but don't translate it into your head as saying that the electrons are moving close to the speed of light.

1:41:45.0 SC: Christoph Radomski says, "'It's just a theory,' some science deniers say, not understanding that theory is a well-established hypothesis backed up by strong evidence. I feel uneasy when scientists misuse this word when talking about hypotheses of various credibility. Unfortunately, you also did that at least once when relating to uncertain hypothesis. That's why I came up with some idea. Do you think it would be useful to introduce a hypothesis credibility level system? By this I mean some well-established and measurable system that would let us say this hypothesis is level 3 credible, which would actually mean the credibility of it is within some well-described range."

1:42:24.2 SC: So no, I do not think that's a good idea and I did not make a mistake when I was speaking myself. I think exactly the opposite of what you're saying here. I think that words mean their uses and you have to understand how words are used. Scientists mostly understand this, but there's certain non-scientists who think that there's this hierarchy of credibility between conjecture, model, theory, law, whatever. That's just not true. That's just not how the words are used, nor is it what they mean. All of these words are used essentially interchangeably by physicists, by scientists more broadly, and that's fine. In fact, very often you have things that are super-duper well-established, like the past hypothesis, which is 100% established, or as close to 100% as you can get, given a tiny little bit of credence otherwise. But it's still called a hypothesis, whereas Newton's laws of motion are known to be false because we know that classical mechanics isn't right. Bode's Law is the name given to this relationship between the distances of planets around the solar system, which is 100% crazy wrong, right? But it's still called a law. So that's okay. The correct answer to this question is not, let's change language so that the words mean different things, so that you're immediately telling people your level of credibility by what word you use. It's just to tell people what level of credibility the concept that you have actually deserves. You will notice, if you had tried to have some system for assigning credibilities to different hypotheses, that people don't agree on what the level of credibility is. So it's going to be impossible to actually make that an objectively statable fact.

1:44:20.6 SC: Isaac Osterber says, "Do you think it is better if a graduate student has a concrete set of topics and possible experiments when applying to a PhD, or is it better if they are more open to whatever work a research group has funding for? Does receiving a fellowship change that answer?"

1:44:35.2 SC: Well, this is an excellent question, a super important question to be asking if you're a prospective graduate student. I can't speak very directly for experimentalists because they live in a different world than I do in terms of funding and what needs to be done in the lab and so forth. But I actually, for whatever reason, I've been asked this question or questions like it by undergrads more generally recently. And so I have the following idea, which is that as an incoming graduate student, there's sort of a happy middle ground you should be looking for. It looks bad either to graduate school admissions committees or to prospective grad student advisors if they ask you what you want to do and you say, "I don't know. I'll be able to do anything." That looks bad. That looks like you haven't really thought about physics and you're not really passionate enough about any particular thing. But it also looks bad if you say, "I want to write this thesis. I want to write this paper. I want to investigate exactly this thing." That makes it look like you're not open to taking guidance from the department you're in or from your prospective advisor. An advisor is going to be looking for students who are passionate about a relatively specific area but willing to learn about what is going on in that area and adjust their research agenda accordingly. That's at least true for theorists. For experimentalists, they're generally going to have a very, very specific idea about what experiments need to be done, what needs to be done in the lab, et cetera. So I think that the best thing to say is, "I'm excited by what is going on in your lab and would like to be part of it."

1:46:16.9 SC: Lars Krueger says, "I'm fascinated by the idea of hidden sectors in particle physics. How likely do you think it is that dark matter is either one of multiple types of particle of a hidden sector that does not otherwise interact with normal matter?"

1:46:31.2 SC: Yeah, I think it's a fascinating idea. I'm a little bit on the fence. I've written papers about the idea. I wrote papers with various people, various collaborators about dark radiation, so charged dark matter, but the dark matter is charged under a new kind of electromagnetism that only interacts with dark matter, not with ordinary matter. So it's not ordinary electromagnetism, it's dark electromagnetism, and you have the dark photon, dark magnetic fields, the whole bit. You can even imagine dark atoms and things like that. I think it's worth exploring because it might be true, number one, and that's always a good reason to explore it. And number two, maybe by exploring it, you're led to realize that there are some new phenomena that you might not have otherwise imagined if you didn't think about it. And what we looked at were truly massless dark photons. Just so you know, if you hear the phrase "dark photon" out there in the literature, it's more commonly assigned to ideas where you have a photon-like particle that does have a mass. So I would call that something like a dark Z boson, but I'm not the emperor of physics. Not everyone listens to me there. Just so you know, there's no evidence that this is right.

1:47:46.6 SC: Some of these ideas are chased down because they lead to interesting experimental consequences. I think that's an excellent motivation for chasing them down because a theory investigated very carefully can let you realize there's an experiment you wouldn't otherwise have done. And if you do that experiment, you might see the thing you're looking for and therefore become famous and discover something new about the universe. And you have to do the theoretical work first. But if I were to place bets on whether or not the dark matter sector is actually complicated, I would bet that the chances are pretty small. And the reason for that is just because once you get complicated dark matter sectors, it's easy for it to be noticeable. It's easy for there to be interactions and couplings and different dynamics, which we might have very well noticed already. It's also easy to avoid them, don't get me wrong, but in the world of complicated dark matter models, many of them are already ruled out by the fact that we haven't noticed anything interesting about the dark matter dynamics. In yet other words, the dark matter acts exactly as if it would act, as far as we know, if it were just a single particle, no interactions, nothing interesting going on. So I think that's, again, the way to bet. We should change our credences if we find any data that points otherwise.

1:48:10.5 SC: Mike Bishop asks a priority question. It's been a while. We haven't had a priority question. But again, once per your life, you're able to ask a priority question. I'll do my best to answer it. He says, "Is your undying Patreon gratitude subject to Cantor's diagonal?"

1:49:28.6 SC: So I think that your priority question needs to be phrased a little bit more clearly because I'm not quite sure what you mean. What I am going to guess is meant by this is the word "undying," which maybe I've used, I'm not sure, implies a kind of infinity, right? If something truly never dies and time itself goes on forever, then the gratitude must go on forever. Cantor's diagonal is Georg Cantor's way of proving that there are different kinds of infinity, right? That there is the countable infinity, the uncountable infinity, and so forth. Basically, you can't put the number of, let us say, irrational numbers in one-to-one correspondence with the number of rational numbers because there are more irrational numbers than rational numbers. So no, [chuckle] I don't think that my gratitude is truly undying. I think that's a rhetorical flourish. Therefore, whether or not it is a countable infinity of gratitude or an uncountable infinity is not really an answerable question. Sorry about that.

1:50:30.6 SC: Henrik Jacobson says, "Let's say we find a strong unified theory where a lot of things fall into place neatly and perhaps with independent evidence speaking for it too. Let's say it's also predicting an inevitable and high probability of vacuum decay and therefore is dismissed. But with the many-worlds interpretation, could this theory survive? And would the evidence for it even lend credence to a many-worlds view? Any observer must, after all, be in a branch that is not yet decayed, no matter how thin those branches get." So I think this is a decent question that I'm not 100% sure I have a simple answer to, because my views on how to use anthropic reasoning have been changing a little bit. So I think that the short answer is it wouldn't be very good evidence. So just to back up and to help everyone understand what's going on here, it's not that hard to contemplate theories of physics where the vacuum state that you're in at any one moment is unstable. Basically, you can have a bubble of a different phase of the quantum vacuum that would have a lower energy. And there is a probability per unit space... I should say, per unit volume and per unit time that such a bubble spontaneously appears. And because it has lower energy, this is the true vacuum bubble, and our vacuum that we live in is the false vacuum. So that bubble of true vacuum would expand at nearly the speed of light. And if many bubbles appeared around us, they would take over and wipe us out, basically.

1:52:07.7 SC: So the idea is you're comparing two theories, one in which... You're comparing two different ways of thinking about quantum mechanics: a single-world theory, which says that that bubble nucleation is truly stochastic and just unpredictable, and a many-worlds theory, which says that really there's many different branches of the wave function, some of which have the bubble here and there and others of which don't. So would the fact that we live in the universe and have not yet been wiped out by a bubble of true vacuum be evidence for the many-worlds view? Because the probability of getting wiped out might be very high, but in the many-worlds view, there's at least somewhere in the universe where it hasn't happened yet. Okay. I think it's tricky. It's very tricky because in the same view, you want to compare the probability that someone like you could exist somewhere in the universe.

1:53:08.0 SC: What is true is that in the Many-Worlds version of this story, someone like you could definitely exist in one branch of the wave function of the universe. Whereas in the single-universe story, if there was really a high probability of all these bubbles being created and wiping us out, then there would be a very small probability that someone like you could exist. Fair enough. But there's a special wrinkle with quantum mechanics, right, which is that not all of the worlds count equally. If the amplitude for the world in which you could exist is very, very small, then in some sense only a tiny slice of you exists, or a version of you exists that doesn't count as much as a version of you that exists in just a single universe would count. If you don't believe that way of thinking, then you can't get the Born rule right in many-worlds quantum mechanics. So my inclination is to say that it's not evidence in favor of Many-Worlds, but I'm not 100% strong on that. I've not really thought it through all the way, so I think it's a little bit of an open question. By the way, I should say that the very fact that we're still here now is evidence that there is not a strong probability of the vacuum decaying. Indeed, 'cause if there were a strong probability, it would have decayed a long time ago. We're right on the boundary where if it's going to decay at all, it better decay soon. Otherwise, we're in a regime where the phase transition does not percolate, as we say. Even if bubbles of true vacuum come into existence, there's more and more ordinary false vacuum growing with time and you never completely complete that transition.

1:54:51.8 SC: Melita S. says, "When physicists say that particles are really excitations of quantum fields, what picture should someone actually have in their head? Is there a mental model that's accurate enough to build intuition without being misleading?"

1:55:05.3 SC: Maybe not. [chuckle] I think... I do talk about this kind of issue from time to time, and I think I have a pretty consistent take here, which is that it's super useful and helpful to have images, pictures, visualizations in your head when thinking about things going on in physics. But it's never the point. It's never the thing you should be aiming for, because ultimately the equations are what tell you what's really going to happen. And that's because pictures are a little bit floppy and manipulable, and equations are just clean and crisp and you can't argue with them. Sometimes in modern physics, you have things happening for which there's no picture, for which there's no easy way to visualize. When I say that the quantum state of a field theory consists of imagining the space of every possible profile the field might have and assigning a complex number to that, which is the amplitude for that profile, that shape of the field throughout all of space time, throughout all space, that's hard to visualize. You can have a picture in your head of a little two-dimensional field vibrating up and down and then a little arrow pointing to the complex numbers or something like that, but it's really not at all what's happening in this space we're trying to visualize. So if someone says a particle is a vibration or an excitation in a quantum field, if you want to think of a little vibration in a two-dimensional portrait of a field, go ahead. It's just to give you a warm and fuzzy feeling. It doesn't really help you build that much intuition.

1:56:50.0 SC: Aaron Fisher asks a priority question. He says, "Does the positron have its own field as distinct from the electron field? In Quanta and Fields, you explicitly write of the positron field, page 210, as separate from the electron field. But AI is very annoyingly confident that this isn't what Sean meant and that positrons and electrons, in fact, do exist in one field, the electron Dirac field."

1:57:17.5 SC: Well, that's the thing about the AIs, is that exactly where questions become slightly subtle is where they're the worst. They're not good at subtlety. They're very, very good at conveying information for which there's a complete consensus about what the right way of talking is. So the answer to the question is there is a single field that includes positrons and electrons, and in fact, it includes the spin-up part of the electron, the spin-down part of the electron, the spin-up part of the positron, and the spin-down part of the positron. This is the four-component Dirac field. Okay? But the components are different. They're separate things. And indeed, you don't need to use Dirac notation to write down fermionic fields. There are two-component versions of the things you can write down. But the more important thing is the positron, in a correct representation, in an appropriate representation of the Dirac spinors, the positron is represented by different components of that field than the electron is. It's like saying the x-component of the electric field and the y-component of the electric field. I can talk about them separately. I can talk about the X-component of the electric field. No one's stopping me from doing that. Or I could talk about the electric field as the combination of all three components, x, y, and z, right? No one is stopping you from doing it either way. When I talk about the positron field, I'm just talking about the components of the four-component Dirac spinor that represent the positron. There's literally nothing whatsoever wrong with doing that, and what I meant was what I said.

1:58:59.6 SC: Jeffrey Segal says, "In the last AMA, in responding to Nigel Benjamin's question about Laplace's demon and when a uranium atom undergoes fission, you said that in Many-Worlds, Laplace's demon would be able to say when the atom would undergo the fission. Could you explain this further? The probability of fission is a continuous function, and so in many-worlds, I would have thought that there would be constant splitting off of worlds as the fission occurs at various times in proportion to the probability function."

1:59:32.1 SC: Yeah, I'm not sure what I said. I never remember what I said. It's certainly not true that Laplace's demon, simply in a many-worlds context, knows when something that is a quantum uncertain event is going to happen. What Laplace's demon would know in this very, very hypothetical situation where Laplace's demon knows the entire wave function of the universe is the evolution of the wave function of the universe. So it knows the set of all possible worlds and the set of times when everything happens in every possible world. Okay? That's what Laplace's demon knows. So you can put that into whatever question you want to know. Whenever you have an infinite-dimensional Hilbert space, so you have something that is smooth and continuous and there really is an infinite number of worlds out there, counting the worlds is something that is not a well-defined kind of thing to do, and saying when exactly something happens is not a well-defined thing to do either. But this version of Laplace's demon would know the total set of everything that could happen in all of the worlds.

2:00:36.6 SC: Sonali says, "Could you please explain the equations predicting a cyclic universe proposed in your recent paper in simple words understandable by a non-physicist?"

2:00:47.1 SC: Well, the only equation is the Schrödinger equation, really. That's the point. And I can try to explain it. I can try to explain the relevant part of it. So the Schrödinger equation is the equation that says how a quantum state evolves with time. And it, like we said, it applies to everything: relativistic, non-relativistic, et cetera. The specific context in which we're looking at it in the recent paper is a finite-dimensional Hilbert space, and that is important for the details, but it's not really important today for answering this question. What's important is that the Schrödinger equation is linear. Now, I know you want a non-physicist explanation, but let me just use that jargon word of "linear" because it's super important here. What it means is that the Schrödinger equation just says you tell me the value of the wave function being operated on by this operator called the Hamiltonian, and that directly tells me the rate of change of the wave function. So linear means if I change the wave function by a little bit, the rate of change of the wave function also changes by a little bit. It's not non-linear where a little change leads to a huge effect. A little change leads to a little effect. That's the whole point of being linear. And the nice thing about linear equations is you can solve them exactly. Okay? And basically what you do is if you pick the right starting point, the technical thing to say would be an energy eigenstate, a state of definite energy, but the point is the right kind of state, you apply the Schrödinger equation to it, all it does is there's a little factor outside E to the i omega t. So E to the i omega t means that there is a complex phase that goes around in a circle in the complex plane. E to the i omega t is a complex number that always has length one. It could be 1, or it could be i, the square root of minus 1, or it could be minus 1, or it could be minus i, or it could be the circle on which all of those things live. The basic point is that if you solve the Schrödinger equation, every single state of definite energy doesn't change its shape. All it does is have a complex phase, as we call it, outside that goes in a circle at a definite frequency. And that's true for any system you want to look at in all of quantum mechanics. You can always solve the Schrödinger equation exactly by saying that in a state of definite energy, the only thing that ever happens is its complex phase outside changes at a constant rate.

2:03:27.2 SC: And what that means is that when you have a state that is not a state of definite energy, you can consider it a superposition of different states of definite energy. And they each separately have this complex phase that rotates, but not at the same rate. The rate of rotation of that complex phase depends on the energy of your state. So all of evolution in all of quantum mechanics comes about from interference, both destructive interference and constructive interference, between different parts of the wave function evolving at different rates. And so that's the equation that we have in our paper, the Schrödinger equation with a finite number of states with definite energy. They're all evolving at slightly different frequencies and they sort of beat against each other just like any other kind of wave can interfere or, yeah, can constructively or destructively interfere, I should say. And so that's the whole trick. The whole trick is there's a finite number of states out of which you build the wave function of the universe. They each just sit there rotating, and if you wait long enough, they will come back to the same relative phase that they had in the beginning. And we call that moment the Big Bang or the Big Bounce in our model. I have no idea whether that was understandable to the typical non-physicist, but it depends on what you want. If you want me to say an even simpler thing, I'll just say we solved the Schrödinger equation, and the Schrödinger equation in a finite-dimensional Hilbert space has the feature that you will come back to either exactly or very, very close to where you started. There's nowhere else you can go. You don't have an infinite amount of room to move, so you can solve that equation exactly and find yourself in a cyclic universe.

2:05:21.0 SC: T.C. Jansson says, "I've recently been looking into QCD, SU(3) symmetry, the eight gluons and so forth with some help of AI. SU(3), as I understand it, is a subgroup of U(3). Expanding the SU(3) group to U(3), I seem to be able to create a ninth gluon that is colorless and does not exist in nature. I'm curious about why nature seems to work with the SU(3) subgroup but not the bigger U(3) group and the implications of a ninth colorless gluon. Do we know of a reason or is it just because nature is this way?"

2:05:52.5 SC: Yeah, this is one of the questions that the five-year-old will ask you, like, "Why is the group of the standard model SU(3) cross SU(2) cross SU(1) rather than something else?" You're completely correct. U(3) is a bigger group than SU(3). It has nine generators which correspond to nine gluons in the particle physics way of doing it. But roughly speaking, you could separate it out. Okay? Roughly speaking, the single colorless gluon doesn't need to be related to the eight colored gluons. It wouldn't be confined inside quarks and hadrons or whatever. It could be there, but it's not in nature, like you said. So in some other more grandiose theories of particle physics, U(3) might be part of the overall answer, but there's a symmetry breaking that splits off that ninth gluon from the others. That's entirely conjectural and hypothetical. We don't really know. I think the right answer, the best answer, the most believable and reliable answer is, it's just because nature is that way. Sorry about that.

2:06:57.6 SC: Okay, I'm gonna group two questions together. One is from Mac Was a Great Cat, who says, "In a world where there are things that are brute facts, what's the chain that gets you from 'maybe the very low entropy of the universe is a brute fact' to 'but I think there's probably a reason for it and I'm going to investigate'?" And then James Allen says, "How do physicists know when they've run out of layers of the onion? What suggests that electrons are just electrons, but that protons are composite particles? If muons decay into electrons, what makes us confident that a muon is its own particle and not just an electron wearing a very heavy, electrically neutral winter coat?"

2:07:35.8 SC: I think that both these are perfectly good questions. Given some facts about the universe, you can relate it to the SU(3) question we just asked, given some facts you've noticed about the universe, how do you know that it's a brute fact versus where you go further? And the answer is there's no algorithm. There's no thing you can do to just simply answer that question, like, "Oh, yes, now we know this is the most fundamental level." That's never an answerable question. What you do is you do science to it, which means you construct different hypotheses and see how they work out. The fact that protons and neutrons are made of quarks and gluons and electrons don't seem to be made of any smaller particles, that has not escaped the attention of working physicists. Okay? When you say, "Well, maybe quarks are made of smaller particles, maybe electrons are made of smaller particles, maybe there's a relationship between electrons and muons of some sort," of course scientists have thought about this. They've worked really, really hard to try to come up with a theory that, number one, would have those properties and, number two, would fit the data. They have not been able to do it yet, and maybe they will tomorrow. You never know. But at the end of the day, that's the answer. The answer is, well, where can you go with this? What kinds of theories might have this be the case? What can you do with them? What predictions would they make? How would I know whether they were true or not? In the case of the past hypothesis, the question "Can I come up with a theory that explains it?" is, yeah, I can. I have done it myself, so certainly it can be done by people. My theory might not be the right one, but knowing whether or not it's on the right track is super important, so it's worth doing. This is a running theme through these AMAs and through my podcasts. Knowledge is not foundational. We don't just sort of find some facts that we know once and for all and build everything up from them. We invent hypotheses and explore them. That's how science is gonna work.

2:09:40.7 SC: Shibbs asks a priority question, "I have heard you describe more unlikely worlds, thinner branches, as not mattering as much. I don't know what your ontology of mattering is here, but is it, broadly speaking, the fact that there are just more of the likely versions and therefore we should care about those more? If so, would you consider this a fundamentally utilitarian approach to mattering?"

2:10:02.8 SC: Well, it's not that there are more likely versions. For example, if you have a wave function that is a superposition of two possibilities and one of them has an amplitude square root of 0.99 and the other has an amplitude square root of 0.01, there are only two possibilities. That's the entire trickiness of many-worlds, is that you can't get the counting that you need to get the Born rule right just from counting the numbers of worlds. The worlds don't count equally. They count as much as the amplitude squared. So 99% for this one, 1% for that one. So I use the word mattering just to be indicative of whatever matters to you. So it might be the probability that you find yourself in such a world. It might be how much energy such a world has. It might be the utility function that you have for such a world. Whatever you think is assigned to the word mattering goes along with the amplitude squared in many-worlds. I think that's a good motto to keep in mind to get you through figuring out how to think about this scenario.

2:11:17.2 SC: Cosmic Debris says, "My friend and I argue about basketball versus football, by which I mean soccer. He prefers to watch football as he argues that basketball depends too much on physical characteristics such as height versus things like intelligence or dedication. I claim that it is the same for football or chess for that matter, it is just less apparent than height. Do you think there is a difference?"

2:11:41.3 SC: Well, look, what I really think is it is silly to argue about what sport is better than what other sport. If someone likes a sport, good for them. If they don't like a sport, that's okay too, right? I think that you can do a little bit of talking yourself into appreciating different aspects of different sports. But who cares? Just enjoy it, just like it. And if someone else enjoys it differently, that's also fine. There's no question that basketball has the very specific situation where being tall is helpful, right? Being tall is just so noticeable in a way that other physical characteristics are not quite as immediately evident. For being a sumo wrestler, being heavy is helpful up to a certain point, right? For being a chess player, there's various mental faculties that are useful. But in all of these, it's a mixture of things. Basketball players need intelligence or dedication also. American football players, weirdly enough, offensive linemen are famous for needing to be super duper smart because there's a million different plays with many different moving pieces to all of them. I suspect that soccer/football also requires a whole bunch of physical characteristics. It's true that various soccer players who have been world-class good have been relatively tiny. They would not have been great basketball players. But still, you need to be in pretty good physical shape. And some of that comes from dedication, but not all of it. Some of it comes from your genetics. I think that's okay. You can enjoy playing the sport, you can enjoy being a spectator for the sport. Just enjoy it. You can try to understand why you enjoy it, like I said, but don't tell other people they shouldn't enjoy it as much or they should enjoy this other sport even more.

2:13:35.8 SC: Polina Vino says, "Can you discuss the relationship between chaos theory and complexity theory?"

2:13:42.2 SC: Yeah, I can. It's an interesting little... Slightly fraught relationship. They are both separate chapters in the upcoming book. There was this feeling for a while that complexity happens at the edge of chaos. So what does that mean? So let's put it this way. When you have chaos, chaos is super duper well-defined in a way that complexity is not, to be honest. We all kind of agree on what we mean by the word chaos. If you have some dynamical system, so that's a system that has some quantities that tell you what situation you're in, right? Maybe they're positions and momenta or something else, some variables that you're tracking. And you get chaos when you get a situation where, at least in some regime, there are cases where initially close trajectories deviate exponentially in time. Now, it's not just they deviate exponentially because the space they're moving through is very, very big. In fact, it's usually not that. It's that there is a bounded region of phase space, and so you can't just move off forever and ever. Like, for example, the planets in the solar system, right? The planets in the solar system only have a finite set of places they can be, but nevertheless, the dynamics is, technically speaking, chaotic. The time scale for the chaos to show up is very, very large, so it's not very visible.

2:15:08.7 SC: But in that small bounded phase space, the rate at which the trajectories move apart is e to the some constant times time, okay? And this is quantified by the Lyapunov exponents, and that constant in e to the constant times time is the Lyapunov exponent. So if that Lyapunov exponent is positive, then at least one of them is positive, then you have chaos. That's what it means. In at least some direction in the phase space or whatever it is, the things diverge exponentially in time. Now, typically what happens, and again, I'm not super up on the mathematical theorems here, so I don't know what implies what, but I can say typically when this happens, this chaotic behavior, you get an attractor in phase space. So there's a region in phase space that different trajectories are drawn to. The solar system, which I just used as an example and many people use, is actually a bad example here because it is a... That's just a classical conservative system. Very often you find chaotic behavior in systems that are open systems or sort of not classical isolated systems that are conservative.

2:16:21.3 SC: So in a classical conservative system like the solar system, there are no attractors. But in many chaotic systems, there are attractors, and they are what are called strange attractors. Strange attractors mean it's not just like a circle or a point; it is a fractal in phase space. And so there's a close connection between... But not a necessary connection, right? Just a typical connection between chaotic behavior and fractals. And fractals are sort of scale-free; something is happening at every length scale. And that kind of scale-free behavior is also characteristic of complex systems. You got that? So chaos is closely related to fractals, which is closely related to scale-free, which is closely related to complexity. I think that's the best you can do to draw a quantitative connection between chaos and complexity, because otherwise they're different kinds of things. Chaos is saying that two initially similar conditions will diverge; it's a statement about dynamical systems. Complexity is more about the idea that the system can sort of adapt... Well, there's different... Again, different kinds of complexity, right? Complex adaptive systems happen when you have the ability of a system to use information and to use that to adapt to the world around it. Arguably, the more general conception is when you have many little pieces coming together to make a big piece, but the little pieces are not individually designed like the parts in a car or a radio or whatever. They're somehow individually similar pieces. There's some homogeneity between the intrinsic nature of the pieces, but they begin to serve different functions in the complex system because of self-organization. So arguably, the ants in an ant colony or the neurons in your brain sort of started out similar but found different roles in supporting this complex emergent phenomena. So in that description, words like chaos or even dynamical system don't directly appear. So there is a relationship, but I wouldn't say that there is a close or very clear relationship for that matter.

2:18:37.4 SC: Cyclopropane says, "Why is information treated like a conserved property, like angular momentum or total energy? What makes physicists bristle when considering the idea that information that falls into a black hole is destroyed forever?"

2:18:52.4 SC: Well, again, as we were talking about before, the way that science works is you propose hypotheses and you see what works. Information conservation is treated like an important deal in physics because it is a characteristic of our best physical theories, right? It wasn't first suggested as a principle and then imbued or embodied in the theories. It's that first we invented classical mechanics, and then much later, Pierre-Simon Laplace noted that information is conserved in these theories. So black holes are things that appear in general relativity. General relativity is a theory that conserves information. Black hole evaporation happens because of quantum mechanics. Quantum mechanics is also a theory that conserves information, at least until you make a measurement and collapse the wave function. So the rough idea, and there's a lot of subtleties here that I won't get into right now, the rough idea is that general relativity and quantum mechanics individually conserve information. So therefore we at least expect that information should be conserved in quantum gravity as well. Again, it's not a derivation, it's not a theorem. It's an expectation that you can then go think about. Just like saying the low entropy of the early universe. I could just take it as a brute fact and live with it, or I could take it as a clue and try to explain it. Likewise, the apparent fact that if you're not too careful about it, information looks like it's lost in black hole evaporation might be a clue to learning something about quantum gravity. So that's what we treat it as.

2:20:21.7 SC: Niles Dar says, "I've always found it silly that Hollywood depicted aliens as looking like us, but the other day it occurred to me there might be something to it. How much of the human body plan do you think is dictated by physics versus evolutionary contingency? Given the importance of seeing, hearing, et cetera, if intelligence evolved independently on another Earth-like planet, would you expect something humanoid-esque?"

2:20:44.7 SC: I truly don't know the answer to this one. On the one hand, I am a big believer that there's a lot of variety out there in the world and we don't have a very good grasp on the space of possibilities. So aliens could be wildly, wildly different from us. On the other hand, I do respect the laws of physics and I have been very impressed by people pointing out that this or that feature of biological organisms had to be the way it is because physics makes it the most plausible way to do things. Maybe bilateral symmetry is an important thing. It's not universal in advanced life forms, but it's very, very common. Wouldn't be that surprised to see it even in aliens. But we truly don't know, I think, that kind of thing. The one very simple version example of this question is how big should aliens be? What size should they be? Right? One thing, even in science fiction tropes that try their best to make aliens different compared to human beings, they're usually around the same size, right? Maybe a little bit bigger. They're rarely smaller, actually, in science fiction. And maybe there's a reason for that given physics, right? Given the fact that volume and area evolve in different ways as you scale things up and down. But maybe the environment in which the aliens arrive is so radically different that the scaling is dependent in... Well, I guess the dependence on scaling is different in an important way. So right now, at this stage of my understanding, I would say I'm equally open and would be equally unsurprised if the aliens were really, really different or if they were vaguely humanoid in appearance.

2:22:37.0 SC: Ed Said Stuff says, "I recently heard about a paper from a couple years ago that derived gravity from entropy. What do you think about that idea? And if correct, doesn't that mean that quantum gravity is not necessary?"

2:22:48.0 SC: Well, there have been a bunch of ideas that relate gravity to entropy. I don't want to say derived, like how do you derive something from entropy? That's like saying you're deriving something from the number two, right? You might use the number two in your derivation somehow, but there better be some more steps in there to actually derive something like general relativity. I think the most famous and successful version is from Ted Jacobson, who has been working on what is called the Einstein equation of state. He points out that if you posit a relationship between entropy flux across a surface and the area of that surface, which is a very natural thing to do, you can derive Einstein's equation, right? And so I've written papers about this also. I think it's closely related to my whole program of deriving emergent spacetime from quantum mechanics and from entanglement. It certainly doesn't mean that quantum gravity is not necessary. You still have quantum mechanics and you still have gravity. I know that some people want to keep gravity classical. That just seems entirely barking up the wrong tree to me. Quantum mechanics seems to be the operating system on which the world runs, as far as we can tell. I want to see how it can support gravity. That's my personal preference moving forward. If I'm wrong, then I will change my mind.

2:24:07.5 SC: Joshua Hillerup says, "What's your sleep routine look like? Do you tend to read or something in bed first or try to go to sleep right away? Do you have your phone nearby or keep it out of reach? Do you tend to have problems with insomnia?"

2:24:19.4 SC: Well, I do tend to read and most of my in-bed reading is done with my iPad, actually, because then you don't need to have a light on, right? I do put it in dark mode and whatever, so it's not too bright, not too many photons. In fact, fewer photons hit my retina from reading on the iPad than I would get by keeping a light on in the bedroom. I used to have some trouble getting to sleep. Once I'm asleep, I almost never have trouble continuing to sleep. Maybe as one ages, it becomes a little more likely that one wakes up at 4:00 AM and has trouble getting back to sleep, possibly, that's a thing that can happen, but usually throughout my life it's been more like I have trouble falling asleep because my brain is racing because I'm thinking about something, right? So what I've learned is don't try to go to sleep until you're really tired. So that's why I will stay up reading or something like that until I'm really tired, so that when I do finally put the iPad down and the head hits the pillow, I fall right asleep. And that's actually worked pretty well for me.

2:25:20.7 SC: Sandro Stucki says, "Thank you for the excellent discussion with Bonnie Bassler. It made me think about a discussion I recently had with someone about agency. What criteria does a system need to fulfill to be considered an agent? Are bacteria agents? Are viruses? Can LLMs be considered to have agency? Are they more like a thermostat than, say, a dog? What's your take on this?"

2:25:41.8 SC: Well, I think this is a deep and difficult question. We discussed very similar questions in the recent episode with Christian List, and Christian, given his last name, had a list of different criteria that you might use. He was thinking about free will, but agency is very, very closely related to that. I don't remember what the list is, but it seemed very natural. He's sort of dealing with concepts like you have control over something in your environment based on decisions that you make, right? And of course, you need to define all these words that appear in the criteria, but that's okay. That's just something you have to do. So to me, at least a minimal requirement for being an agent is having, number one, some goals, some preferences, some things you would like to have happen in a sensible, philosophically coherent way, and the capacity to think about how to achieve those goals. Rather than just, a ball rolling down a hill is not an agent, despite the fact that you can talk about it as if it wants to roll down the hill, you really shouldn't talk about it that way. And the evidence is if you stop it from rolling, it won't fight against you, right? It won't just try to find a different way down the hill. It just obeys the laws of physics. Agents have this ability to conceptualize the world in such a way that when their route to achieving their goals is stymied, they can think about other possibilities, right? They can try other ways to do it. I don't think that bacteria, viruses, or LLMs or thermostats count as agents by those criteria. A dog does, right? A dog, you can see being stymied in its desire to achieve its goals and sort of thinking, not very cleverly, but at least thinking a little bit about how to do better. That's not very rigorous, but I do think that's a minimal part of the story.

2:27:40.2 SC: Michael Bright says, "On pages 61 and 62 of Quantum Fields, you discuss a particle decaying into two entangled particles moving in opposite directions, both detectable by the conservation of angular momentum. I am imagining lots of similar particles close to each other, also decaying into entangled particles flying off in opposite directions. If there were two double slits set up on opposite sides of some region of space, many of these newly entangled particles would move through them, creating interference patterns on the screen behind. If you were to put a detector on one of the double slits, thereby collapsing the interference pattern into particle-like patterns on the screen behind, would the screen on the other side also collapse instantaneously into a particle pattern even without a detector? If so, entanglement is even cooler than I thought."

2:28:26.6 SC: So the short answer to the question is yes. Let me try to explain more what's going on. I think it's a great question. It is pointing in the direction that entanglement is even cooler than you think. So in the ordinary double slit experiment, I send a single electron through it. And I've been talking to people and realizing that they don't always realize this. Just send one electron through the double slits, okay, at a time. It's not that you're sending many, but in quantum mechanics, an electron is described by a wave function that is a wave. And so if that's all you do, send one electron through the two slits, that wave will interfere on the other side because part of it goes through the left slit, part goes through the right slit, and both of those parts will wave up and down and interfere with each other. Now, when you observe the electron on the other side of the slits in some detection screen, all you ever see is a dot. So I want to clean up a little bit what Michael says about seeing a particle pattern. If you just send one electron, you always see a particle pattern. You see a dot, right? That's all you can ever see. That's why in most descriptions of the double slit experiment, what you do is you send one electron through at a time, but you end up sending lots of electrons through. So then you see many, many, many dots. And what those many dots reveal is the existence of an interference pattern. There are bands where there's lots of dots, where the two contributions from the left slit and the right slit constructively interfere. And there are empty spaces with no dots in them where there's destructive interference. I mean, that's hard to pull off in real life, but that's the perfect theory version of this.

2:30:04.3 SC: And then you know that if you wait close to the slits and have a detector there, so you can see which slit it goes through, you're collapsing the wave function. You're removing, if you see it go through the left slit, you're removing the part of the wave function that would have gone through the right. So on the other side, there's nothing to interfere with, and you no longer see interference patterns. You just see a bunch of dots. Okay? So what Michael's asking is if I have two... And the bit about two entangled particles, that whole discussion we just had was one particle not entangled with anything. If I have a single particle, like a Higgs boson, that decays into two particles, then what you know is that those two particles, and conservation of momentum tells you, they must be moving in precisely opposite directions, okay? But because it's quantum mechanics, you don't know which direction either one of them is moving in. Or even better, if I'm going to follow my own advice, the wave functions of those two particles don't move in definite directions, they move out in basically spherical waves, moving away from the initial decaying particle. So what you know is that when you measure the location or the momentum of either one of those particles, then you know that the direction the other particle moved and its momentum will be in exactly the other direction. And if you do that carefully enough, then what Michael says will absolutely come true. So you send out two particles. You don't know what direction either one is going in, but they're entangled, so you know that when you measure them, they will be back to back. And one of them hits a double slit experiment, and you detect which slit it goes through, the left one or the right one. And by doing that detection, because the two particles are entangled with each other, you are going to collapse the combined wave function of both of them. And so, since you've detected one particle going through one slit or the other, you know that the other one was moving in precisely the other direction. And therefore the other one is not going to go through two slits. It's going to go through one or the other, because you know exactly the direction it's going in if you did your experiment carefully enough. And therefore, even in the other side, even if it's very, very far away, you're not going to see an interference pattern for those particles because you didn't let that particle go through both slits at the same time. That's how it works. That's how entanglement works. And it is as cool as you can think of it being. And I think that's a good lesson to take home.

2:32:37.9 SC: Tim Gianitsos says, "Scott Aaronson's blog mentions an error discovered by Brent Werness in the original Coffee Automaton paper you co-authored with Scott. He said in 2015 that a new version of the paper would be posted on arXiv within a few months, but the version on arXiv does not list Werness as the co-author or address the error. Is there an updated version anywhere? And does this error invalidate your attempt to show that complexity rises and falls without additional assumptions that are inapplicable to our universe?"

2:33:05.4 SC: Yes, I've mentioned this before. This is just Scott and myself being very bad. There is a new version. It's not quite done, it's a few paragraphs away from being done, but we had to bring on yet another collaborator. But it is very close to being done. I've literally been working on it. It's gonna hopefully appear very soon. None of the conclusions spiritually change. For those of you who don't know, in this paper we looked at little two-dimensional models of cream mixing into coffee, and we measured both complexity and entropy. And you see entropy going up, and you see complexity going up and then down. But we made a mistake in the first version of the paper, and so one of the models where we thought complexity would go up and then go down, it just remained low the whole time. That's completely consistent with what we're saying. We're talking about when complexity can increase, does it increase in this particular way? And so we came up with an even better model, what we call the tectonic model for coffee and cream mixing into each other, and we did the calculation correctly, and the complexity does go up and then go down. So the conclusions of the paper don't change. It's just they are now being applied to different models. And so that's one of the reasons why we weren't working too hard to get the paper out, but we really should get it out. I'm glad that you asked the question because it's a little bit more motivation for us to finish that up.

2:34:35.1 SC: Eric Olavchen says, "What is your current view on how one should rationally form credences under self-locating uncertainty? Specifically, number one, do you still endorse something like Radford Neal's full non-indexical conditioning and the epistemic separability principle you put forward with Sebens? And if so, number two, how do you see these two principles as relating to your preferred answers in the standard Sleeping Beauty problem?"

2:34:56.2 SC: Well, so let's be clear here. My views have not really changed very much at all about that. I think the wider question of how one should do anthropic reasoning, my views are sort of changing a little bit, and I'm trying to get them exactly right in the form of a paper I'm writing with Isaac Wilkins. But the general points of view that I like, the full non-indexical conditioning, which means you should not close your eyes and pretend that you are a typical observer in the universe, that I still believe. The epistemic separability principle is a much more narrow idea that only really applies to Everettian quantum mechanics. And it says that whatever credence you put on being on one branch of the wave function versus another should not depend on changes that are being made very, very far away. Okay, if the wave function is basically the same locally but different globally, it should give you the same credences locally for whatever you're gonna measure when you do an experiment. So that's much narrower in its application than fully non-indexical conditioning.

2:36:01.3 SC: Everett, as we said just before, is a special case that requires extra care when you think about self-locating uncertainty because you can't just count. I'm a believer that when you have two observers who are more or less indistinguishable from each other in a single universe, then they should be given equal credences if you don't know which one of them you are. Except if the universe you're talking about is the Everettian multiverse, because there the universes are not created equal. The standard Sleeping Beauty problem, I really just think is a question of defining things carefully. I think you have to operationalize it. Are you imagining some betting setup where you're gonna bet after every awakening, in which case you should probably be a thirder? Are you imagining some more ontological setup about the actual probability of a coin coming up heads or tails? Maybe you could be a halfer. So I think that you need to tell me exactly, exactly, exactly what you mean by the standard Sleeping Beauty problem before I give you an answer to it.

2:37:04.2 SC: Sean Sullivan says, "I've just finished reading Stephen Hawking's A Brief History of Time and noted that he and Gregory Laflamme believe that the arrow of time would not reverse in a contracting phase of the universe. How does this view differ from your cyclic interpretation? In particular, the paper stated that you have no idea why the universe would possess the low-entropy starting point. Do you have any educated guesses or gut feelings why it may be this way?"

2:37:29.3 SC: Well, Hawking's statements about entropy increasing or decreasing are in the context of a certain model. His model of the wave function of the universe, the Hartle-Hawking wave function, which has a solution to the Wheeler-DeWitt equation and emergent time and blah, blah, blah, blah, blah. Our proposed cyclic model is in a completely different model where you have a finite-dimensional Hilbert space, you don't have the Wheeler-DeWitt equation, time is fundamental, blah, blah, blah. Okay? So they're just two very different setups, and one might be true or the other might be true. What I gotta keep saying about the recent paper about the cyclic universe is we're not putting it forward as the way the universe is likely to behave. What we're doing is we're pointing out a loophole in something that other people have thought about for a long time: quantum cosmology with a finite-dimensional Hilbert space. Okay, ordinarily I would have said that such a scenario is just ruled out by the Boltzmann brain problem. No way out. And in our paper, we point out that there is a way out of the Boltzmann brain problem in that particular situation. It requires a bunch of fine-tuning and a bunch of sort of special pleading, and things gotta work out exactly right, which is why I'm not suggesting that it's the most likely way the universe might be. It's a possible way the universe might be, and it's a very interestingly different way the universe might be than most other people's theories of the universe. So I think it's worth paying attention to because its credence is not zero, but small, but the consequences would be very, very big.

2:39:08.5 SC: Dale Addison says, "What is your read on the Kawhi Leonard trade situation? How much, if at all, do you think the ongoing investigation factored into his decision to return to Toronto?"

2:39:17.7 SC: So again, for those of you who don't know, Kawhi Leonard, super good basketball player in the NBA, has been playing recently for the Los Angeles Clippers, but recently got traded to a team that he used to play for, the Toronto Raptors. The complication is that Kawhi's agent, who I think is his uncle or something like that, is kind of a notoriously shady dealer. And the owner of the Los Angeles Clippers, Steve Ballmer, who made his money with Microsoft, is also kind of known to shade a deal here or there himself. And evidence has come to light that they, some version of the word "they", came up with ways to circumvent the salary cap in the NBA. So the National Basketball Association has very strict rules on the total amount of money that can be paid to players by a team. And it's obvious that one possible way to circumvent that is for the owner of the team, if they're filthy rich, Ballmer is the richest owner in the NBA, which is saying something, to just say, "Okay, your salary is just this certain amount, but my friends or my companies or my shell companies or whatever will hire you to do something not exceptionally onerous and will pay you on the side under the table to do that."

2:40:41.2 SC: This is apparently exactly what happened with Kawhi Leonard, as has been discovered by Pablo Torre, who is a podcaster, sports journalist, who won a Pulitzer Prize, in fact, for audio reporting or something like that. Anyway, they found documents that show pretty darn clearly, as far as I can tell, that Steve Ballmer and/or his friends were paying lots of extra money to Kawhi Leonard for doing nothing. And so this was a year ago, and it's kind of embarrassing that the NBA started an investigation and still hasn't finished the investigation. So I don't know what to think. I mean, I don't know whether that investigation factored into the decision to return to Toronto or anything like that. It's not even at all clear if the NBA does do some kind of punishment, whether the punishment will primarily come down on Kawhi Leonard or on the Los Angeles Clippers. I actually think that the Clippers are the ones who deserve to be punished. They're the ones circumventing the salary cap. If the player and their agent can get them money, then maybe it's naughty, but it's not the worst thing in the world, whereas if the team is offering ways to circumvent the salary cap, I think that's just very bad.

2:41:57.7 SC: Stevie CPW says, "If the Pope agrees to be a Mindscape guest, what are some of the questions you would like to ask him?"

2:42:04.4 SC: I haven't really thought about that. I think the chances are small that the Pope is gonna be a Mindscape guest. You never know. I'd still like for it to happen. I think that for someone like the Pope, Pope Leo, what I would want to talk about is the relationship between ontology and morality. Okay? So basically, the Pope seems like a good guy, as far as I can tell, but he's also the leader of the Catholic Church. And the Catholic Church as an institution has some beliefs that I think are pretty horrible. And he has to go along with those beliefs because he's the Pope. Okay? So they have horrible opinions about gay people and transgender people. They even have horrible opinions about women, right? Women aren't allowed to be priests in the Catholic Church, which is just so obviously a relic of ancient human beliefs rather than the word of God. I would like to know, how much of your beliefs about this stuff do you think comes from your innate humanity and your rational thought versus the doctrine of the church? And I would also, of course, like to know how he thinks about the relationship between science and religion and things like that. I think a lot of people who are both good scientists and religious have this sort of compartmentalized in their brain where they say, "Well, science and religion are two different things. I don't need to fit them together." Other people have this feeling, the sort of natural theology kind of perspective that, "No, no, doing science helps my religious beliefs." And so I would like to know which side the Pope comes down on on that. I think there's lots to talk about, honestly. But since I have to do a single podcast every single week, I don't spend too much time thinking about questions for people who are not the next guest that I'm gonna be talking to.

2:43:52.5 SC: Peter van Vila says, "You once stated that because of the winner-take-all system, some people can go unrepresented for a lifetime if they happen to live in a predominantly blue or red county. In Europe, I'm mostly happy with the proportional distribution that we have after elections. At best, it allows for substantial switches in policy between legislatures and to several parties having to meet halfway in order to obtain a majority. At worst, it can lead to negotiations dragging on for months without effective government. A byproduct is the emergence of single-issue parties. While most elections are fought between traditional left-right factions, social or liberal or religious, what also exists is, for example, the Animal Party for animal rights, 50PLUS for senior citizens, the Pirate Party against copyright, the Farmer Party, and so forth. In the unlikely event that the United States should abandon winner-take-all system, what types of possible ideological or new political movements would you expect or hope to see emerge from the ranks of the Democrats and Republicans that up to now remain underrepresented in the mainstream consensus?"

2:44:55.8 SC: Well, I mean, that's a good question. You're asking a practical question that I'm really not best equipped to address. But it's an important one. So for those of you who don't know, proportional representation is the following idea. In a usual United States election, we have some geographical region like a state or a congressional district, and its majority wins, right? There's two candidates, or maybe there's three candidates. But the US has a constitutional system that strongly favors a two-party system, so it's usually two candidates. And then you vote, and whoever gets more votes wins. It's actually not majority rule; it's plurality rule, because if there are three candidates, then whoever gets the most votes wins, even if they don't get completely a majority. And the problem with that, obviously, I mean, you can imagine a situation, it's a little bit contrived, but imagine that in all 50 states, 49% of the electorate was in one party and 51% was in the other party. Then basically the country is split almost 50/50, 51/49, but the legislature you end up electing is 100% from the 51% party, right? So this kind of mechanism does not result in a fair representation of the actual opinions of the electorate.

2:46:18.0 SC: Proportional representation, which can happen in different ways, but the most straightforward way is some sort of party-based proportional representation where in each geographical area you vote for a party. Americans hate this idea because Americans don't like political parties, and they've been trained against them since the Federalist Papers. But in other countries, they don't have that visceral anti-party reaction. So you say, "I'm gonna vote for the Democrats, the Republicans," or whatever. And there might be... There will generally be more than one representative from each geographical area. So if you have, let's say, 10 representatives from a certain geographical area and 40% of the vote goes to one party and 60% goes to the other one, then one party gets four representatives and one gets six. Okay? And the political parties then decide, and usually they will advertise this ahead of time, but in principle it's their choice," However many votes we get, here are the individual people we will give the seats to." Okay?

2:47:24.1 SC: So, a lot more power in the hands of political parties in that kind of situation, but also much closer representation to what the actual people want in principle. So in general, it's a better system. Political scientists at least think that this is a much better system than plurality vote or first past the post, as it is sometimes called. Now, there are downsides. These weird little parties can absolutely get represented in the legislature. And I don't, to be honest, I don't understand why that's true. And this is why I'm reluctant to actually give you a direct answer to the question. I can absolutely understand that, let's say, you think that animal rights is the single most important thing. So you think that you're gonna vote for the Animal Rights Party. But the obvious problem with that is that most votes that a legislature will take are not about animal rights. Okay? So even if you think that animal rights are the single most important issue, you should be voting for a party that agrees with you on that issue and hopefully also agrees with you on other issues, 'cause the other issues are gonna come up. I mean, I would think that there would be some other party that would say, "We're in favor of animal rights and also we have all these other elements of our platform," and that party would gain more votes. So I'm not quite sure how that... Why it is true that this kind of voting system gives rise to these tiny single-issue parties. I do think that simply invoking the magical words "proportional representation" doesn't say that much by itself. There's different ways of doing proportional representation. There's the party, like I said; there's other ways of implementing it, though. So there might be some ways of doing it that could get rid of that problem or help it out. This is something that now you've exactly gone past anything that I know very well. So I'm thinking about it a lot because we're thinking about... It's a coarse-graining question, right, from the physics point of view. Just like you are coarse-graining the positions and velocities of atoms to get temperature and density and things like that, you're taking a microscopic description, which is the preferences of every single voter, and you're coarse-graining them to a macroscopic description, which are the people representing you in the legislature. How best should you do that is an excellent question that I'm thinking about, but I won't promise that I actually have all the right answers to that right now.

2:49:52.8 SC: Ankar Sadakar says, "I recently started my first postdoc after finishing a PhD. Do you have any advice for how to smoothly transition from being responsible for doing the groundwork, like cleaning data, running simulations, to assuming a more mentor-like role in a team, giving feedback, suggesting ideas? I still feel the urge to do all the things by myself, but given my increased responsibilities, it's not always possible."

2:50:17.2 SC: Yeah, I know this is gonna be another thing where different, very, very specific conditions are different. I was never someone mostly cleaning data, running simulations. I was coming up with ideas myself even as a grad student, just not because I'm extra special, just because that's the kind of physics that I do. That's my job. It's mostly coming up with ideas and then calculating the consequences of them. I would say, though, that you're right to have the intuition that assuming a more mentor-like role is part of your job as a postdoc, and it's also just a good thing to do. But even more than that, I would say keep in mind that there's this sort of lagging education that you get, as I like to think about it. The things we want you to do as a postdoc are not the things we train you to do as a grad student. And more importantly, now that you're already a postdoc, the things that we train you to do as a postdoc are not what we want you to do once you become a faculty member. So I would not only think about what I need to do to be a great postdoc, although there's that, you should also think forward to what will I need to do to get a faculty job and then become a good faculty member. And so it's not just being a mentor to those under you, but really being an intellectual leader about what projects should we be working on? How can we make progress on them? The team as a whole might have some people who make progress faster than others. How can we pick them up, the ones who are lagging behind a little bit? How can we apply for money, apply for grants? What conferences should we go to? What speakers should we bring in? Who should we be collaborating with? All of these faculty-level questions you might as well start thinking about right now. I have no easy advice on how to do it well. These are all things that you gotta get in there and do to get better at. So practice makes perfect in this case.

2:52:23.8 SC: David Kudaverdian says, "Is it a sensible question to ask what the shape of a photon is, say one emitted from an atom? For example, I've heard a lot about the shape of a hydrogen atom's electron orbitals, s, p, d, f, et cetera, but I've never heard anything similar about an emitted photon."

2:52:41.9 SC: Well, the reason you have not heard that is because those electron orbitals are confined, not confined in the sense of quark confinement, but they are bound states in the atom. So they trail off to infinity, but they trail off when you get far away fairly quickly. So in a very realistic sense, those orbitals are sort of stuck inside some region close to the nucleus, whereas a photon isn't stuck anywhere. You can put it in a cavity, right? You can sort of seal it up, a single photon, and then there's something called cavity QED. And then you can talk about the shape of... It's not good to talk about the shape of a photon. Talk about the shape of the wave function of the photon. That's what the orbitals are for the electron. So in empty space, the shape of either a photon or electron wave functions can be anything at all. They're typically approximated by wave packets or plane waves or spherical waves. Generally, if you have something like a photon that is emitted from an atom, roughly speaking, it depends on details about angular momentum and so forth, but roughly speaking, it will spread out in a spherical wave. It will spread out in all directions. And that's the closest thing that you can come to saying what is the shape of the photon. Which means that if you are observing a photon emitted from the sun, then suddenly there was a wave function that spread out, I don't know, 93 million miles in radius around the sun that you just collapsed just by looking at it. Now, in the real world, probably that photon bounced off various things in the atmosphere, so it's not quite true, but in principle, things like that can really happen.

2:54:26.2 SC: Henry Jacobs says, "I've been working on a project to help grow activist movements in the likely event that certain people decide to meddle in the midterm elections. However, when demoing the project to friends, I don't get pushback on the project itself, but instead on the very idea that activism is effective. Of course I think it's effective, but there seems to be a rise in skepticism after the No Kings protests came up empty. Do you care to comment?"

2:54:50.3 SC: Yeah, I think that people are impatient and they shouldn't be. I mean, what is the purpose of skepticism or the reasoning behind the skepticism after No Kings protests? We don't have a king. It worked, right? But I think more importantly, there are very slow-moving variables describing the global opinion of people in a country or in the world that don't just change overnight, but do get affected by the existence of demonstrations and movements and things like that. I think that historically demonstrations and activist groups have been absolutely crucial in making things happen. There is the George Bernard Shaw quote about how it's true that reasonable people know you can't change the world, therefore all change comes from unreasonable people. And I think that in a certain sense that's true. A lot of people will in the world, we have a lot to do. We have to worry about our own stuff, about earning a living, about our own families and whatever. And it can be hard to devote time and energy and money to movements or efforts that don't get immediate payoffs, but they're crucially important. So people like you, Henry, who are doing this are just absolutely central. And if some of your friends don't get it, just keep asking other people, look for other friends. To me it's like having sex and then the next morning you say, "Well, I don't see any babies here. I guess it didn't work." Sometimes it takes time for things to pay off. You have to lay the groundwork well in advance and it will eventually have a good outcome.

2:56:34.2 SC: David Maxwell says, "Are you excited about the medium-term Moon plans? Not just going there, but establishing a presence and developing a way to explore the solar system more effectively?"

2:56:44.9 SC: I wish I could be more excited than I am. I'm someone who is a believer in exploring outside the Earth, whether it's the Moon or Mars or elsewhere in the solar system, et cetera. I think that very, very long-term that's gonna happen, it's gonna be super important, and there's a lot of benefits to it. It's also super-duper hard. And I think that there's a lot of people who are just closing their minds and covering their eyes about how hard it is and they're sort of not really realistic about it. And worst of all, the whole thing has become kind of politicized in ways that are unpleasant. The discourse around exploring the solar system is being dominated by a bunch of people who don't know what they're talking about, don't want to do it for the right reasons, and are unreliable in their predictions and their understanding of what exactly is going on. So I'm both in favor of doing it, the Moon and elsewhere, but also not at all trusting that it's gonna be done well. The federal government is the most likely place to get people having plans for doing something with the Moon because there's not gonna be any immediate payoff. Even if it's private corporations that do it, they are gonna need to be heavily, heavily subsidized by the government. And I just don't have any faith in our current federal government to make good decisions about this. So I like the idea, but I'm kind of very lukewarm about the actual implementation in the near term.

2:58:17.5 SC: Chris Mason says, "In your opinion, what makes a good research question?"

2:58:23.1 SC: Well, that's a good question in it of itself, and it's gonna be one where very different answers are perfectly reasonable. There's different kinds of good research questions. Some research questions involve embarking on decade-long projects, building the right experiment and chasing things down and being very patient and assembling a team. Others are things you can do in a weekend, right? So there's not gonna be any one simple description of what makes a good research question. I think that a lot of factors come in. Better than saying what is a good research question, a better thing to ask is sort of what should I keep in mind when contemplating whether a research question is good or not? The very down-to-earth, almost too simple to be dwelled upon things are, it has to be a question that other people want to know the answer to. You're always losing if you're... Because I do this all the time, this is how I know, if your research question you're very fascinated by, but nobody else cares, right? If you say the answer to it, they're like, "Oh yeah, I knew that," or "I'm not really sure why that's interesting." When I wrote this paper with Chris Shalu last year about what happens, what do you see, what does Hawking radiation look like when you fall into a black hole? And I've been fascinated by this question for decades, and I knew that it would require an enormous amount of very technical calculation, and I didn't quite have the stomach for it. And eventually Chris did, and he did it, and he deserves most of the credit for that. And we found a result that makes perfect sense and is kind of intuitively clear after you say it.

3:00:03.5 SC: But look, I know that I've asked this question and talked about it to many, many very smart theoretical physicists before we wrote our paper, and none of them gave me the right answer. And then once we came up with the right answer, I explained it to very good theoretical physicists, and they're like, "Yeah, I guess that makes sense. I probably would have said that." No, you wouldn't have. But okay. If you want a question that actually gets people excited, it has to be one that people recognize as an interesting question ahead of time. And then it has to be one that you can do, right? A lot of questions sound good, but you just don't quite have the either ideas or the technology or whatever to answer them effectively. So that's very, very simple-minded. The third quality is that you better be interested in it yourself, right? So you better think it's interesting, the world better think it's interesting, and you better have the wherewithal to actually do it. Those are hard to find. It's easy to find questions we don't know the answer to. What happens very, very, very, very often when you are a young scientist is that it's easy for you to find questions that you can answer. This is more of a problem than you might think. When you're an undergrad, typically there aren't any questions that you can come up with that you instantly know how to answer, otherwise you'd be writing research papers. But once you've got a PhD, now you're trained in one little area on how to get answers to certain questions. And it can be way too tempting to simply spend time and effort on questions because you can answer them rather than because they're interesting. So I think finding that dividing line is really, really important. I specifically, explicitly shifted my research focus 10 or so years ago from questions I could answer about scalar fields and dark energy and whatever to harder questions that I think are more important, that are more foundational. And other people might not think they're important. That's okay. I'm old enough to just focus on what I think is interesting. But young people gotta take the rest of the world into consideration.

3:02:13.0 SC: The Great Deceiver says, "Could an event be considered a complex system? Are there any theories on that out there? I went down the JFK hole a few years ago, and I kept thinking this has to be one of the most studied events ever. And now you've brought some of the ideas of complex system sciences to me, so it seems like a natural question. Can an event be studied as a complex system?"

3:02:36.1 SC: Well, I think the traditional answer would be no, in the following sense. A complex system is something that itself has an evolution over time. Like you can ask questions like, "If I perturb the system in a certain way, how will it respond?" Right? That sort of dynamical question is central to what we think of when we talk about complex systems. Events, and I know that what you mean are the sort of everyday meaning of the word event, not the general relativity meaning of the word event, which is a point in spacetime. Points in spacetime are not complex systems. You mean a happening in the world, something that occurs through the buildup of some set of occurrences with some causal structure. The thing is that the event as a whole, like, even if all sorts of conspiracies are going on and all sorts of moving parts, but the whole thing plays out, it's done. Right? It's there. It's not changing. You can't perturb that system and ask how it's going to change with time because it's already in the past. So it might involve complex things going on, but it is not quite the traditional notion of a complex system. I do believe that there's many sort of different notions of complex systems, and maybe system isn't what you want. Maybe you can talk about the complexity of an event even though it's not a system, right? You can apply the idea of complexity to many different things: to strings, to images, to questions in computer science or whatever. So I don't know if anyone has studied sort of a way of quantifying the complexity of a process that has led up to an event, but maybe they could do that.

3:04:17.1 SC: Ken Wolf says, "In your talk with Marc Berman, he made a distinction between directed or top-down attention and involuntary or bottom-up attention. I couldn't help but thinking of this as a distinction between things that are productive and things that make us happy. In the modern world dominated by economic thinking, we like to think those things go together, but this distinction seems to suggest they are inversely related. We give up the game when we speak of having a happy, productive life, which implies the two things are distinct. Does this suggest that there is an inverse relation between productivity and happiness, or am I reading too much into the distinction?"

3:04:51.6 SC: Well, I think you're reading too much into the distinction in this case. I would argue that there are some things that require directed, top-down attention and can make you productive that also make you happy. It happens to me all the time. Right? Here I am making a podcast, and if making the podcast didn't make me happy, I wouldn't be doing it. There's other things, like earning your beloved Patreon support that helps me do the podcast, but I wouldn't do it if it didn't make me happy in some way, and it absolutely requires some attention. Okay? I don't think that's the right distinction. I don't think that's the distinction that Marc was trying to make. There might be some way of thinking about these things which almost wants to sort of denigrate top-down, directed attention, like you're working too hard, you should just live life in the flow, in the moment from time to time. I don't believe any of that. I'm sorry. I think that's part of living a happy life, but if that's all you did, I wouldn't be happy at all living life that way. I think that happiness is much more about finding the right mix of when you're giving directed, top-down attention, when you're giving involuntary attention, and when you're not giving attention at all. Okay? When you have those different kinds of things going on in your life. Other people will find happiness elsewhere. Some of them might find happiness purely in this involuntary, in-the-zone kind of thing. Others are only gonna get happiness out of directed, top-down attention. And I think all that's fine. Let people be happy in the way that they find happiness.

3:06:36.5 SC: Peter Solfest says, "In the last AMA, you said that to an outside observer, the particles across the event horizon can have negative energy, which is how Hawking radiation shrinks a black hole. Would an observer inside the event horizon see a positive energy particle, or is the negative energy particle just a mathematical artifact used to describe the source of Hawking radiation?"

3:06:57.1 SC: Yeah, the people inside the event horizon would see a positive energy particle. The whole point of the discussion is people in different reference frames assign different energies to particles. That's the least surprising thing in the world, right? Even in special relativity, even without gravity, two people moving at a velocity with respect to each other assign different energies to different particles. The thing that gravity and general relativity and horizons make possible is that that energy can actually go negative. But to a person in the rest frame of the particle, it will always have a positive energy.

3:07:33.3 SC: Graham Snowden says, "What do you think the American scientific community could or should be doing to push back against the relentless and increasingly effective attacks from the Trump regime on the place of science in the USA?"

3:07:50.8 SC: Well, I think that the Trump regime is in power right now, and they're not open-minded. They're not gonna listen to advice or anything like that. I think that the American scientific community should do what a good, politically engaged set of people should always be doing, which is reaching the people who can be reached. I think the American scientific community should make the case for the importance of science to the broader public, and it should do it in many, many different ways. It should do it through TV shows and podcasts and books and events and any possible way it can do it. We have to spread the message of science and reason and the Enlightenment and all of those things, and we never are gonna stop doing it. It's not gonna be something like, "Oh, we'll do it for a year and we'll win, and then we can go back to doing research." This is just part of life in a democratic society, that you have to make the case for what you think is important. And I don't think we've been especially good at that. We had a good situation for much of the 20th century where the importance of science was simply unmistakable. In some sense, now it's even more unmistakable. It's less mistakeable. But nevertheless, people manage to deny it because we've invented all of these mechanisms by which people can deny reality for an extended period of time. Eventually, the price for that is gonna have to be paid, and this is why you get people getting measles and explosive diarrhea and so forth. But in the meantime, a lot of people are just gonna keep denying science. So you have to let some people be like that and reach the people who are reachable. We don't do a good job. We don't value it as scientists. We look down on people who do it. We don't put systematic efforts into doing it, and we're beginning to pay the price for that. And it's not just scientists. This is something that lots of people don't want to do, don't want to spend time and effort making the case for the importance of what they do. I think that we have to do that.

3:09:59.6 SC: Stuart Haynes says, "When you decided to leave Caltech for Johns Hopkins, how did you negotiate the transition of your grad students? Was that difficult?"

3:10:06.2 SC: I don't have much of an answer to this because none of my grad students transitioned. I knew that I was gonna be leaving Caltech before I did, and also we had a pandemic, remember? Like, I left Caltech in 2022, so you can do the math. So I had already stopped taking grad students, and my last Caltech grad student, Ashmeet Singh, graduated during the pandemic in, I think, 2021, maybe 2022, I forget. And then I started picking up new grad students when I got to Johns Hopkins. So I didn't actually have to go through that myself, nor did anybody else. It is a thing, just so anyone out there who might be interested. When professors move from one place to another and they have grad students working with them, it is often the case that they negotiate with wherever they're gonna go, "Can I bring my grad students with me?" And then there's a separate negotiation: "Can I just literally, physically bring them, but they're still students at the old institution, or can they join the new institution?" And all the different possibilities work out in different ways. It just depends on all the details. So I personally didn't have to go through that.

3:11:13.4 SC: Jonathan Cart says, "I recently saw Ed Copeland doing a video about reheating after inflation. He talked about how the inflaton field would decay via a couple mechanisms, perturbative and resonant, and that would generate the particles we see today. So my brain has accepted that heavy particles decay into lighter ones because they're in a lower energy state. But I have no intuitive understanding for why the inflaton field would decay. Is it a similar lower energy state mechanism that I need to wrap my head around?"

3:11:42.8 SC: Well, remember, what's a particle, right? What is a particle? A particle is a vibration in a field. That's how quantum field theory works. So the truth is that when inflation ends and the inflaton reaches the bottom of its potential, it rocks back and forth in the potential. It's sitting near a minimum and it's rocking back and forth, gradually decaying in amplitude for a couple of reasons. But you can interpret that field configuration as a set of particles. Since there's very little variation in the field from place to place, there's very little spatial gradients, the momentum of the particles is close to zero. They're particles sitting at rest, okay. And there's a lot of them. It's actually, technically, it's a Bose-Einstein condensate of particles. And if you want to think about the creation of ordinary matter and radiation, quarks and electrons and photons and so forth during reheating, you can, if you want to, just think of those individual inflaton particles as being heavy particles that decay into lighter ones. That's it. Now, there are subtleties that that picture doesn't capture because the field picture is more fundamental than the particle picture. So the details that Ed was talking about with perturbative and resonant mechanisms and so forth, these are all details that are more clearly seen mathematically from the field image rather than the particle image. But if you just want to have an intuitive understanding of how do the quarks and electrons get made, you can think of them as arising as decay products from the inflaton particles themselves, which are excitations of the inflaton field.

3:13:24.4 SC: Michael Honey says, "There have been several mathematical conjectures disproven by LLMs in recent weeks. Are there similar conjectures in theoretical physics that might be amenable to LLM investigation?"

3:13:35.8 SC: I could have grouped this, I guess, with the earlier question, but roughly speaking, no. I say that tentatively because the state of the art is improving very rapidly and maybe things are gonna change. But the nice thing about math is it is very clean and crisp to know when you have a disproof of a conjecture or even a proof that you could plug into a formal proof-checking thing like Lean, like we were talking about before. In science and physics, you don't have that, right? If I say, "Oh, look, I have a new theory for what the dark matter is," prove it, right? The LLM could prove a mathematical conjecture about that theory of dark matter, but it can't prove that your model is the dark matter. That requires experiments, okay? And LLMs are not very good at that. So the kinds of conjectures we have in theoretical physics are not ones that are conjectures about the properties of mathematical axioms or postulates. The word conjecture means something a little bit different in math and physics. In math, it means this statement may or may not be a provable theorem based on these axioms. In physics, it means this statement may or may not be a correct description of nature. So it's a very different task that is much harder to use an LLM to investigate.

3:14:55.5 SC: Chris says, "The other day you mentioned your enjoyment of playing silly games on your iPad for a mental break. Over the years, what have been your favorite games?"

3:15:03.7 SC: I think my favorite game over the years has been the Kingdom Rush series that I have on my iPad. It's a what is called a tower defense game. So the enemies come through and you build towers to prevent them from getting to your kingdom or whatever. It is a very silly game. What I like about it is the game is good, and there's a series of them. So I can play one and then whenever I want to play again, I can play a different one, and it's not just playing the same thing over and over again. My goal for playing silly games on my iPad is not to do too much intellectual effort. My intellectual effort goes somewhere else. So a lot of games require an investment, right? Like you need to take notes and really understand the world at a deep level and get into it and get immersed. And I love that idea, it's just not for me. I'm putting my intellectual effort elsewhere, and when I turn to my iPad to play a game, I want a break. I want something I can sit down, do for 10 minutes, complete a task, move on to something else. I'll also play traditional games: backgammon, poker, solitaire, whatever. Those are also things that I will play. I've tried various other kinds of genres, but it's a very fine line. You don't want to be too silly and stupid, you don't want to be too effortful and intellectual, and that is very hard for me to judge from the descriptions of the games on the App Store or whatever. So I have a difficult time figuring out whether I would actually like a game or not, and that makes me just stick to my old favorites.

3:16:38.3 SC: Anonymous says, "Are there thought experiments or intuition pumps neutral to the interpretation of quantum mechanics that can allow us to understand entanglement, or why multiple particle quantum wave functions are not separable by particle?"

3:16:52.7 SC: I mean, short answer is I don't know. I don't know, it's a matter of what will satisfy you, right? I don't think that there is once and for all a single answer to that. My attitude is that you should really take quantum mechanics seriously for what quantum mechanics says. And what I mean by that is, instead of starting with the classical world and trying to describe it in quantum mechanical terms. This is why people get in trouble with saying we can't know the position and velocity of a particle at the same time because of the Heisenberg uncertainty principle. Instead of what they should say is there is no such thing as the position or the velocity of a particle in general. There are certain specific quantum states that have definite answers to the question, "What would I observe the position or velocity to be?" and they're not the same states. That's the uncertainty principle. But if you think of what is real as the quantum state itself rather than your observational outcomes, and you understand that the quantum state of several particles is a single quantum state, not separate quantum states for separate things, that's not surprising at all. There's no difficulty in that. The x, y, and z components of the location of a particle are not separate things, they're part of one thing. Then the fact that different possible measurement outcomes are correlated because of entanglement, in my mind, is just like, "Okay, sure, of course they are." That's not hard to see. I think all of the difficulty in understanding entanglement comes from thinking of the measurement outcomes as fundamental rather than the quantum state as fundamental. And then if you start with measurement outcomes and work backward to describing them using a quantum state, then entanglement seems really, really weird and mysterious. If you just start with the quantum state, then I think everything becomes much easier to understand.

3:19:02.5 SC: Peter 42 says, "Clocks run slower the stronger the gravitational field they find themselves in. When we place a clock at the center of the Earth, the planet pulls on it equally from all sides, so there's no net gravitational field. Does a clock at the center of the Earth therefore run faster than one on the surface, or does it run slower because it has the mass pulling from all sides that slows it down?"

3:19:22.8 SC: So, clocks do not run slower the stronger the gravitational field they find themselves in. Clocks, unless they're broken, run at one second per second. This is the kind of thing that I say over and over again, and people kind of roll their eyes and like, "Why are you bothering? You know what we mean." But questions like this are exactly why it is better to say the true thing than the false thing. The true thing is that clocks run at one second per second no matter where they are. The true thing is that you can only compare the total amount of time that has been measured by two clocks when those two clocks start at the same position and time in spacetime and go on journeys and end at the same position and location and time in spacetime, start at the same event and end at the same event. The true thing to say is that if one clock starts outside a gravitational field, outside a black hole or whatever, and just stays there, not moving, and another clock starts at the same place and goes to visit the gravitational field and hangs out there and then comes back, the one that hung out for longer in the stronger gravitational field will read less time than the one that just stayed home. The way to remember it is staying home makes you always measure the more time. It's the flip side of, the shortest distance between two points is a straight line. The longest time between two events is just staying home and not doing anything. Okay? So it doesn't have any... There's no meaning to the question, "How do we compare a clock at the center of the Earth to a clock outside?" You have to start them at the same point and then travel. And you'll notice that if one stays outside and the other travels inside to the Earth, it has moved through a spatially dependent gravitational field. And then you can very easily ask the question, "What is the total amount of time elapsed?" And the amount of time elapsed from the clock that visited the center of the Earth will be less than the one for the clock that stayed outside. Perfect example of where saying it correctly helps you understand what to predict.

3:21:35.8 SC: Copernic says, "As a thought experiment, let's assume that infinity is purely a mathematical construct and that the real world is finite in every way. What interesting or significant things might that change, if any, with how you look at cosmology or theoretical physics?"

3:21:50.6 SC: I think this is a thing that people should be thinking about. I don't think that people have thought about it in any very careful way. There was a workshop organized by former Mindscape guest Justin Clarke-Doane a year or two ago, I'm forgetting now, but it was written up in Quanta Magazine that you can check out. It was by a bunch of people who are interested in ultrafinitism or finitism, I don't know how to even say it. The idea that mathematics should be based on just a finite number of things, not an infinite number of things. And you can do it, right? You can invent a new system of axioms that make a finitist arithmetic work. And then the question is, is that relevant to the real world? And on the physics side of things, very few people take this seriously because quantum mechanics does not say the world is made of little pixels or anything like that. It just absolutely doesn't say that at all. It says the world is made of quantum states. And how many quantum states are there? An infinite number. Even for a single qubit, there are an infinite number of quantum states. It might be a two-dimensional Hilbert space, but it's a vector space. There are an infinite number of vectors in that two-dimensional Hilbert space. So I was invited to the workshop because I wrote this paper a couple years ago on, "Could you discretize quantum mechanics?" and some other people had tried to do that before, but they didn't do it in a very good way, I thought. And indeed, this is kind of an interesting thing. I wrote the paper saying, "Look, you can, if you do things in exactly, precisely this way, come up with a truly discrete and finite version of quantum mechanics." But it requires a lot of what would look to us like fine-tuning. And the real problem is it has a Boltzmann brain problem because there's a finite-dimensional Hilbert space and time would go on forever, or at least it would be cyclic in some way.

3:23:55.9 SC: And the more recent paper that I wrote with Saakshi Dulani and Nadiia Diachenko pointed out that there's a loophole there. And in fact, you can avoid the Boltzmann brain problem in precisely that situation. So now I think it is possible. And when I say possible, I mean both mathematically/logically possible and compatible with all the data we have of the universe as we know it, that you only need a finite set of states to do physics in a reasonable way. So I think that's very interesting. Again, I don't think that there's any evidence that it's true. I think there's certain mathematical difficulties in the whole thing, but it's allowed. And I think that there's obvious benefits to it. Infinity can be hard to deal with. It's hard to make predictions when there's an infinite number of possible things that can happen. Having everything be finite would make many, many things much simpler. So if we did have any evidence it was true, it would be really nice, and we'll have to see how that works out.

3:24:56.8 SC: Kevin's Disobedience says, "Did Dirac accidentally invent string theory only to abandon it? Reading his biography, I was surprised to learn that he once conceived of electromagnetism as quantized lines of force, the opposite charges manifesting on each end. I realize it isn't open and closed vibrating strings that are giving rise to fundamental particles, and that it doesn't require extra dimensions, negative cosmological constant, supersymmetry, anything like that. But am I wrong to think that this sort of approach to quantizing a field theory is akin to string theory? Or is this octagonal", I think you mean orthogonal, or maybe it's a joke, "to the whole project?"

3:25:32.6 SC: Yes and no. So you can think of what Dirac did as a deep, deep precursor to string theory. He didn't invent string theory. And in particular, if you want to think of a positively charged particle and a negatively charged particle as being connected by a quantized line of force, the thing is, that line of force has no energy density whatsoever. It's completely a mathematical artifact. So that's very different than the fundamental starting point of string theory. String theory thinks that the physical, real stuff are the strings. Dirac never really thought that. He thought it was a way to quantize gauge theories. In fact, he was most interested in, I'm not even familiar with him ever talking about electric charges that way. He talked about magnetic charges that way. And that's a very famous construction called, unsurprisingly, the Dirac string.

3:26:24.7 SC: Those of you who know a little bit about Maxwell's equations, Maxwell's equations for electromagnetism might remember that there's an asymmetry there because there's a term that represents electric charges, but no term that represents magnetic charges. There are no magnetic monopoles in ordinary Maxwellian electromagnetism. So of course, again, every physicist is gonna think, "Well, why not? Maybe we could make it work." And Dirac is smarter than the average physicist, so he did make it work. And one of the big obstacles to making it work was that we have this idea that the electric field and the magnetic field are not the fundamental variables, right? There's something called the vector potential field from which you can derive both the electric field and the magnetic field. But if you believe that, what you find is that it's an immediate mathematical identity that there cannot be magnetic monopoles, that just the math doesn't allow you to do it. So Dirac figured out how you could do it by having singularities that you dealt with in a very delicate way, et cetera, et cetera. Later... So we have the Dirac monopole, Dirac magnetic monopole connected to the Dirac string, which is sort of a mathematical artifact. Later, the whole thing was made much more respectable in spontaneously broken gauge theories, where you can have the symmetry that was spontaneously broken be restored at the center of a magnetic monopole, and you don't need strings or singularities or anything like that. So Dirac did important work which did eventually feed into string theory, but he was not suggesting that the fundamental stuff of nature is made of strings in any way. So he should not get credit for inventing string theory.

3:28:08.9 SC: Okay, the last question comes from Matt, which is, "What are your thoughts on the 76ers after the Jaylen Brown trade and the new signing of LeBron?"

3:28:18.0 SC: Now, I know that I said in the introduction I wasn't gonna talk about that, but this is the last question. We're gonna let ourselves have some fun. Now I don't feel constrained. What I wanted to talk about... It was an obvious thing to do to talk about the Sixers in the intro to the podcast, but I know that it's not everyone's thing. So, and I could easily talk about it for a very long time, so I thought I would put it at the end. And I'm giving it to Matt's question, but many other people asked related questions to this. You should all consider yourself getting credit for this question. So where are we? To lay the groundwork here, the Sixers have been... There's a long history here. I could literally write a book. Books have been written, in fact, so I don't need to write the book. But there was this famous experiment in the NBA where the ownership of the 76ers fired their general manager and hired Sam Hinkie, who was an assistant general manager with the Houston Rockets. And he famously, he was very driven by analytics. That is to say, we should really look at the data, really learn how to figure out what is working, what is not working, rather than just going by our gut feel and the eye test. It is the basketball version of Moneyball, which is associated with Billy Beane and the Oakland A's. In fact, Sam Hinkie's boss with the Houston Rockets was Daryl Morey. And sometimes the basketball version of Moneyball is called Moreyball because Daryl Morey pioneered it. And Daryl Morey, as you know, is a Mindscape guest a little while ago. He later became Sam Hinkie's not quite successor, but successor of successor as leader of the Philadelphia 76ers. And a few months ago, he was fired. Sorry to hear that.

3:30:04.6 SC: But the reason I'm bringing this up is because it has been quite a journey. So Hinkie looked at the at-that-time roster of the 76ers and realized there was just no way with the assets they had that they could build a championship contender. And in his mind, the job was to win championships. Not to be pretty decent year after year, but to actually win the NBA title. And they were good enough, but not good enough... They were good enough to win some games, not good enough to win the championship. What are you gonna do? Well, in order to win the championship, you need superstar players. How do you get superstar players? You can try to trade for them and whatever, but if you don't have any superstars, how can you trade for them? You can try to attract them as free agents, but unless you're a good threat to win the championship, that's hard to do. So the other way to do it is to draft, get them through the NBA draft, which means you need a high draft pick. And in the NBA, high draft picks go to the teams that do badly. So the way to success is to get rid of all your good players, be bad for a couple of years, and then get a lot of good draft picks, hit on the lottery and get some superstars, and then win some championships.

3:31:15.8 SC: Now, this generated a huge amount of controversy. Hinkie was basically forced out... There's just so much detail here. Hinkie was basically forced out by the NBA leadership because they thought that he was losing too many games and making the Sixers look bad. His successor, Bryan Colangelo, was a nepo baby hire. He was the son of Jerry Colangelo, who was a big name in NBA circles. And he had to resign, Bryan Colangelo, he had to resign in a scandal where he was using burner accounts on Twitter to criticize his own players, and his wife was also doing that. And the whole thing was just horrible and messy. And meanwhile, the Sixers did become good because they were able to draft some good players. They also had some huge mistakes in the draft. That's okay. Hinkie himself never thought that once you get a good draft pick, you're set. You want to give yourself as many bites at the apple as you can.

3:32:18.2 SC: And the one big success they had was Joel Embiid, who has been an amazing player, won the MVP of the league, won multiple scoring titles, but has also had a lot of injury history. So he's about... I think he's gonna be either 32 or 33 years old in this upcoming season. So that's the beginning of your decline, typically, as an NBA player, but still in your prime, but beginning the decline downward. So he's been around for a while. But it has been a hilarious series of misadventures with the Sixers. They had a draft choice which was a local hero, Mikal Bridges, who played at Villanova right outside Philadelphia, whose mom worked for the 76ers. Everyone was so happy about this. Mikal now plays for the New York Knicks, just won the championship. And they immediately traded him away for another player who then had a peanut allergy, and he ate something that had peanuts in it unbeknownst to him and became debilitated by this and never played NBA basketball again. Like, how do you predict these kinds of things?

3:33:28.2 SC: They picked Markelle Fultz as a first round, as a very first choice, who was clearly the best basketball player to be picked that year. They put a lot of effort into picking him, and as soon as he arrived in the NBA, he could no longer shoot the basketball. And it's sort of an open secret that he clearly was injured. He got into some automobile accident and injured his shoulder. It's clearly a physical ailment that prevents him from shooting anymore. And what can you do? Like, how can you predict things like that? Anyway, here we are in the current era, and last year was Daryl Morey's last year, and he had made a big gamble. He had gone through a whole bunch of drama with Ben Simmons and James Harden or whatever, and he had finally convinced Paul George, who's a very, very good basketball player and was the leading free agent on the market in a certain year where they had money, to come join the 76ers. It was a great coup. Everyone was very happy about it. But the thing is that Paul George and Joel Embiid combined had a lot of money. They're senior basketball players. George is like 36 now, and Embiid is now, was 32 last year, 33 this year, I think. Don't quote me on those numbers. But they both were injured all the time for different reasons, and the team never quite got off the ground.

3:34:50.6 SC: Now, last year, the other two important players they had were Tyrese Maxey, who they completely lucked into. The thing about luck is sometimes it goes your way, sometimes it doesn't. So completely true, Daryl Morey was able to draft Tyrese Maxey relatively late in the draft, and other teams didn't pick him, in part because that was in 2020, the COVID year, and the NCAA March Madness basketball tournament was canceled. So a lot of teams didn't do their good scouting and missed on Tyrese Maxey, and the Sixers got him, and he's blossomed into an All-NBA superstar, which is great. And then the previous year, they had so many injuries with Embiid and Paul George and everybody else that they really just bottomed out again and got lucky in the lottery and got a high draft pick. They were able to pick third, and they picked V.J. Edgecombe, who has turned out to be an amazingly good player. Someone who easily could have been the very first pick in a different year, but there were two even more promising players chosen ahead of him. Perfectly fair, they're also superstars. But basically, so the Sixers had four really, really good players, two of whom had trouble staying on the court. And then just because everything weird happens to the Sixers, Paul George was suspended for 25 games for violating the league's drug policy. [chuckle]

3:36:18.7 SC: We don't know what drug it was. We're not sure exactly what that was all about, but it just throws another spanner into the works, as it were. So they entered the playoffs, and here's the thing, and I'm gonna talk a little bit more about this, but you have to address the elephant in the room, which is Joel Embiid is injured a lot. When he is playing at his peak, he is honestly as good a basketball player as I have ever seen. He won the MVP, I think, three years ago, and the year after he won the MVP, he was playing even better. It was just miraculous thing to see. He had a 70-point game. He would often score 50 points before the fourth quarter and they would sit him down because they were winning by too much. And then he suffered a horrific injury where another player from another team landed on his knee and twisted it in a terrible way. And since then he's never been quite the same.

3:37:12.6 SC: But there's a reputation that Embiid has acquired of just always being injured, never being available. And you can see if you don't follow the Sixers closely and instead just look from outside at the numbers, it's absolutely true that Joel Embiid has missed a tremendous amount of time by being injured. But there's being injured and being injured. There's sort of a kind of being injured where you're injury-prone in the sense that there's a part of your body that is just liable to give out. And there's also the possibility that you're just kind of unlucky. And I will defend the claim that Joel Embiid has just been unlucky. He missed the whole first two years of his NBA career because he hurt his ankle. He broke his ankle and the original fix was screwed up. It was bad surgery. So he ended up missing not just one year, but two years. But his ankles since then have been fine. The most recent thing that has really slowed him down was the knee, which is a whole different part of your body. And it was clearly someone fell on it. We all know why it happened. And that first year after he hurt his knee again, it wasn't well fixed, but he got a second surgery that really did fix it. And his own claim and the team's claim and the evidence on the court is that his knee is fine now.

3:38:38.7 SC: But meanwhile, in other years, it's just been a comedy of errors. One year he ran into... His face ran into the shoulder of one of his own teammates and broke his face. Like, literally broke the bone in his face and he had to wear a mask and it was just not the same. Another year he had Bell's palsy, which meant that the left side of his face was drooping and he could only see out of one eye. And these always happen just before the playoffs start. Okay? This year he had appendicitis just before the playoffs started. He had an appendectomy. He had surgery two weeks before the playoffs start. He was expected to miss all the playoffs. He heroically came back and he led the Sixers to a thrilling come-from-behind victory in the first round of the playoffs, where he was clearly the best player. So, on the one hand, yes, Joel Embiid has undoubtedly been injured an enormous amount. And if all that is what you knew, you would be smart to play the odds and say he'll be injured again. On the other hand, they're all weird injuries. There's something called a gambler's fallacy where you say, "Well, I've been having bad luck for a while, therefore I am due to have good luck." That is a fallacy. That is not true. But it's also a fallacy to say, "I've had bad luck again and again and again, therefore I will just keep having bad luck." Okay?

3:40:03.7 SC: I would argue that it would not surprise me at all for Joel Embiid to be perfectly healthy this year. In fact, at the beginning of last year, before they started playing the games, Embiid said very explicitly, "I'm gonna be back this year, but I won't be 100% healthy until next year." And he was right. And this is the year where he predicted he will be 100% healthy. So that's all preliminary to saying that last year didn't quite work. They got to the playoffs... Actually, they didn't get to the playoffs. They got to what is called the play-in, where four teams play for two slots. But they won their play-in game and they faced off against the hated Boston Celtics. And they went down 3-1. And Embiid sort of worked himself back into shape after the appendectomy and willed them to a thrilling 4-3 victory after being down 3-1. The only time in the history of the Boston Celtics franchise that they have lost a series after being up three games to one.

3:41:05.6 SC: But honestly, it was pretty clear that all of that heroism to win the series after being down 3-1 left them completely wiped out. And then they hit the New York Knicks in the second round and walked into a buzzsaw and got destroyed four games to nothing. And it didn't help that Nick Nurse, their coach, really played all of their players very, very heavy minutes. So that... Like, you literally saw, because I was watching these games, they were competitive in the first three quarters and then they would get destroyed in the fourth quarter because they were just too tired. So, okay, that's fine. So that's where we are at the end of this season. And both Paul George's contract and Joel Embiid's contracts are so big that even though they're All-Stars, they're first-ballot Hall of Famers once they retire, they're considered that no one would ever take these contracts on because they're not able to play enough games.

3:42:01.9 SC: So the Sixers fired Daryl Morey. I feel bad about that. But they hired this guy, Mike Gansey, who, gotta give credit, has been working miracles. The first miracle he worked was he traded Paul George to the Boston Celtics for Jaylen Brown. Jaylen Brown is another All-NBA level player. He's younger. I think he's gonna hit 30 this year. Again, I'm not so clear on the ages. He has his own issues, which we can talk about. But clearly at a basketball level by itself is a clear upgrade to Paul George. And they had to send in a couple of extra draft picks to do it. But everyone in basketball land thought that the Sixers got the better end of that deal. And because of that, there was one domino waiting to fall, which is that LeBron James, who is among the top basketball players ever to pick up a basketball, let's just say that, arguably the number one best basketball player of all time, reaching the end of his career. He's showing us superhuman levels of longevity. In addition to being an amazingly good basketball player in his prime, he's been able to keep up his super high-level play very, very long. He's gonna be 42 this upcoming year. There's never been an NBA All-Star level player at age 41 or 42. But he was absolutely playing, not at a top-five player in the NBA level, but at a top-20 player, no doubt, last year. But he was in Los Angeles and they had pulled off an amazing trade for a much younger superstar, Luka Dončić. And it was clear that LeBron was no longer in their plans. And so he basically, in some sense, he said, "You can't fire me, I quit." And LeBron said, "I'm leaving." And he also said... He's like the wealthiest NBA player ever. He has a lot of money because he's been doing well for a long time. He says, "I don't care how much money I make. I just want to go to a good situation." People thought that LeBron might want to go back to either Cleveland or Miami, where he had previously played, or maybe to Golden State to join his buddies Steph Curry and Draymond Green to have a last hurrah. But instead, to certain people's shock, he joined the Philadelphia 76ers.

3:44:22.1 SC: Now, there's a lot to say about this. A lot, a lot to say about this, but the short version is we are all over the moon here in Philadelphia 76ers fandom. And I'm certainly over the moon too, because not only is he LeBron James, one of the best people ever to play basketball, but his skill set is exactly what the Sixers needed. And it's very funny for me to see people who are professional basketball commentators who clearly haven't watched a lot of Sixers games because they're saying a lot of stupid stuff about how LeBron is gonna be yet another superstar who needs the ball all the time, and that's gonna hurt the development of other players on the Sixers and whatever. It's just completely backwards. At this stage of LeBron James's career, the thing he most brings to the table is his super high basketball IQ and his ability to both literally pass the ball, but also just more generally orchestrate the offense and the defense in real time, which is exactly what the Sixers were lacking. Tyrese Maxey, who is an amazing player and love him to death, also just one of the most likable players in the NBA, has been the Sixers point guard. He's been the one delivering the ball, the passes, as well as scoring himself. And so people have this idea that he needs the ball to be successful. But any Sixers fan will tell you that when he is at his best is when someone else is the point guard and he gets to catch the ball and shoot and make what we call attacking closeouts, where you catch the ball, someone runs at you, and Tyrese Maxey is literally in the conversation for the fastest player in the NBA. So if he does get the ball to shoot and someone tries to close out on him very quickly, he can move past them and get a layup like nobody's business.

3:46:15.7 SC: LeBron James is exactly the person you want to have playing that role. And his best position these days, LeBron's, is power forward, which is exactly the one position that the Sixers did not have a great player at. They have very good players, but they're backup level players. And so LeBron fits into this scheme perfectly. The other guy, V.J. Edgecombe, who was a rookie last year, was a super talented rookie and he's gonna get better this year. So basically you have four players in the starting five who are established All-Stars. At least two of them, Embiid and LeBron, are easy first-ballot Hall of Famers. And the other two, Jaylen Brown and Tyrese Maxey, are likely Hall of Famers. And then V.J. Edgecombe is gonna be in his second year of a promising young career. It is arguably the best starting five ever to take the court in the NBA. Now, that means nothing. That means nothing as far as results are concerned. I think that there's all sorts of reasons to think it will work. And I... Like I said in the intro, I choose to be an optimist and a fan and think that everything is gonna go great, at least until it doesn't. Now, I'm also not completely irrational and I can be a realist when I want to. So there are some things that could go wrong. I think that there are three things that could go wrong that are reasonable things to worry about, and I will list them in what I consider to be increasing order of concern. One is Embiid's health. I think Embiid is healthy and I think that he's gonna surprise a lot of people. But the numbers don't lie. He has managed to find weird, bizarre ways to get injured many times in the past. So it could happen again. That's absolutely something to worry about, and especially because we don't have great backup centers. That's the one position when it comes to the bench where the Sixers are not very well stocked right now. So it's a... But what can you do about that? I think that that's just something we'll have to roll with and see how it is. The second.

3:48:24.4 SC: The second thing is LeBron James is gonna be 42 years old. It's unheard of, literally unheard of for someone to play high-level basketball in the NBA at that age. People have played, a couple, a handful of people, but they were clearly benchwarmers and whatever. We're relying a lot on a 42-year-old guy to keep it together and play at a high level, he's been declining, but the decline has been so gradual, it's been weird. It's absolutely possible that this year will be different and he'll have more serious decline. I think that there's reasons to believe he'll be energized by the new situation, by the new challenge, by the new prospect of winning an NBA championship with a fourth team. He's won championships in Miami and Cleveland and LA already. So I'm optimistic that LeBron will step up to the plate here. But realistically, being 42 years old, things can happen. You don't wake up quite as bouncy as you used to. I can vouch for that.

3:49:26.9 SC: The big question for me is Jaylen Brown. So Jaylen Brown is a hugely talented basketball player, but he's a very kind of specific basketball player and he's not actually an obvious perfect fit with the rest of the operation. His talents are... Like last year, his running mate, as it were, his fellow superstar on the Celtics, Jayson Tatum, was injured most of the year and Jaylen got to sort of be the alpha dog. And he was very good at that. The Celtics won far more games than anyone predicted, led by Jaylen Brown. And part of the reason why he was willing to be let go by the Celtics is because he wanted to be the leader of a team all by himself rather than playing second fiddle. Arguably now he's playing fourth fiddle. So that's not the direction in which he wanted to move. And the specifics of his game are very much about him having the ball and being the alpha dog and having an isolation called for him and working his magic. And so there's good parts and bad parts to that.

3:50:29.0 SC: The good part is sometimes you just need someone to step up and get a bucket all by themselves with individual heroism, and he's the perfect person to do that. The downside is often when all the other players on your team are world-class good, you need to be a little bit more part of a system, not just an individual hero. He's not especially great at shooting from long distance, three-pointers. And in a very real way, a younger Paul George is a better fit to this team than Jaylen Brown is right now. But two things about that. One is that Paul George is not young anymore. And number two, I could see by watching last year, I could see why Paul George rubs people the wrong way a little bit. I mean, he's incredibly talented. He has the ability to step up and just control a game for brief stretches, but he doesn't. He kind of blends in. And when he shoots and scores in sort of low-impact moments more often than in high-impact moments. Whereas Jaylen Brown is both younger and bouncier and is pretty good at rising to the occasion. But what I can see the possible problem is Jaylen not wanting to fit in. If Jaylen Brown decides that he wants to do what it takes to win an NBA championship and really dedicate himself to playing defense and to sharing the ball and playing smart, high-IQ basketball with these other super-talented teammates, the Sixers will be unbeatable next year. I mean, if he chooses to do that, there's just no one that can stand in their way. But he might not. And I don't know. I truly don't know. That's a completely valid concern in my mind. To be fair, we have three players now who are a little bit of drama queens between Joel Embiid, LeBron James, and Jaylen Brown, none of these people take it quietly when things are not going their way. And I think they're all good people. I think their hearts are in the right place, but they're not like the Tyrese Maxeys of the world. So, Tyrese can just play with anyone and be a success, and everyone loves him. And these three are more prickly pears a little bit.

3:52:55.3 SC: I think that winning solves a lot of problems, and if they find themselves winning a huge number of games, everyone will be happy. But that is something that I would at least worry a little bit about. Otherwise, I'm not worried about a lot of what people are worried about. They're worried like, "Oh, there's only one ball." That's fine. The ball will be in the hand of LeBron James, one of the best players to ever play the game. And think of it this way: if you're the other team, when you're playing five-on-five basketball, the first thing you do as the coach of a team on defense is say, "Okay, some of our players are better at defense than other players. Every team has some people who are not good defenders. We're gonna put our worst defender on the other team's worst scorer." Right? That's a very natural thing to do.

3:53:42.2 SC: The new Philadelphia 76ers have no bad offensive players. There's nowhere to hide a weak defensive player. The obvious choice is on V.J. Edgecombe, the guy who was the rookie last year, who in his very first NBA game scored 34 points. He scored far fewer than that on average because he specifically deferred to his older, more established teammates, not because he couldn't. He's an amazing scorer from all the different parts of the basketball court. So good luck with that. There's other people... So basically, LeBron James is gonna start the offense. He's gonna push it around and choose your poison. You can have Joel Embiid in the post or at the nail, and you can have Jaylen Brown and V.J. Edgecombe and Tyrese Maxey zipping around the perimeter, making screens for each other. You can't double-team anybody because you can't leave anybody open on this team. Good luck stopping them from scoring. You might worry about the defensive side, but again, I think this is something where people don't watch the Sixers, and that's why they have these opinions. I think that Joel Embiid and LeBron James both have the property that they can be super good defenders when they want to be. Now, admittedly, these days, they don't usually put their energy into that. They put their energy elsewhere. But guess what? On this team, they can put their energy into that because there's so many players that can score the basketball. So if they really... They might coast a little bit through the regular season, but when the playoffs come out and they really need to shut people down, both LeBron and Joel can absolutely do that.

3:55:28.6 SC: Tyrese Maxey is a little tiny guy. He's like my height, okay? And that is a disadvantage. But he's super dedicated to the defensive end, and if you look at the numbers, he's an absolutely average NBA defender. Jaylen Brown, like I said, has all the tools to be a well-above-average NBA defender. Sometimes he does that, sometimes he doesn't. V.J. Edgecombe is an elite NBA defender already as a rookie, if you look at the numbers. He's been tasked with guarding the other team's best players and has done an incredibly good job. So I'm not even worried about the defensive end. I'm very excited about the upcoming season. I think that the Sixers and the city of Philadelphia and their fans deserve something to go right for them after all the trauma of the last decade or so. No way of knowing whether it's gonna happen or not. If we knew, it would be less exciting. It's living through the experience that is worth it. I can tell you that I'm gonna be going to some Sixers games this year if I can afford them. Who knows? They might be out of my price range these days.

3:56:37.0 SC: But I'm very excited to see what happens. I love the team, both as it's constructed and the individual members of it. And I'm gonna try my best not to devote too much AMA time in the upcoming year to talking about the Philadelphia 76ers. But if they win the championship, I'm not gonna be held to that promise. So thank you all for listening to this month's AMA. Thank you all for supporting Mindscape. Thanks especially to the Patreon supporters for doing more than just moral support. Hope you enjoyed this one. I'll talk to you next time.

[music]