Welcome to the November 2026 Ask Me Anything episode of Mindscape! These monthly excursions are funded by Patreon supporters (who are also the ones asking the questions). We take questions asked by Patreons, whittle them down to a more manageable number -- based primarily on whether I have anything interesting to say about them, not whether the questions themselves are good -- and sometimes group them together if they are about a similar topic. Enjoy!
Listen or watch Rolex Planet Visionaries wherever you get your podcasts. #ad
Got to https://scribe.how/mindscape and mention Mindscape for you first month of Scribe Capture free on select plans. #ad
Get 30% off your first year of the Base44 Builder plan at https://Base44.com/mindscape. #ad
Use code MINDSCAPE at https://monarch.com/ to get your first year of Monarch Core half off at just $50. #ad
Look your best and support men's health at https://www.amazon.com/b?node=223476965011&ref_=MOV26_pod #ad
Support Mindscape on Patreon.
Click to Show AMA Questions Only
AMA questions November 2026
Peter
If you could have a super power, what would it be? Flight, super strength, super speed, teleportation, telekenesis (could be a fun party trick!) etc.
James McBeath
In the September 2026 AMA, you said:
The eternal Schwarzschild solution has all sorts of weird properties. It has not only a black hole, but it has a white hole in the past out of which things can come but nothing can go in. ... But the Einstein-Rosen bridge and the white hole are not thought to really exist in the world, but we study them because there's a simple mathematical setup to look at first.
In other podcasts you have said that scientists should have the courage to accept what the math is telling them. How do you decide when to accept the math and when to assume the predictions don't really exist?
Friedrich
As a European (German/Swedish) I regularly have to watch with horror how the American election systems create really poor outcomes for everyone involved. But from a physics of democracy perspective, aren't you stuck somehow? The politicians who could change anything are the ones who are heavily dependent on its existence. And psychologically, most voters can probably not fathom any alternative without having lived abroad for some time. Do you see a path that could work to move to enable better outcomes without destabilizing society on a catastrophic level?
lucas pollard
Have you ever significantly changed your mind about a scientific or philosophical question without encountering new evidence, simply because you came to understand the existing arguments differently?
Roo Phillips
I loved the recent episode with Greg Caruso. I know you're not a fan of free will discussions, but keep them coming! I'm curious how you would compare and contrast your compatibilist views with Greg's free will skeptic views, specifically on what changes to the US legal system should theoretically be pursued. I'm having a hard time understanding the differences between compatibilists and skeptics like Greg, and I'm hoping that when applied to a specific case like how we might modify the justice system might make it a little more clear.
Jeffrey Segall
I was wondering whether you think it is just as well that there does not seem to be the possibility of faster than light travel or communication. As I think von Neumann pointed out, since our species has evolved rather late in the development of the universe, if faster than light travel were possible, it is likely that a species that developed earlier might have already spread throughout the galaxy, including earth, and blocked the evolution of humans. Similarly, with faster than light communication, a more advanced civilization might be able to hack our technology (along the lines of the three body problem novels) and doom humanity. Given those possible dangers, do you nevertheless hope that faster than light travel/communication may be possible or do you breathe a sigh of relief that it is not?
half-a-crackpot
My question for you is as an author. Let’s say, hypothetically speaking mind you, that I own copies of a few of your books. I get to see them sitting on my shelf in all their glory every time I walk into my home office. Then one day I happened to find PDFs of those same books that fell off the back of a truck. I tried my best to chase them down to return them, but alas, I didn’t even manage to get their plate. Would you personally be offended, or feel wronged, if I used those digital copies that were not paid for?
Perry Whan
I am having trouble thinking through the differences between an infinite universe (repeating?) and a continuous universe that just bubbles off from itself. Is there an intuitive way to think about these distinctions?
Kevin’s Disobedience
You don’t strike me as someone who is inspired by any one particular philosopher, but if you were going to teach a course on one person’s philosophy who would it be?
********************************************
Alex West
There seems to be split opinion on OpenAI withholding their Yang-Mills existence proof from public view. Thoughts?
Alexander Knochel
I would love to hear your opinion about the latest Yang-Mills claims by OpenAI. I'm personally very unhappy how the results are communicated, but am of course very interested in learning more about rigorous definitions of Yang Mills and the Standard Model.
p.s. how do you view the significance of the actual result, irresepective of how it was communicated
Tyler Whitmer
i have a sort of meta-question about the yang-mills openai announcement. i was struck by the different ways the various constituencies engaged by openai in the math and physics worlds asked for the discoveries to be handled by openai. do you have any thoughts on how you’d prefer these things to be dealt with in the future, especially if they directly affected an area where you’re a published expert?
**************************************
Caleb Yaryan
My son is a Jr. In highschool and has expressed interest in pursuing a career in physics, I have an engineering background and have practical advice for him if he wanted to pursue engineering but I am curious what advice you have for a highschool student that is considering physics?
Philip Obbard
Do you think there may be something in Nietzsche’s idea of eternal recurrence if the block universe is to be taken literally? After all, if the entire history of the universe is “out there”, there should be nothing to stop you experiencing your life anew after you die (?).
Joshua Hillerup
Would anthropic reasoning mean we should favor with Fermi paradox explanations one that maximize the number of intelligent species out there who are capable of wondering about the Fermi paradox?
KalessinTheFirst
If big wormholes a-la Star Trek were real, what would be the safest way to enter them? In media, wormholes act like objects on a 2d plane and spacetime bends inward perpendicular to that plane. The spaceship then enters from a direction perpendicular to the plane, right into the middle of the curvature. But in real life, they would be spherical and spaceships could not enter at an angle perpendicular to the plane. So could one enter from any direction? It's science fiction so maybe anything goes.
Tom B. Night
PRIORITY QUESTION: Apologies in advance because I know you don't like questions about free will. It seems like many people in this debate, for example your recent guest Gregg Caruso, try to eat their cake and have it too. They acknowledge someone's lack of free will when they make a bad decision, like committing a crime, then assume others have free will when deciding how to feel or respond. It seems the practical, logically consistent position is to treat everyone as if they have something resembling libertarian free will. What do you think?
David Commander
How does a Boltzmann brain continue to survive and if it can't why isn't ongoing experiences in the moments of now insufficient to prove I am not one? In other words, if I can have false past memories can I also have false ongoing experiences or the illusion thereof?
Chris Kaltwasser
I have an unpopular theory, and that is that the three-point shot should be removed from the NBA. It rewards a less exciting game. Who cares how good somebody is at shooting a long shot? I would rather see really cool plays that involve strategy, great passing, getting inside, and that kind of thing.
I have yet to do the research, but I don't think the three-point shot actually pays off as often as people think. I feel like players miss them more often than they make them, giving up possession without making a shot that might have taken more strategy but had better odds.
What are your thoughts on this topic?
Nikola ivanov
If cosmic birefringence is conclusively detected, how could we distinguish observationally between a fundamental violation of Lorentz invariance, as explored in your early work, and birefringence caused by an evolving field in a Lorentz-invariant universe?
Rich Hogg
As one with a strong naturalistic bent,I would tend to think someone else's supernatural experience was a hallucination or other misfiring of his brain. But I want to keep an open mind while not forgetting that it could happen to me. If I see a ghost tonight, how do I as a "maybe-crazy Bayesian" decide whether I should update my priors, or make an appointment with a neurologist?
Michael Bright
Since learning even just a little bit about Einstein's explanation of gravity I've always been bothered by the cartoon image of a ball on a trampoline. It makes it look like space has a "down". If you agree with my frustration, is there a more preferable cartoonish image or set of images in your mind to visualize what spacetime curvature actually is?
Zach McKinney
The prevailing cultural narrative around AI is one of inevitability and runaway proliferation. However, this chain of reasoning never seems to reckon with the planetary boundaries of finite material resources and energy.
On what time scale (if any) do you foresee the proliferation and diffusion of AI to become constrained by such limits?
Michael Aster
Since you've talked about how you like the game, how would you feel if Kingdom Rush reached out for a sponsorship on the podcast?
David Sotolongo
In the April 2023 AMA, you said: "I think we need to build a lot more housing. I think it is very, very obvious, that's the best way to make housing more affordable, build more of it."
I agree, and am curious how you think we should go about building more housing? Should the government step in to build, and maybe manage, such housing? And what, if anything, would you do about existing homeowners who oppose new housing because it would decrease the value of their property?
Allan Lubel
Your intelligence and generosity of spirit has really made an impact on me, asking if you feel you received these qualities from your parents or any other influences when growing up and if you could give a couple of examples.
Miron Mizrahi
In the last AMA, when you discussed neutrino oscillations, you said something that sounded a lot like an uncertainty principle: there is no state when the neutrino has both definitive mass AND definitive flavour. I get how this can work for a single particle or field. But how does it work for three? or is there one field for all neutrinos and the mass/falvor combinations are just different values of the field?
Dylan Samuel
I'm a teaching assistant for a mechanical engineering class and I'm pretty sure that a very large percentage of the class is using AI to do their homework, as their test scores tend to be vastly worse than their homework scores. Recently, a UC Berkeley professor made news saying that they no longer see the point in grading their students' homework. Apologies if you've answered something similar before, but what are your thoughts on this, and generally how are you adapting to teach with students knowing they'll likely use AI? It seems counterproductive to pay for college and then not learn anything!
Tucker Hiatt
Am I mistaken to interpret Newton’s Second Law to include the notion that forces CAUSE accelerations, and to want something similarly causal in a QFT worldview? Please don’t make me give up on causality!
Matt Patterson
I’m an avid Mindscape follower and was intrigued by your discussion with Jenann Ismael and her philosophical argument about why it’s OK to kill cows but not humans based on their Value. Value, in this case, being measured by one’s memory of one’s life, having plans for one’s future as well as these same items in others that one has influenced. For cows, this kind of Value is essentially zero and for humans it is much greater. It occurred to me that the Value of individual humans can vary over a wide range (and might even be negative - e.g., in the case of a serial killer). Given that thoughtful humans do decide to kill other humans it seems that considering the Value of those that would die might be one of the considerations in evaluating the moral rightness or wrongness of making such a decision. My question is: Do you think that the Value of potential “death victims” should be one of the considerations when assessing the morality of decisions to kill them? Or is there some moral, or other, objection to this idea?
Karoly Kantor
How likely do you think it is that the observable universe is such a minuscule part of the total universe that our observations are just local phenomena and say little of use about the whole?
Calvin Firth
In mad-dog Everettian QM, where there is a single wavefunction of the universe that evolves with time according to some Hamiltonian, what exactly is that Hamiltonian, or where does it come from? In what I have learned so far, as a beginning physics grad student, the Hamiltonian comes from some potential, which is usually classical-looking (ie Hydrogen atom) or at the very least is a function of space (ie inf. square well, SHO, etc.). However, in the mad-dog Everettian picture, I take it that space is emergent. So where does the Hamiltonian come from? Where does it "live"?
Felivel Camilo
Do you believe technologies such as LLMs or realistic image generation arrived at an inappropriate moment? Might humanity have benefited more from developing them only after achieving some form of enlightenment?
Redlynx
What is your opinion about AI successionism, the idea that if machines will become conscious it would be ethically right to let them replace humanity?
Edwin Miles
I have been wondering recently about ‘value’ and wanted to ask the following:
Is the value of something only and always contingent on there being someone or something to value it? Or is there some intrinsic ‘value’ to something separate to any human bias, preference and valuation etc.
scott
Do Feynman diagrams have ontic content or are they purely computational? I understand that all we measure are scattering amplitudes, not full histories of particle interactions that Feynman diagrams might visually imply. But it seems that among working physicists, these diagrams are essential vocabulary for discussing and visualizing processes. It makes me think there is more to them.
**************************************
David Khudaverdyan
could you please talk about pointer states? Which aspects of them are already explained and understood by the physics community, and which aspects still await an explanation? You once mentioned that finding the classical world is still an open question in the many-worlds interpretation.
Henry Jacobs
On the question of “why do we never observe a superposition”. Is it plausible that observing always involves large systems becoming entangled with the small quantum spin or whatever, and we don’t consider the system “observed” until it has evolved into one of the eigenstates. This seems circular but I don’t know if that’s a bad thing
***********************************
Daniel Zaccariello
Baseball has no clock; outs are its time. And its rulebook forces an exact conservation law: every batter ends the inning as an out, a run, or a man left on base. Physics derives its conservation laws from the form of its rules too, except it had to guess the rules. Is that the only difference?
Anonymous
If you were starting over today, what assumption in our current formulations of quantum theory would you be most inclined to question—not because it is demonstrably wrong, but because removing it might open a genuinely new direction?
Daniel Shulman
It seems to be accepted dogma that singularities exist inside black holes and some time just before the Big Bang. If we treat spacetime as having some torsion, instead of zero torsion, we get Einstein-Cartan gravity, which looks exactly like GR except at the small scales where singularities matter, and the repulsive forces from the torsion in Einstein-Cartan gravity eliminates singularities. Is it just a historical accident that Einstein didn’t know enough geometry to introduce torsion that we still believe in singularities, or is there a reason why physicists don’t give more credence to Einstein-Cartan gravity?
Mat Becker
Do you enjoy gardening or have any other outdoor hobbies?
Jeroen
Do you think there are interesting dynamics in quantum field theory that are under explored because they don't map well to particles?
MrG
It was recently reported by Sludge that former podcast guest Cornel West's 2024 presidential campaign was indirectly funded by Elon Musk through the "Interstate Priorities" election fund. While its not clear that West knew of this involvement, its difficult to reconcile the purportedly pro-democracy rhetoric of someone like West with the practical effects of his, and similar ostensibly pro-democracy political actors, actions working against American democracy.
How should liberals act towards those who claim to support democracy but actively work against it?
Rob Butler
If a potential solution to the fine tuning problem is a multiverse, and if they all expand like ours, what prevents the universes from expanding into each other? If they had different laws of physics would they interact, or rather pass by like ships in the night?
David Maxwell
Like you, I've always believed in letting ideas flow and trusting people to sort through the truth. Recent years have made me question that is media landscapes have enabled systems where large percentages of the population only see total fabrication and never see the true ideas. Have your views on censorship changed?
josh
You’ve mentioned playing poker a few times, so I’ve gotta ask…assuming it’s Texas Hold’em, what’s your favorite starting hand? Are you a big pocket pair kind of guy, or do you get more excited seeing suited connectors?
Alen
In the previous AMA you characterized disagreements about free will as largely semantic, despite arguing just minutes earlier in the same episode that emergent phenomena are genuinely real. You also seemed to endorse the view that free will emerges from agency in the recent Christian List episode. Can you reconcile the apparent contradiction or clarify whether I’m missing something here?
Paul Torek
How important is it for Everettians to *derive* the Born rule? Obviously, it would be *nice* to reduce your theory's number of assumptions by 1. But don't most other "interpretations" of QM help themselves to an empirically-validated Born rule, rather than deriving it?
Gustavo Chaves
In the March/2026 AMA, while answering a question about your take on the notion of "computational functionalism", you said that you "think that the underlying processes that lead you from the input to the output are crucially important in whether or not those processes involve something that we can call consciousness."
In his recent essay, The Mythology Of Conscious AI (https://www.noemamag.com/the-mythology-of-conscious-ai/), Anil Seth challenged the notion of "computational functionalism" arguing, among other things, that there are functions that Turing Machines cannot compute perfectly, such as continuous or stochastic functions. As I understand it, he argues that perhaps consciousness depends on processes that a Turing Machine cannot compute, but only simulates in a way that isn't precise enough to allow for consciousness to emerge.
Is this more or less what you're thinking about when you mentioned the importance of the "underlying processes" happening in our brains or was it something different?
******************************************
Krisztian Dobo
Do you think AI will be able to produce new theories in physics just by the sheer fact that it has access to a vast amount of papers, findings and theories and is able to hold them in its head and freely recombine them? I understand your point about physics, that given its experimental nature (with regards to proving things, anyway) it's less susceptible to AI breakthroughs, but i have a feeling there might be low hanging (yet important) fruits hanging around that are not discovered because of the immense specialization and amount of knowledge in science.
James Alan
What do you predict would happen if you constructed a training data set for an AI that only contained data from quantum mechanics experiments. Could you build up an agent who’s intuition about the world were quantum mechanical instead of classical? do you predict being able to extract novel insights from such an agent?
*********************************************
Jonathan Kleiman
PRIORITY QUESTION:
Sean, how do you distinguish genuine academic consensus from people being afraid to disagree? Has that fear contributed to public distrust of universities, and what should they do differently?
P Walder
How do you think AI will inform Theoretical Physics in the short term - 3-12 months?
C Jurlando
In your April AMA, you discussed Habermas’s concern that economic and administrative systems can colonize the lifeworld, displacing communication and shared meaning. A successor in that critical tradition, the sociologist Hartmut Rosa, has developed a related critique through his concept of “resonance”: relationships with people, work, nature, or art in which we feel addressed, can respond, and are changed by the encounter. In such moments, the world seems to come alive and "crackle," drawing us into encounters we cannot fully control. Rosa proposes resonance as a criterion for a good life beyond wealth or the mere satisfaction of preferences. I find this compelling: a society could become ever richer and more efficient while making resonant relationships increasingly difficult to sustain - and that is bad! Do you think resonance could provide an objective standard for judging a society’s success, or at least a standard that goes beyond subjective preferences? What would justify treating it as such?
Robert Ruxandrescu
Have you thought about adding some perks for the higher paying members of Mindscape? For example, you could hold a trimestrial or yearly two hours Zoom meeting with the higher tier paying members and have a live chat with them, like a live AMA where you could have short back and forth conversations. What do you think?
the_priority_question
How much of an impact do you think that social media and their algorithms promoting echo chamber bubbles has on the current political polarisation, and is this a bad thing?
Peter Huncar
I really enjoyed your episode with Gregg Caruso. It got me thinking about the “who” behind the choosing, so here’s a two-sided question:
Who is Sean, and when is Sean?
Narrator, brain, or body, and over what timescale?
Over seconds you look autonomous, but over decades your zip code does a lot of the choosing.
Where do you draw the boundary of the self that chooses? Is that a fact or a modeling choice, and at what point does “Sean did it” become just “the universe evolved”?
Sid Huff
There has been considerable discussion in the news media recently about the concept of designing a “kill switch” that would shut down an AI system if it were determined to have “gone rogue.” From your understanding of current and future AI, would a kill switch be an effective solution to potential AI malfunction?
Michael Helgerson
With the NBA season in full swing and your work on the Physics of Democracy underway, what would your outline for The Physics of Basketball book look like?
Owe
I was a little surprised in your reflection after talking with Chandra Sripada at how strongly you seemed to take his perspective. One thing I found mostly missing in the conversation was what's built in to language. My question is: do you think that Sripada was correct in ascribing so much intent to the models (for example, he mentions "the model is doing something like implicit Bayesian inference", or instead is something like Bayesian inference being built in to the weights between words because humans use Bayes in constructing the sentences that LLMs are trained on?
Brandon Lewis
I'm still trying to understand how to think about particles in QFT. I get that the fields permeate all of space. The fields are constantly oscillating within their degrees of freedom, but also coupled to each other, so the oscillation can move between fields. When that happens, we observe particles decaying or fusing. We can only observe discrete measurement outcomes of a continuous wave function, the infinite other possibilities lost to us once our universe has made its choice. Are particles in QFT just when the wave function has sufficient energy or density at some location?
Joie
As often heard on Mindscape, the early universe is described as very low entropy that increases with time, which is also how/why we experience time linearly. As a total physics novice who only knows about it from this podcast and other lay sources, my question is, can entropy be zero at time zero? If so, what would that look like and mean?
Marie Rausku
With the upcoming Halloween and people talking about ghosts, haunting and paranormal stuff I have realised that so many people have easy to believe in ghosts, demonic possession or afterlife. But then the Many World Interpretation of quantum mechanics is called or counterintuitive or "bonkers" and "much harder to believe in" (than ghosts)- not just by people believe in ghosts but by other scientists too. Why is that, what do you think?
Rob Gebele
You consider Boltzmann brains a serious problem for cosmology because they may vastly outnumber ordinary observers. Yet in Everettian quantum mechanics, observers on branches with arbitrarily small amplitudes are also real, and you appeal to Born-rule weights rather than simply counting them. If Boltzmann brains themselves arise from rare quantum fluctuations, why can’t we resolve the BB problem by applying an appropriate quantum measure that suppresses their contribution? What makes the quantum measure principled but an analogous cosmological measure illegitimate?
Daniel Bagley
A recent article in Aeon states, “If panpsychism doesn’t resonate with you, it’s worth asking if there truly is a reason for everything – if the world is ultimately intelligible after all.” Do you think a belief in consciousness as an emergent thing invites the view that there is no reason for the universe to exist? I take it that “reason” is a function of consciousness, and panpsychists believe that the reason the universe exists is that consciousness exists, but I’m not sure if it’s correct to think that a person who doesn’t believe consciousness is fundamental would also believe there is no reason for a universe that produced conscious beings to exist.
Mindscape
Do you have a description or definition of how or what you define as consciousness? I'm enjoying reading the many ways consciousness is currently described by others in various fields. I'd enjoy hearing your thinking on the subject. Do you align with a particular definition?
Jonathan Good
Could 2 branches on the path from a low-entropy past contain different conscious experiences but still become 1 branch in the future if the differences were never recorded in a way that became entangled with current observers' degrees of freedom in a way that produces lasting decoherent differences?
Dan Butler
PRIORITY QUESTION:
Can you talk a bit about Lorenz invariance in emergent spacetime? Is it a challenge to come up with quantum theories where Lorenz invariance emerges rather than is just assumed? Is it a helpful constraint on theories of emergent spacetime, or not really?
Moshe Feder
In the late 90s, Randall and Sundrum proposed two variations of a five-dimensional anti-de Sitter space model to solve the hierarchy problem of gravity’s weakness relative to the other forces.
What did you think of their proposals at the time and has any further work been done on them since?
Nate D
I’m a bit confused about the relevance of the various “interpretations” of quantum mechanics for experimental physics. Are the formalisms underneath all of the interpretations so consistent as to make one’s theoretical preference as an experimentalist—let’s say, for Many Worlds over Pilot Wave—irrelevant? Or would “believing” in Many Worlds over Pilot Wave potentially generate different experimental arrangements? Basically, this is just a question about the relationship between theory and practice in contemporary quantum mechanics.
Sonja
What are your thoughts on the recent elections in Serbia and the key role played by university students and members of the academic community?
Benjamin Zand
In general relativity spacetime has a definite curvature at every point, described by the Riemann tensor. If we measured a traveler’s speed relative to this curved geometry, with each point having its own curvature, wouldn’t that give us an absolute standard of motion? Or do the components of the curvature tensor change with the observer’s velocity in a way that keeps the principle of relativity intact? If the latter, where/how does observers velocity enter into the equations?
Jake
PRIORITY QUESTION: In a previous AMA, you argued that the morality of painlessly killing hypothetical kittens depends on the reason, and that killing them purely for enjoyment would be morally wrong.
Would you also consider it morally wrong to kill an animal for a tastier meal when a nutritionally equivalent plant-based alternative is available, since in both cases the reason for killing is enjoyment? I understand why taking pleasure in killing might be morally objectionable, but does the seemingly innocent pleasure of eating justify the killing itself?
-
People at AI labs appear to be freaking out that AI research looks like Go or theorem proving. AIs suck at most things, but in the narrow domains where AIs can get fast and objectively true signal on their performance (they won/lost the Go game, their hacking exploit worked/failed, their LEAN proof was valid/invalid), AIs can run an ungodly number of trials and teach themselves to outclass humans through brute force. The bottlenecks of AI research seem to have fast and objectively true signal (“prediction loss went up/down”), and the labs are now spending billions per quarter trying to train AIs to do AI research. Should we be freaked out?
Dennis Briggs
I tell my students the linear systems don’t know when enough is enough. Is Many Worlds the ultimate example of that aphorism, or am I missing something?
George Hampton
Given the state of our democracy, I've been thinking about Gödel's Loophole and am wondering if this is the 'inner inconsistency' he was thinking about. Do you have any thoughts on this?
Click to Show Full Transcript
0:00:00.3 Sean Carroll: Hello everyone, and welcome to the November 2026 Ask Me Anything edition of the Mindscape Podcast. I'm your host, Sean Carroll. As you're listening to this, we have had an election in the United States, November 2026, a midterm election. The questions in this month's AMA don't involve that election, mostly because they came in before the election happened, but I'm recording it after the election happened. For those of you 500 years from now who don't know what we're talking about, the Democratic Party, which was not in power, gained a huge amount of power. In particular, they won majorities in both the Senate and the House of Representatives in Congress. Still don't have the presidency, of course; there was not a presidential election. Still don't have the Supreme Court. Supreme Court is supposed to be completely apolitical, but that is hilariously untrue. The Supreme Court is very political and very much on the conservative side these days. But they did, the Democrats, gain both houses of Congress.
0:00:57.0 SC: We're not gonna go into details about the election results, except to mention a structural feature that is kind of amusing. This election had a couple of very explicit examples of what is known as a dummymander. Again, for those of you who aren't following US politics or don't know how things work, there's this idea called a gerrymander, which is when you have certain conditions in place, namely number one, you have winner-take-all voting system. First-past-the-post, it's sometimes called. So in any one geographic district, or any one district in general, whoever just gets the most votes wins. Secondly, you have geographic districts, so you have boundaries drawn in literal two-dimensional maps that determine what district is being represented by somebody. And then third, you have the ability to draw these maps put in the hands of perhaps partisan governments. So if you're already in the majority, if you're in control of some state, you can under certain circumstances draw the maps however you want to decide what counts as a certain congressional district.
0:02:10.3 SC: All of these features, majority vote, geographic districts, putting the drawing of the districts in the hands of a perhaps partisan majority, are bad. They're not good features of a working government, but some of these things get in there, and once they're in there, they're very hard to get rid of. So we have all these in many places in the United States, and that opens the door to what is called gerrymandering. We actually talked about gerrymandering here on the podcast from a mathematical point of view with Jordan Ellenberg quite a while ago. How well can you do it? How can you avoid doing it? How can you look at a map and know whether or not it's been gerrymandered or not?
0:02:48.0 SC: So the idea of gerrymandering is the following. Imagine you have some number of districts, 10 districts in a state, and imagine that 50% of the voters typically are one party and 50% are the other. You'd expect that on average, of those 10 districts, five of them would end up being represented by one party and the other five by the other party. But what you can do is you can use the kind of detailed demographic information that people now have to carefully draw the districts so that the party that is not in power gets put in a small number of districts where almost all their voters are. So you could have, you know, two districts with a huge number of Party A, and then eight districts with a small majority for Party B. So that even though in population terms it's 50-50 Party A, Party B, in majority in the district terms, Party B is now way ahead. Okay, that's a gerrymander, because these districts that end up doing that end up looking very complicated. And Elbridge Gerry was the politician who became notable for doing it, and one of the districts that he drew, or I forget whether he was the one who drew them or was voted into it, but it sort of looked like a salamander in the eyes of a political cartoonist.
0:04:07.3 SC: Anyway, the dummymander is people aren't actually stuck forever voting for one party or the other. They can change their minds. And in a midterm election, which is very often pushing against the party in power, people just don't like the party in power. That's a very standard feature of democracies. So it can be the case that even though overall, usually on average it's 50-50, maybe in this election, there's a strong majority of people voting for Party A over Party B. And what you've done is arrange it so that your party has a majority usually in most districts, but that majority is small. So if there's sort of a wave election that pushes many people to the other party, suddenly you can actually have fewer seats than you think you deserved, because in the districts that you tried to make 60-40, they're actually majority the other way. This is called the dummymander.
0:05:14.2 SC: And there's evidence that precisely this happened in various cases in this year's election, which, of course, is bringing some guilty pleasure to the party that benefits from this, some sadness to the party that loses. But it's certainly an advertisement for trying to do better when organizing your political system so that things like this should just never happen in the first place. And that's where I come out here.
0:05:39.8 SC: Anyway, we're not talking about that during the AMA, but I thought it was something to bring up because we're still very much in the period of time where we're dealing with the implications of these new election results, pro and con, people reacting in different ways. It's kind of amusing from a spectator's point of view. Thanks as always to the folks who support these Ask Me Anything episodes. Those are the folks who pledge on Patreon. That could be you. You could go to patreon.com/seanmcarroll. And these days we've switched over Patreon's way of paying things, so you pay by the month, not by the episode, but it's essentially the same thing. And if you think you get a dollar or two of pleasure or reward from a typical Mindscape episode, we would love it if you pledged that much support. And as a benefit, you get ad-free versions of the podcast. Unless you like the ads—you're still allowed to listen to the ads, I suppose. And you also get the right to ask these Ask Me Anything questions.
0:06:43.7 SC: So all the questions you're about to hear come from Patreon supporters who we greatly appreciate for their support. And with that, let's go. Peter says, "If you could have a superpower, what would it be? Flight, super strength, super speed, teleportation, telekinesis, et cetera." Classic question. Mostly a classic question when you're 15 years old. But that's okay. We all have the inner 15-year-old who really cares about these things. My standard answer would be teleportation. Teleportation. But this is an old person's answer. Okay? It's just a pain to have to travel in these awkward ways. Just last week, as I'm recording this, I zipped down very quickly to Santa Fe to visit the Santa Fe Institute, where we were having a small workshop on timelessness. At SFI, I have been serving on a small committee that's been organizing a series of events about time and the arrow of time and things like that. And this was sort of the last event in that series that's been going on for years.
0:07:52.1 SC: And so I felt I had to go. I had an obligation to go. Of course, it's fun to go. I don't really have time to go, but it was on timelessness, which is a certain delicious irony. So I went. And the point of it is, I was trying very hard to get there in a hurry to sort of miss the fewest classes that I'm supposed to be teaching while also making it for as much of the meeting as I could. And I came very, very close to missing a connection while I was flying out there. In fact, I came very close to missing connections both ways, flying out there and flying back, because I was flying directly to Santa Fe. Sorry, the word directly is wrong there. I was flying to Santa Fe rather than to Albuquerque, which is where I usually go. Usually you go to Albuquerque, rent a car, drive an hour to Santa Fe. That's perfectly convenient. And there's a lot more flights to Albuquerque than there are to the tiny little Santa Fe airport. But I wanted to get there in a hurry, so I flew to Santa Fe, which means there's no direct flights from Baltimore. And what that meant was that I was sitting in O'Hare working on my talk, as one does, and they switched the gate on me. And I never noticed until very late.
0:09:04.1 SC: Anyway, I did notice in time. So I got to the correct gate, which is in a completely different, not terminal, but region of the terminal. And so that was a close call. And again, irony, because I was going to a workshop on timelessness and the existence of time was very important. And then likewise, when I came back, there was rain or snow or something. It couldn't have been snow. It must have been rain and fog in Denver, which delayed my flight into Denver, but not my flight out. So once again, I came very close to missing that. So anyway, the point is, with teleportation, presumably, I'd be able to avoid all that, right? I could just walk out the door and be in Santa Fe instantly later. Now, there's practicalities. If I was the only person in the world who could do teleportation, then are you sure when you teleport somewhere that you know exactly where you're teleporting to and that there's not a person standing there and you land inside their body and explode or something like that? I think you'd have to work out a lot of the practicalities there. So maybe teleportation is less good than it seems. But if everyone could teleport, I mean, the world would be a very bad, messy place, I think. That would be good in some ways, but bad in other ways. So maybe I should grow out of my thought that having teleportation ability would be the best.
0:10:24.3 SC: Again, since I'm an old person now, I don't know if it's a superpower, but health would be a wonderful... My health is fine right now, but it doesn't last forever. And if you had these healing factors like Wolverine has, and you could just get better... I don't know, does that work for a stomach flu and things like that? I don't know. It works when you're stabbed, but I don't know what the general theory of the healing factor would be. But to bounce back from small indignities in your body would be very, very useful right now. I think, honestly, thinking about telepathy and things like that, seeing into other people's minds, I don't want anyone to see into my mind and I don't want to see into anyone else's. I think that talking is a perfectly good way of doing things. Flying was... Maybe flying is the compromise. If you could fly very quickly, I think once again, there's practicalities. It gets cold up there and there's a lot of drones flying around in the atmosphere these days. So if you had flying without the ability to heal, you might bump into a drone and hurt yourself very badly. So this superpower thing, it's less good than the comic books make it out to be. There's all sorts of things, higher order questions you need to take into account.
0:11:41.6 SC: James McBeath said, "In the September 2026 AMA, you said the eternal Schwarzschild solution..." That's the solution to Einstein's equations in general relativity for a spherically symmetric empty space distribution. "Anyway, the eternal Schwarzschild solution has all sorts of weird properties. It has not only a black hole, but has a white hole in the past out of which things can come but nothing can go in. But the Einstein-Rosen bridge and the white hole are not really thought to exist in the world. But we study them because there's a simple mathematical setup to look at first."
0:12:12.8 SC: And then James continues, "In other podcasts, you said that scientists should have the courage to accept what the math is telling them. How do you decide when to accept the math and when to assume the predictions don't really exist?" It's an excellent question and it points to a little bit of sloppiness on my part in saying things like you should accept the math. The important thing is you should accept the math... You should never accept the math in the sense that you just believe it no matter what. What I think is that you should take seriously what the math about your theory is trying to tell you to the extent that you think your theory is right and you've set up the mathematical formulation of the problem correctly. So in the case of the eternal Schwarzschild black hole, the math isn't telling you that the eternal Schwarzschild black hole exists. The eternal Schwarzschild black hole is a mathematical solution to Einstein's equation in an otherwise entirely empty universe. But we don't live in an otherwise entirely empty universe.
0:13:18.2 SC: The black holes that we care about in our universe are created from stars, right? Stars collapsing or other things that are analogous to stars. So the point is that there are absolutely black hole solutions that are relevant in the real world, but they're solutions that don't start with a white hole. They start with a star or some collection of matter and then the black hole develops over time. So the eternal Schwarzschild black hole is a mathematical solution, but not one of any physical relevance. And it's not a matter of not having the courage to accept it. It's just not an expression of the physical situation we care about. That's in contrast to the very existence of black holes, which was kind of implicit in Schwarzschild's original solution in 1917. And no one believed it. No one really paid attention. No one even figured out what it was trying to say until the 1950s and people like David Finkelstein.
0:14:15.1 SC: And also things like the Big Bang. Oh, there's a singularity at the beginning of the universe. Things like the multiverse, if you believe in Many-Worlds interpretation of quantum mechanics. Now, all of these are cases where physically realistic initial setups imply that the equations are telling you something dramatic. So there you either need to have the courage to accept what the math is telling you or you change the math. You change the theory, you believe in a different kind of theory, and that's also fine. I'm not against changing the theory if the theory doesn't seem to be leading you where you want it to. What I'm against is having a theory that you accept, having a well-formulated idea of what initial conditions are and what solutions should be relevant, and then just not choosing... Choosing not to quite accept the implications of your theory given those conditions. That's not what's going on with the eternal Schwarzschild black hole and its wormhole.
0:15:13.8 SC: Friedrich says, "As a European, I regularly have to watch with horror how the American election system creates really poor outcomes for everyone involved. But from a physics of democracy perspective, aren't you stuck somehow? The politicians who would change anything are the ones who are heavily dependent on its existence. And psychologically, most voters can probably not fathom any alternative without having lived abroad for some time. Do you see a path that would work to move to enable better outcomes without destabilizing society on a catastrophic level?" I think this is a super good question. I think it's more just a democracy question and an institutional structure question than a physics of democracy question, but still an excellent question. If I can just rephrase it in much more general terms, we had a period here in the United States 250 years ago where we were thinking about how to structure the government. This is something actually in the book, The Physics of Democracy, that I'm supposed to be working on. I mention this as a contrast.
0:16:14.4 SC: Between physics notions of emergence and coarse graining versus political notions of emergence and coarse graining. So part of the book is saying that certain analogies or metaphors are useful. They shed light, they illuminate on what's going on in political situations. But also I try to say, okay, the following things don't work. I have a little screed in there why quantum mechanics is really not relevant to political or social situations at all. And the story of emergence and taking individuals in a large number and having them collectively work to get some goal or something like that is very similar in politics and physics. But there are important differences. And the most important difference is in physics, the correct way to coarse grain, that is to say, the particular ways in which you throw out some microscopic information to keep certain macroscopic variables, is fixed by the physical situation. It's fixed by the demand that you get a predictive theory that uses only the macroscopic variables that you've coarse grained away from the microscopic variables.
0:17:31.2 SC: In politics, you're not interested in a predictive theory. I mean, maybe you want a predictive theory of politics, that's fine. But when you're inventing a voting or representation system, the goal, the standard of success, is correctly representing the will of the people, whatever that means. And you have to decide what mechanism it is to do that. And the ability to predict the future is just not a standard for that, for the success of representing what the people want. So 250 years ago, people were debating this, and it was an interesting kind of discussion because democracy theory was still relatively primitive and young compared to where it is now. The United States in its current system of government is actually old compared to many other current systems of governments in other countries. And so they made some choices, and some choices they made were good. And some choices they made were bad.
0:18:29.4 SC: Overall, the people, though, were a fairly impressive group of people. And they were fairly thoughtful. And of course they disagreed with each other about important things, but they were trying their best to do a good job, I think. And now, as Friedrich points out, there are problems. So one issue is just this issue of winner-take-all, majority rule, right? It's a little bit different. So there's sort of two things. One is your voting system, which is to say, how do you pick the representative? But then there's also your representation system, which is what is the set of representatives that you're picking? I think to Americans, the idea that you have a geographically chosen district, like you draw a line on a map and inside is your district and outside is everything else, and one person represents that district, that's almost baked in. They don't even imagine there could be a difference from that system to something else.
0:19:32.0 SC: So they will examine ways of choosing those representatives with different voting systems. It doesn't need to be plurality votes or majority votes. It could be ranked choice voting or approval voting or there's other systems that are interesting there, but the idea that there should be a representative for a district is kind of seen as self-evident to a lot of Americans. It's not self-evident. You can do other things. You can do systems of proportional representation. If you have 60% of the people in one district members of one political party and 40% members of another one, you could have 10 representatives for that area. Maybe the area is 10 times as big as a typical US district. But okay, we can debate that. But you could have 10 people represent, and then you could have situations where rather than voting for individuals, the voters vote for a party. And if the party gets 60% of the votes in a 10-representative district, they get to pick six representatives, and the 40% party gets to pick four.
0:20:36.9 SC: Now, Americans don't like that. You tell them that, you even suggest it, and they kind of recoil, in part because they don't like political parties. And that's in part because political parties are kind of decorative in some sense in the United States. I mean, we have political parties, we split up into them, but the party apparatuses typically don't have the level of power and decision-making that they do in parliamentary systems.
0:21:00.1 SC: So anyway, this is all to say there are alternative systems, like Friedrich mentions. Many European countries use various kinds of alternative systems. These alternative systems, by the way, have their own issues. Because when you have a system, for example, like proportional representation, then, okay, what if you don't have two parties? What if you have more than two parties? It's still very possible for people to be not represented effectively, and there's different sorts of dynamics going on that get in the way. But I think... Sorry, I'm just sort of talking around the problem. The actual question Friedrich is asking is effectively, could it get changed? Could you imagine, I don't know whether for some questions you'd have to amend the Constitution to the US, for other questions you could do it just through the ordinary legislative process, but could you implement a better system?
0:21:56.3 SC: And actually, I kind of very much feel the worry that Friedrich is implying here, that I kind of don't want to upset the applecart too much because I don't have trust that the people who would do the redesign would be as smart and as civically minded as the people who did the designing 250 years ago. I do think there's no question in my mind that there are absolutely better systems than the United States has. And I don't know what the perfect system is, but in many ways, the one that we have here is obviously bad and could be improved. Do I have the faith or trust that the people in power now would effectively improve it? No, I do not. Do I have the trust that there is any mechanism that could choose people to be put in charge and, let's say, have a new Constitutional Convention that would completely from scratch rewrite the Constitution of the United States? Oh my God, no. That's a very, very frightening prospect to me. I just don't think that the people who would be chosen for that would be good people.
0:23:03.6 SC: It could be. It's possible to imagine good people really, really improving the governmental system of the United States. And by good, I don't mean on my side of the political divides. I just mean smart people who know about politics and who are motivated by what's best for the country, not what's best for their particular partisan side. I'm just not sure how we would ever make that happen. So I think that the fact that the United States Constitution allows for itself to be amended, but that amending it is difficult, I think that was a smart move on the part of the authors of the Constitution. You should be able to fix it, but it should be hard to fix it because it's too easy to break it. And even though it's clear that it could be improved, not changing it is better than making it dramatically, dramatically worse. So maybe I'm being too pessimistic there. I don't know what the real political scientists would think of that. I would love it if we had a slightly more sensible political system.
0:24:01.2 SC: Lucas Pollard says, "Have you ever significantly changed your mind about a scientific or philosophical question without encountering new evidence, simply because you came to understand the existing arguments differently?" Sure, yeah, all the time. I mean, I'm not even quite sure how to come up with an example for that, but I mean, I think it's absolutely true that as one thinks about things, one can change one's mind. In fact, I do think I'm not gonna be very good at giving you examples just because there are too many. I mean, maybe moral objectivity is an example for me. Like, I used to think years ago that you could be scientific about morality and derive principles of morality roughly based on the idea that we had evolved through natural selection to survive, and the things that would make us survive and flourish can be mapped onto moral principles.
0:24:55.0 SC: Now, just by thinking about it, and more data came in, but just by thinking about it, reading and listening to other people, I understand that that was wishful thinking. That, of course, we do evolve through natural selection, and from that we get preferences and intuitions and inclinations, but those are not morally objective facts, suppose I should say objective moral facts. They are the instincts that we get from natural selection. Those are different than what our moral principles should be. So now I'm not a moral realist anymore. But I think that more generally, there's an interesting question here about what it means to be a good Bayesian. We often mention being a good Bayesian here on Mindscape, and we cast it in terms of: you have credences over a set of propositions, and new information comes in, new data, and you update your credences based on the likelihood of getting that data.
0:25:51.4 SC: Of course, new ideas or new thoughts should count as new data. But do they count as new, how do we count them as new data? This is an excellent question that I don't know the answer to. I almost said the answer is not known, but probably someone knows the answer, or at least people have very well-thought-out opinions about these things. I am not up on the best writings or thinkings about these exact issues. In fact, now that I'm thinking about it, I have a better example. A classic example is, believe it or not, the electroweak theory. Steven Weinberg, in, I don't know, 1967, I think it was, wrote his famous paper, "A Theory of Leptons," where he showed how you could unify electromagnetism and the weak nuclear force. And no one paid any attention. It was a very short paper, appeared in Physical Review Letters. Weinberg himself didn't pay... He never sort of cited his own paper in the papers that he wrote in the years after that. And we know why. Well, we know part of the reason why.
0:26:55.5 SC: One reason why is just it was a speculative idea. But another reason why is there was a theoretical argument that said it can't work, namely that the theory seemed to be non-renormalizable, which is another way of saying that if you did calculations, you would get infinities that you can't get rid of. So it didn't seem to be physically plausible or physically promising. And then it was approximately 1972 that Gerard 't Hooft, and in part in collaboration with Tini Veltman, showed that theories of that form, with a Higgs mechanism and spontaneous symmetry breaking and things like that, are renormalizable after all. And then suddenly it became hugely popular. So five years later, it went from no one paying attention to hugely popular. No new data had appeared. All that had appeared was an improved understanding of what the theory was really about. So things like that happen all the time in science, not to mention in philosophy. And so that's an example where it's not just me changing my mind, it's the community as a whole.
0:28:01.9 SC: Rue Phillips says, "I love the recent episode with Greg Caruso. I know you're not a fan of free will discussions, but keep them coming. I'm curious how you would compare and contrast your compatibilist views with Greg's free will skeptic views, specifically on what changes to the US legal system should theoretically be pursued. I'm having a hard time understanding the differences between compatibilists and skeptics like Greg, and I'm hoping that when I should..." Sorry, I read that without the correct pausing. "I'm having a hard time understanding the differences between compatibilists and skeptics like Greg. And I'm hoping that when applied to a specific case, like how we might modify the justice system, might make it a little more clear."
0:28:39.6 SC: Yeah, unfortunately, or for the question, but not unfortunately for the world, I don't think there is much of a difference. I think what came out in that discussion, and by the way, no free will discussions coming up anytime soon. I did Christian List and that was enough for a year, but he was on one side and I thought it was fair to have someone from the other side. And Greg is a very well-spoken and well-known articulator of the other view. So we did both of those. That's enough for a while. We can cleanse our palates now and not have more free will discussions. But I think that what came out from talking to Greg was his motivation for declining the offer to use free will vocabulary to discuss human beings making choices, which he admits human beings make choices even though they're completely physical creatures. He just doesn't want to call it free will. Why not? Well, because to him, calling it free will implies this kind of moral desert. It implies a kind of way of thinking about human beings where if you make the wrong choices, you are bad and you deserve to be punished.
0:29:51.8 SC: To me, that just doesn't follow at all. That's a separate thing. Okay. I can absolutely imagine applying the vocabulary of human beings thinking things through and making choices, making the wrong choices, or bad or evil choices, whatever you want to call them, and dealing with them in ways that do not simply involve punishing them because they're bad. That just is a non sequitur. They're separate questions. One is the question of how we analyze and talk about human behavior. The other is how we organize the justice system and our ideas of what people deserve and things like that. So I don't see any real difference. As a compatibilist, I'm completely happy with a point of view that says we should try to, number one, rehabilitate people who have behaved badly. Number two, deter them from doing further bad things. And those should be the central focuses of the justice system, not punishing people because they are bad.
0:30:56.0 SC: Jeffrey Siegel says, "I was wondering whether you think it is just as well that there does not seem to be the possibility of faster-than-light travel or communication. As I think von Neumann pointed out, since our species has evolved rather late in the development of the universe, if faster-than-light travel were possible, it's likely that a species that developed earlier might have already spread through the galaxy, including Earth, and blocked the evolution of humans. Similarly, with faster-than-light communication, a more advanced civilization might be able to hack our technology and doom humanity. Given those possible dangers, do you nevertheless hope that faster-than-light travel or communication may be possible, or do you breathe a sigh of relief that it is not?"
0:31:35.4 SC: So I think I wouldn't exactly characterize von Neumann's point in quite that way. It is true that we have a speed of light barrier on both travel and communication, but it is not true that this is getting in the way of colonizing the galaxy or spreading throughout the galaxy. The galaxy is big, but it's not nearly as big in light-years as the age of the universe in years. In fact, the size of the galaxy, it depends on... The galaxy doesn't have a sharp edge, so you can debate how big the galaxy is, but the typical scale is tens of thousands of light-years in linear size. So the universe is billions of years old. And you can say, well, it's hard to travel near the speed of light, but maybe that would be hard even if the speed of light were not a speed limit. It's still hard to build giant rocket engines or whatever it is.
0:32:34.5 SC: So to make it easier for you or more acceptable, maybe you say, okay, well, people in advanced civilizations can build rocket ships that can go 1% the speed of light. And if you have a 10,000 light-year sized galaxy and you go 1% the speed of light, okay, it takes you 10 million years to get across the galaxy. And even if you say it takes time for civilizations to evolve and whatever, 10 million years is nothing. It's just nothing compared to the age of stars and solar systems and things like that. Even life on Earth started 4 billion years ago, right? So it's not really a barrier. Like that's... I think, that was Fermi's point. I don't know whether von Neumann said it or not. But even moving at a relatively slow pace, even if you stop... It's more likely, rather than saying that you move at 1% the speed of light, it's more likely that you move at almost the speed of light in a realistic rocket ship, but then you stop and build some settlement on a planet or whatever and then you move on. Still, it wouldn't take billions and billions of years to fill the galaxy if that was what was going on.
0:33:44.9 SC: But okay, putting that aside to Jeffrey's question, I'm not sure that it makes a big difference that there's a speed of light barrier. We're not building spaceships that go anywhere close to the speed of light, right? So it's a little bit science-fictiony to think that, oh, that's what's holding me back. Like if it weren't for special relativity, I'd be flying across the galaxy by now, right? I mean, there's other barriers out there. I think it's fine. It doesn't really... My personal preferences are not relevant to there being an ability to go faster than light or not. I think that the more you dig into how physics works, the existence of some kind of speed limit does play a role, whether it's the speed of light or not. But the fact that if I poke the universe right here, the influence of that poke does not spread throughout all of spacetime instantly is important to how things work.
0:34:41.0 SC: I'm not worried about people hacking our technology. I think that that's a little bit harder than people make it out to be. And who knows? By the time we human beings do have the ability to travel to distant stars within our galaxy, maybe we'll just be a different kind of species than we are now, whether biologically or even just morally, right? Maybe we won't want to live on other planets that might someday develop life. Maybe we'll just build habitats out in interstellar space or interplanetary space around some star, using materials from planets that have no chance ever of developing life. I don't know. And I don't know whether the other species would feel similarly to that or not. So I don't think that anyone ever will design faster-than-light travel, and it doesn't really either bother me nor give me a sense of relief one way or the other.
0:35:38.0 SC: Halfacrackpot says, "My question for you is as an author. Let's say, hypothetically speaking, mind you, that I own copies of a few of your books. I get to see them sitting on my shelf in all their glory every time I walk into my home office. Then one day I happen to find PDFs of those same books that fell off the back of a truck. I tried my best to chase them down to return them, but alas, I didn't even manage to get their plate. Would you be personally offended or feel wronged if I used those digital copies that were not paid for?"
0:36:06.1 SC: Well, look, I don't think that the point is whether I would be personally offended or not. I am all in favor, for example, of people reading books in libraries, even though authors don't get as much money if their books are read from libraries rather than purchased at a bookstore. But I think there's an economy. This is a collective action problem, right? One person getting a book or an album or whatever, or a movie, and not paying for it is no big deal. That's not the moral problem. The problem is if you like things, if you like things that other people create and take time to do that, whether it's, again, books or music or whatever, then you want to incentivize those people to do it. Once those people are multimillionaires, the incentive might be a little bit less, but still, the basic idea is that you want to pay people for the services that they provide you, if only not for moral reasons, but because that keeps them providing those services.
0:37:08.4 SC: There's also moral reasons. I mean, I think that there is some legitimacy to saying that it makes sense morally for there to be an exchange of goods and services, et cetera. We live in this weird world where a lot of stuff is available for free, some stuff isn't. There's this whole model of advertising, right? Most, the vast majority of people who listen to this podcast do so for free, but they have to listen to ads to do it. And they will nevertheless complain about the ads, even though they're listening to a free product, and even though they also have the option of just paying for the product by going to Patreon and then not getting the ads. They will still complain about the ads. I think that people just should generally have a feeling that you pay for things that you get, and that's not bad. And of course, different people are in different circumstances. Some people can't afford to pay or whatever. The world is complicated, but I don't think that there's any moral rectitude in taking stuff for free that you should be paying for. Of course, that's not to say that various corporations or whatever don't try to bleed you dry, and that's bad, and you can talk about that. But the general idea of an economy of goods and services where people pay for what they get, I think is perfectly good.
0:38:26.6 SC: Perry Huan says, "I'm having trouble thinking through the differences between an infinite universe that may or may not be repeating and a continuous universe that just bubbles off from itself. Is there an intuitive way to think about these distinctions?" Well, I'm not sure what you mean by a continuous universe that just bubbles off from itself. There are models, including the one that I explored with Jennifer Chen, of a universe that does have what we call baby universes, child universes, that bubble off from the original ones. But the baby universes are not exact copies of the original universe. They sort of go their own way. So I think that there's a big difference between a repeating universe and a sort of ever-growing universe. Those are two different things. There's different ways to be infinite, basically. You can be infinite by simply repeating yourself over and over again. You can be infinite by constantly changing. I think that in the case of the universe or cosmology, both of those possibilities are 100% on the table.
0:39:29.8 SC: I propose theories. I have papers where I write and I sketch out some cosmological model. And sometimes people who are not scientists, professional physicists or whatever, seem to think that because I'm investigating a model, I must be advocating that it's true. I mean, it does have a certain sensibility, right? Like, why would you suggest this model if you didn't think it was true? Why would you suggest a false model? The point is that it's not true or false. We don't know. We're exploring the space of possibilities, right? So to know that something is a viable possibility is important, useful information. In the case of the most recent paper I wrote with Saakshi Dulani and Nadiia Diachenko, we have a cyclic universe that we claim does not have the Boltzmann brain problem, unlike other cyclic universes do. And we made the point that we're pointing out a loophole, that if you'd asked me before, I would have said that this kind of cyclic universe is not viable, because of the Boltzmann brain problem, we realized that under certain circumstances, it can be. That doesn't mean I think the universe goes that way. It means I think the universe might go that way.
0:40:41.3 SC: And sorry for Perry, I'm going away from the original question, but it inspired me to mention this idea that this is what we do as theoretical physicists. We propose scenarios. Like Stephen Weinberg in 1967 didn't actually think, because he proposed the electroweak theory, that it was correct necessarily. He thought it was a chance of it being correct. It turned out to be correct and one of the most important contributions in the history of theoretical physics. But you don't know, and that's okay. So likewise for the universe being infinite it might be infinite, and it might be infinite in a way that is never repeating, or it might be infinite in a way that is repeating, or it might not be infinite at all. All of those are very viable. We should be very humble when it comes to thinking about the very, very large-scale structure of the cosmos.
0:41:27.6 SC: Kevin's Disobedience says, "You don't strike me as someone who is inspired by any one particular philosopher. But if you were going to teach a course on one person's philosophy, who would it be?" That's a good question. You're right that I'm not inspired by any one particular philosopher. I'm not... In the sense that I'm not a follower of any one particular philosopher, I borrow ideas from whoever I think has good ideas. And also, the other thing to keep in mind is that if you teach a course on somebody, that doesn't mean that you're a follower of them, right? Doesn't even mean you're temporarily a follower of them. Maybe you just think it's interesting to talk about what they're saying. So honestly, there's no philosopher right now who I think I could do a good job starting tomorrow teaching a course about. Like I said, I pick and choose from what different philosophers have to say.
0:42:20.3 SC: I mean, unless it was a very modern contemporary philosopher of physics, right? I could teach a course on David Albert or David Wallace or Tim Maudlin or whatever, but that would be weird. Those people are still alive. They should teach their own courses. So probably I would go the other way around. Just for my own interest, I would teach a course on Plato or Aristotle or a group of old philosophers, the Atomists or the Sophists or whatever, because I think that'd be fascinating and useful. I would love to teach a course in ancient Chinese philosophy, but I'm incredibly woefully underqualified to do that, so I wouldn't dare. I know just enough about Plato and Aristotle that I could learn enough to teach a course. And someone like Plato is very, very entertaining in the writing. He had some crazy ideas. Aristotle's ideas were largely, I think, closer to reality than Plato's, but he's not nearly as engaging a writer. So it goes back and forth.
0:43:19.7 SC: And of course, then it also depends. Do you count Isaac Newton as a philosopher? Was Galileo a philosopher? Teaching a course on Galileo, I think, would be amazingly fun. Newton would be much harder because the Principia is really hard to read. Newton was not a good writer in any sense, not a good explainer of what he was talking about. So Newtonianism is very interesting, but I don't think I could teach a course on Isaac Newton. I could teach a course on classical mechanics in a modern formulation of it, but not the original stuff. No, that's hard to figure out what's going on there.
0:43:55.2 SC: Alright, I'm gonna group together three questions. Alex West says, "There seems to be a split opinion on OpenAI withholding their Yang-Mills existence proof from public view. Thoughts?" Alexander Knuckle says, "I would love to hear your opinion about the latest Yang-Mills claims by OpenAI. I'm personally very unhappy how the results are communicated, but of course I'm very interested in learning more about rigorous definitions of Yang-Mills and the standard model. PS: How do you view the significance of the actual result, irrespective of how it was communicated?" And finally, Tyler Whitmer says, "I have a sort of meta question about the Yang-Mills OpenAI announcement. I was struck by the different ways the various constituencies engaged by OpenAI in the math and physics worlds asked for the discoveries to be handled by OpenAI. Do you have any thoughts on how you'd prefer these things to be dealt with in the future, especially if they are directly affecting an area where you're a published expert?"
0:44:49.1 SC: So for those of you who don't know, there is a famous math/physics problem, one of the Millennium Problems, about the rigorous existence of quantum Yang-Mills theory. So Yang-Mills theory is just a generalization of electromagnetism. These are what we talk about when we talk about gauge theories. If you've read Quanta and Fields, my second volume in the Biggest Ideas in the Universe series, you know what we're talking about when we're talking about Yang-Mills theories. QCD, the theory of quantum chromodynamics, is the paradigmatic example of a Yang-Mills theory. In some sense, good old electromagnetism is a Yang-Mills theory, but it's too simple, so we don't really think about that too much.
0:45:35.3 SC: In another sense, the electroweak theory we were just talking about, where you have the Higgs boson and spontaneous symmetry breaking, is also a Yang-Mills theory, but it's a more complicated kind of Yang-Mills theory, precisely because of that symmetry breaking from the Higgs boson. So that's why QCD, the strong interactions with gluons and quarks, is the paradigmatic version. It's complicated enough to be interesting, not extra complicated by symmetry breaking, et cetera. And what you find, well, there's also different versions. You can think about what Yang-Mills theory, what QCD would be like if there were no quarks, right? If there were only gluons. You still get an interesting theory because gluons interact with each other. Photons, at the simplest level, at the first approximation, photons don't interact with each other. In the real world, photons interact with each other a little tiny bit, simply because photons interact with charged particles and charged particles interact with photons back, and therefore you get virtual quantum-induced indirect interactions between photons and other photons.
0:46:42.8 SC: But putting that aside, the gluons just interact with each other directly. So there's a whole bunch of questions you can ask in this world of rigorously defining quantum field theories. And I'm not in this world, really, because this is a heavily mathematical world, and it's one where there's a bunch of results that are sort of a little mysterious. There are theorems in what is called axiomatic quantum field theory where you can prove things that are a little bit surprising, things that don't seem to line up with our informal, less rigorous understanding of quantum field theory, and therefore there are debates back and forth about whether axiomatic quantum field theory is missing something important, whether we should just forget about being rigorous and use effective field theories and stuff like this. Anyway, that's just a little bit of the background.
0:47:33.5 SC: The other point is that if you choose your parameters and your fields correctly, then we know that a theory like QCD has this property called asymptotic freedom. Asymptotic freedom means as you look at the interaction between particles, in this case gluons, as a function of energy, the idea of renormalization and the renormalization group tells you that the strength of the interaction depends on the energy of the particles when they're interacting. And asymptotic freedom means the more energetic they are when they're interacting, the less of an interaction there is. So asymptotic means higher and higher energies, freedom means less and less interaction. So there is a couple of informal, and that's very important, informal implications of the idea of asymptotic freedom. The flip side of asymptotic freedom is infrared slavery. So as you get to lower and lower energies, the particles interact more and more, so much so that they become confined. So the fact that you never see a gluon or a quark all by itself out there in the wild is supposed to be related to this fact that the interactions become stronger when the fields or the particles become further apart from each other.
0:48:53.9 SC: In quantum field theory, large distances means low energies. So when we talk about low energies, the interactions become strong. The interactions become strong when the two particles, two quarks or two gluons, get pulled apart from each other. So the energies become so large that it's easier to make another quark-antiquark or gluon-antigluon pair, and therefore you can never just take apart a single particle from a strongly interacting collection of particles, and that's confinement. Along with confinement, there is a related feature that there is what is called a mass gap. So if you have a single photon, a single photon has zero mass and therefore it has whatever energy you like. Its energy depends on its wavelength: longer the wavelength, lower the energy. And you can imagine, at least as a thought experiment, arbitrarily long wavelengths and therefore arbitrarily low energies. And in principle, you could go all the way to zero. You can't quite reach zero because otherwise that photon wouldn't exist if it had exactly zero energy, but you can get as close as you want.
0:50:01.6 SC: So if you think about the set of possible energies that a single photon can have, they go from just above zero to whatever you want. With gluons, it's not like that. So you ask yourself, in the theory of gluons, what is the lowest mass that a particle can have? And you might have been told that the mass of the gluon is zero, but there are no gluons all by themselves. Okay. So what we're asking is, what is the lowest mass particle that actually exists all by itself? And in QCD, well, you would have to define, do you mean QCD with quarks or without quarks? With quarks, we know there are things like protons and neutrons. There's also mesons like pions and kaons and so forth. The mesons for these purposes actually don't count because they're not stable. They're not really particles that can exist arbitrarily long all by themselves.
0:50:59.1 SC: But a proton is stable in just QCD. In the electroweak theory, it becomes more complicated, okay, but otherwise, for a very good approximation, protons are pretty darn stable. And they don't have zero mass; they have a non-zero mass. And even if you didn't have quarks, you could imagine making something like a glueball, which is a collection of gluons made up into a single particle that could exist by itself. And the idea is that its mass would not be zero. Okay. In the real world, we think that if you can make glueballs, they would also be unstable, but if you didn't have any quarks, there would probably be stable glueballs lying around.
0:51:35.2 SC: Okay, so all of that is to say, in Yang-Mills theory, there's something called a mass gap because there is a smallest possible mass; it is not zero. So the photons, electromagnetism, have no mass gap. You can get as low energy as you want. In QCD, you think that there is a mass gap. So there's confinement and a mass gap. These are the characteristics that you think are important for your quantum versions of Yang-Mills theory. But we haven't been able to show any of these things in terms of rigorous proofs. We haven't even been able to show that Yang-Mills theory exists in the sense of what particle physicists mean by that. It exists in the sense that you can write down the basic principles of the theory, but you haven't proven that it exists in the sense that those basic principles add up to a well-defined, 100% rigorously defined theory. Turns out that because quantum field theory is hard and there's infinities everywhere, there's an infinite number of degrees of freedom, proving rigorous statements is just super duper hard. So one of these big open math problems is prove rigorously that there is a quantum Yang-Mills theory that exists, and if it's true, prove that there's a mass gap, prove there's confinement, things like that.
0:52:51.5 SC: So again, I don't do that kind of thing. It's a very, very niche, specialized topic, actually, because most physicists just assume that it works and get on with their lives. It seems to work, right? There are protons all around us. We can make predictions, calculations using QCD for things that happen at the Large Hadron Collider or wherever; they come out matching the observations. So this sort of mathematically rigorous level proving that the theory works is of less immediate relevance to the working physicist than to mathematicians. Still interesting and important, don't get me wrong. I'm just explaining that even though I think about theoretical physics, I don't spend a lot of time thinking about the existence or rigorous proofs involving Yang-Mills theory.
0:53:38.6 SC: So as with many interesting math problems recently, in the past couple weeks and months, there have been AI-assisted proofs. OpenAI in particular, as a company, has devoted some millions of dollars to throwing CPU cycles at proving mathematical questions one way or the other. And one of these, apparently, is the existence of Yang-Mills theory, a big, very, very important problem for mathematical physics. Okay, so that's the background. The extra bit of background is, as I understand, 'cause I've not followed it super duper closely, so anyone can correct me if I've gotten it a little bit wrong here. But the basic idea is OpenAI says they've proven it, and they say there is a verification of the proof. And what that means is you can write down the proof step by step and have it checked by a computer so you don't think you've made any mistakes. That's what people mean when they say they have a Lean verification. That doesn't mean they have a skinny verification; it means that there's a program called Lean that can check the logical consistency of your mathematical proofs. And so they say they've not only proven it, but that it's been checked by Lean, which means it's correct.
0:54:52.6 SC: Okay, but there's one thing to say that, there's one thing even for it to be true, there's yet another thing to understand what is going on. And that both means understanding as human beings what were the important missing steps that this computer used that the human beings who've been trying to prove this didn't do. What are the important conceptual moves that were necessary to getting this result? But also, are we sure that the result is correct even if it's been checked by the computer? All the computer can check is that the conclusions follow from the assumptions. Are you sure that the assumptions you put in are the ones that we care about and the ones that we would consider to be the right ones? And so what OpenAI did is rather than just releasing the proof to the world on some webpage or whatever, they assembled a small group of very, very high-level smart physicists led by David Gross, my former boss, Nobel Prize winner at KITP in Santa Barbara. And he had another bunch of people who were working with them, and they were given the proof. And as I understand it, they're gonna go through it, try to understand it, and then try to explain it to the rest of the world. And meanwhile, the rest of the world doesn't get to look at it.
0:56:10.1 SC: So if that is correct, if my understanding that far is correct, I see no problem whatsoever with this. In fact, I think it's kind of absurd to think that there's any problem with this. Look, if I do get a result as a theoretical physicist and it's in a file that I've typed up on my computer, I'm under no obligation to immediately release it to the world. In fact, I don't immediately release it to the world. I write a paper about it, I think about it, I decide what the implications are, what the best way to present it, and make figures and cite people and stuff like that. And that's what these human beings are doing. That seems 100% valid and, in fact, 100% morally right to me, as far as I can tell. I might be missing something; I might not know about some sketchiness going on there in the background. But to me, the idea that human beings are given access to the proof, they're now collaborators on the project, and they are trying to make sense of it and will explain it to everybody, that's just how science works. That's a completely everyday, conventional way of doing things.
0:57:18.5 SC: Now, we'll see how this plays out, and we'll see how it plays out in the future, et cetera. But I think that we're still not in any equilibrium phase in terms of how we balance what the AIs are able to do and what human scientists and mathematicians care about. That's something that's still very much in flux. This is an experiment in trying to figure out how to do that correctly. And look, I'm not a fan of OpenAI. I think I can criticize what they do for a lot of different things. This does not seem to be one of those times that I see anything going wrong.
0:57:53.8 SC: Caleb Urian or Yaryon says, "My son is a junior in high school and has expressed interest in pursuing a career in physics. I have an engineering background and have practical advice for him if he wanted to pursue engineering, but I'm curious what advice you have for a high school student that is considering physics."
0:58:10.3 SC: Yeah, so I'm asked questions like this all the time. I think I'm just gonna... I'm probably repeating myself. I don't remember what I said and what I don't. The simplest thing to say is it's very hard to give advice because everyone's situation is different. What is your son good at? What does he care about? What does he want to be? What kind of physics might he want to do? I do think that I feel comfortable saying that pursuing physics is generally good if you're interested in it. It may or may not lead to a career doing physics, but knowing physics and studying it and sort of tackling all the abstract difficulties of it is a worthwhile pursuit. So there's no reason to say, "Oh, no, don't do that. You're wasting your time." Even if you end up doing something else, the training is extremely valuable.
0:58:59.1 SC: In terms of him getting ahead, there are debates about whether you should sort of study things by yourself before you get to them in your regular educational curriculum. I'm a big fan of doing that. By all means, study things by yourself, especially nowadays when there's so many resources online. There's online courses and lecture notes and whatever. I don't see any reason to wait. When I was taking calculus in high school, our calculus teacher said, "Don't study anything that I haven't taught you because you'll learn it wrongly." I think that was just terrible advice. I think exactly the opposite of that.
0:59:35.3 SC: In terms of preparation, learn about math, learn about physics. It's hard when you're in high school to learn sort of interesting physics, right? You don't have the mathematical tools to learn the kind of physics that physicists do at a research level on an everyday basis. You're stuck doing like inclined planes and classical mechanics and things like that. That's fine. There's a certain amount of patience that is absolutely useful and necessary here. A lot of people get tripped up because they want to skip right to the most interesting stuff, and you gotta put in your hours, right? You have to learn the basics, the background before you can get there.
1:00:15.1 SC: So learning math, learning calculus and geometry, these days learning something about computers. How the computers are going to be useful is different from month to month, so I have difficulty giving you advice about that. Don't let the computers do the work for you. Use the computers to teach yourself by all means, but make sure the ideas get into your head and make sure you're able to use those ideas. I know that the computers can solve an equation for you, but you need to understand, you being your son, need to understand what it means to solve an equation and how you would do it.
1:00:51.3 SC: Because one plausible future of the relationship between humans and computers in physics is that humans are mostly interested in deciding what are the interesting problems and what are the directions in which you should look for a solution, and then they team up with the computers to turn those vague ideas about directions to look for a solution into a rigorous solution or to a careful solution. So working with the computers, both AI and not AI, but just computers very generally, is absolutely part of the skill set.
1:01:24.6 SC: Otherwise, you gotta go to college, so apply to college. I'm a pretty big believer that you can do well as a physics student as an undergraduate at most colleges. I think that people worry too much about getting to the absolute best places as an undergraduate. As a graduate student, the total number of people is much smaller. Who you're working with as your advisor is much more important. And choosing which graduate school to go to is crucially important and changes your life. When you're an undergraduate, as long as you're at a place that is big enough that they have enough resources, you cannot possibly take advantage of all the resources as a student there for four years studying as an undergraduate. So it's much more in your hands as the student than it depends on where you go to school.
1:02:12.0 SC: You can absolutely get a world-class physics education going to a big university, state school, right? Penn State or the University of Texas or whatever. You need to do it. It's really in your hands. If you go to a small liberal arts school, that can be a wonderfully useful and impactful experience in terms of becoming a good human being. It's not as good, I just have to be super honest here, even though it breaks my heart, it's not as good for being a great physicist. Because the definition of a small liberal arts school is that there's no graduate school there. So you can get a very good undergraduate education, but you can't dip into more advanced topics. You can't talk to grad students and postdocs, you can't see the professors doing research with grad students and postdocs, et cetera. There's a whole part of the experience that you just can't get.
1:03:07.7 SC: And as I say that, I say that as someone who was in that position. We didn't have any grad students, graduate programs in physics or astronomy or math when I was an undergraduate at Villanova. So I think that was a problem for me. I didn't get to take any advanced classes that I might have been able to take if there were graduate classes around. So that's the only thing that I would, if you want to be a research-level physicist, I would point to as something to really care about in your choice of universities. You can always overcome it. There's no yes or nos here. There's things that nudge you in one direction or another, and you have to decide how the nudges sum up for your personal situation.
1:03:48.7 SC: Philip Abard says, "Do you think there may be something in Nietzsche's idea of eternal recurrence if the block universe is to be taken literally? After all, if the entire history of the universe is out there, there should be nothing to stop you experiencing your life anew after you die." I probably should have grouped this one with the previous question because I don't think that the two things follow from each other. Philip's asking an interesting question, but the block universe is just a way of saying, as far as existence is concerned, all moments of time participate in existence equally: past, present, and future. There's nothing special about the present time other than the version of you that is in this time is calling that now. That's the only thing special about it.
1:04:31.8 SC: But there's nothing in that block universe perspective or eternalism that necessarily implies any opinion about eternal or finite. You can have a block universe that goes forever. You can have a block universe that is finite. You can have a block universe that goes in cycles. You can have a block universe that goes without cycles. All of those are possible. So I don't think there's any necessary connection there.
1:04:54.4 SC: The other thing I just want to warn, and I've warned many, many times, you shouldn't say things like "nothing to stop you experiencing your life anew after you die." You are a collection of particles in this particular region of spacetime, okay? That's who you are. If there happens to be at some future or some faraway, distant part of the multiverse, or distantly in the past, another collection of particles that is very, very similar to the you that is listening to this podcast right now, that's great, but it's not you. Because you're here, okay? That is another person who just happens to share a lot of characteristics with you, but it's not you experiencing your life again in any sense whatsoever.
1:05:46.4 SC: Joshua Hillerup says, "Would anthropic reasoning mean we should favor, with Fermi paradox explanations, ones that maximize the number of intelligent species out there who are capable of wondering about the Fermi paradox?"
1:05:59.2 SC: Well, people argue about the anthropic principle, so I'm not gonna say what the consensus is because I don't know what it is. My own view of the anthropic principle is that what it can do is to say that given who you are, okay, given what you know about your location in the universe and what you see around you and how you reconstruct your environment, all that stuff, it's fair to ask, given different cosmological models that you're trying to say, what should my credences be in these different versions of a theory of the universe? You should update them as a good Bayesian by the likelihood that at least one version of you with your knowledge and your environment and all that stuff would exist in this universe. If there are a million versions exactly the same or indistinguishable from each other, you don't get any benefit for that. You don't know which one of those you are. You can go on to ask what are the probabilities that I'm different versions of these people who are indistinguishable from me, so I don't think you get like a million times more probability just because there's a million more observers. I do think that you get extra probability or extra credence depending on the probability that at least one version of you exists.
1:07:16.4 SC: Now, the easiest and cheapest way for there to be a probability of there being at least one person like you is for there to be many, many, many people or many, many close calls in the universe. So I do think it's absolutely legit to have some boost to bigger universes, larger cosmologies over smaller ones, but not an absolute boost to an infinite number of observers over a finite number of ones. And having said all of that, I can easily imagine scenarios where in, let's say, the observable universe around us now, which has a few hundred billion galaxies, it's completely possible that we are the only intelligent species. And yet the number of regions like ours, observable universes the size of ours, is 10 to the 10 to the 10 to the 10, in which case there'd be lots of people like you, but only one in each person's observable universe. So I'm not quite sure that it necessarily has any connection to the Fermi paradox, although there's sort of an indirect connection you could think about.
1:08:23.5 SC: Callesin the First says, "If big wormholes à la Star Trek were real, what would be the safest way to enter them? In media, wormholes act like objects on a 2D plane and spacetime bends inward perpendicular to that plane. The spaceship then enters from a direction, sorry, perpendicular to the plane, right in the middle of the curvature. But in real life, they would be spherical and spaceships could not enter at an angle perpendicular to the plane. So could one enter from any direction? It's science fiction, so maybe anything goes."
1:08:57.9 SC: No, I think you're basically on the right track. A wormhole would be like a spherical region of space that had the property that when you look into it, you're seeing another spherical region of space on the other side, okay? So you're seeing a 360-degree view of a different part of the universe from the 360 degrees around the sphere you're looking at. And the sphere is not like a hard surface that you bump into; it's a tunnel, okay? So you literally just see the stars that would be on the other side of the sphere or whatever. So it's easy to imagine that a wormhole could be out there and be almost invisible, right? If the other end of the wormhole were in a region of spacetime that didn't have a lot of stars around it, it would just look black. You wouldn't even see anything. It would not be like the glowy thing. This, by the way, was one point of contention when they were making the movie Interstellar. Kip Thorne, who played a big role in the movie, tried to make them understand that wormholes would look a certain way.
1:10:03.3 SC: And the movie makers said, "Look, I know, but people have heard of wormholes before. They have a certain set of expectations, and if we completely violate those expectations, they're not gonna be happy." So there was some compromise on trying to make the wormhole look wormhole-like and yet be physically realistic. But basically, it's a sphere, and you can enter the spherical region in any direction you want. If you enter it glancingly, then you come out the other side glancingly. That's all that really happens.
1:10:34.9 SC: Tom B. Knight asks a priority question. I forget if I've already said this, but priority questions, you get a priority question once per your lifetime as a Patreon supporter, and I'll do my best to answer it because usually we get too many questions for me to answer all of them. So Tom's question is, "Apologies in advance because I know you don't like questions about free will. It seems that many people in this debate, for example, your recent guest Greg Caruso, try to eat their cake and have it too. They acknowledge someone's lack of free will when they make a bad decision, like committing a crime, then assume others have free will when deciding how to feel or respond. It seems the practical, logically consistent position is to treat everyone as if they have something resembling libertarian free will. What do you think?"
1:11:18.7 SC: Well, I think a couple of things. Greg Caruso, I shouldn't speak for him. I know people want me to explain the arguments put forward by my podcast guests, but I do that very reluctantly because hopefully their arguments are non-trivial and I shouldn't be the person repeating them. But I think Greg would say he doesn't believe in free will for anyone, making good decisions or bad decisions. A lot of his motivation for not believing in free will is that when people make bad decisions, he doesn't want them to be thought of as morally lesser, and therefore deserving of punishment or something like that. So I think that he's logically consistent. I think that he treats everyone the same; he just doesn't believe in free will.
1:12:02.5 SC: There's another logically consistent position, which is treating everyone as they have free will. Now, you say "something resembling libertarian free will." That makes me a little nervous. I'm not quite sure what that means. What does it mean to resemble libertarian free will? Libertarian free will, remember, is precisely when you think that your actions are not only incompletely specified and unpredictable at some higher emergent macroscopic level, but truly independent of even the microscopic laws of physics. A libertarian about free will believes that even Laplace's demon couldn't predict what you were going to do next, or even probabilistically given quantum mechanics or anything like that. So I'm not sure what use or what role the word "libertarian" is playing in your sentence. I do think that it's consistent and logical to treat everyone as if they have something resembling free will. And honestly, like I said, I don't care if people don't want to use that terminology, as long as they treat people as rational decision-making agents, that's fine with me. I certainly would not say that they resemble having libertarian free will.
1:13:11.5 SC: David Commander says, "How does a Boltzmann brain continue to survive? And if it can't, why isn't ongoing experiences in the moments of now insufficient to prove I am not one? In other words, if I can have false past memories, can I also have false ongoing experiences or the illusion thereof?"
1:13:28.5 SC: So I will say what I always say when it comes to Boltzmann brains: brains are not the issue. The idea of the Boltzmann fluctuation is that you tell me what it is you want to wait for to exist in the universe. Maybe it's a brain. Maybe it's a person who lives a hundred years. Maybe it's a whole solar system that lasts 4 billion years. I don't care. Whatever it is, in a randomly fluctuating universe, that thing will most often arise as a random fluctuation, not as a sensibly thermodynamic, rigorous evolution from an initially low-entropy state like we think actually happened in our actual universe. So everything is possible. Everything that is allowed by the laws of physics can randomly, in this scenario, can randomly appear. And yeah, it could... You could have false ongoing experiences. You could also have true ongoing experiences. You could fluctuate into something that was much lower entropy than the surrounding chaos and long enough for you to live for whatever time you want and have real experiences, and they'd be perfectly valid.
1:14:44.0 SC: But in addition to that, your extrapolation before the moment that you randomly fluctuated into existence would be wrong in principle. Now, there'd be plenty of fluctuations into things that were sort of kind of brain-like, or kind of observer-like, or kind of whatever, where they didn't think that they had wrong impressions of the world. That in fact is much more likely to not be blessed with these false memories. But that's not the point. The point is that given that we're asking about people who have certain memories, the chances are overwhelmingly large that those memories randomly fluctuated into existence. So don't get hung up on the brains. When it comes to the Boltzmann brain problem, the real problem is for whatever observers you care about, they're most likely to have randomly fluctuated into existence.
1:15:36.4 SC: Chris Kaltwasser says, "I have an unpopular theory, and that is that the three-point shot should be removed from the NBA. It rewards a less exciting game. Who cares how good somebody is at shooting a long shot? I would rather see some cool plays that involve strategy, great passing, getting inside, that kind of thing. I have yet to do the research, but I don't think the three-point shot actually pays off as often as people think. I feel like players miss them more often than they make them, giving up possession without making a shot that might have taken more strategy but had better odds. What are your thoughts on this topic?"
1:16:09.8 SC: Well, I think so there's two different issues going here. I mean, you've definitely asked two different questions, but that's okay, we'll allow it this one time. One is what makes for the more exciting game, having a three-point shot or not. And the other is given that there is a three-point shot, what makes for the better strategy, using it or not. As far as the strategy is concerned, I think it is pretty clearly demonstrated that given the rules of the National Basketball Association as they are currently construed, shooting three-point shots is very good, is absolutely a good strategy. I mean, this was Daryl Morey, former Mindscape guest and former president of basketball operations at the Philadelphia 76ers, pioneered this approach very explicitly. Other people have sort of done versions of it, but he really made it very clear: the two best shots you can take are either a shot right next to the rim, a layup or a dunk, or a three-point shot. As you go further away from the rim, but not all the way to the three-point line, your probability of hitting the shot continuously goes down. And then what happens at the three-point line is that even though the probability of hitting the shot keeps going down, suddenly the number of points you get goes up discontinuously. So you get an extra bonus there.
1:17:25.0 SC: And Chris, I think that your intuition was absolutely shared by a large number of people in the NBA, both sort of at the level of coaching and playing in, let's say, the 1980s and '90s. They just didn't shoot a lot of three-pointers because they thought, well, number one, usually you miss. Number two, there's long rebounds and you don't get the rebound and it's just a less effective shot. But since then, people have run the numbers and that's just not true. The benefit you get from hitting the shots, for one thing, we have better shooters now, right? A good three-point shooter will be hovering around 40% of the shots going in, which is lower than the 50% for a good two-point shooter. But guess what? 0.5 times 2 is less than 0.4 times 3. So you get less points per shot for a mediocre two-point shot than for a good three-point shot. So I think that... Of course there's a lot of extra details there. Do you in fact have a lower percentage of offensive rebounds with a three-point shot than a two-point shot? And I think people have worked that out, but I don't know it myself.
1:18:34.4 SC: As far as the interestingness of the game is concerned, I do think, and in fact Daryl Morey has said, that there are too many three-point shots. I mean, he's interested in winning basketball games, so he encourages his players to shoot them, but he would prefer if the rules were changed so that the three-pointer was less dominant than it is now. I can see the arguments for that. I haven't really thought it through myself. It's completely misleading, I think, to people who are not basketball fans to suggest that there's no strategy involved in shooting three-pointers. Everyone knows, both on the offensive side and the defensive side of the ball, that a good open three-point shot is very valuable, and therefore the defense doesn't give that up to you. They will go to great lengths strategically to prevent people from getting uncontested three-pointers.
1:19:27.5 SC: And these days, the level of detail in tracking what happens on the court and converting it into data is very, very high. So we know what percentage of three-point shots are contested or uncontested or open or not open, high-percentage shots, whatever. It's a little bit more likely to hit the shot if you're in the corner of the court rather than at the near mid-court or whatever. They know all of those things. So there's a huge amount of effort that goes into scheming so that your good three-point shooter is left alone, able to shoot without a hand in their face. And you may or may not enjoy that aspect of the strategizing, but it's definitely there.
1:20:07.8 SC: Nikola Ivanov says, "If cosmic birefringence is conclusively detected, how could we distinguish observationally between a fundamental violation of Lorentz invariance, as explored in your early work, and birefringence caused by an evolving field in a Lorentz-invariant universe?" So for those of you who have not been following along, cosmic birefringence is this phenomenon you can get whereby a photon traveling through space, photons have polarizations, right? Like if you have polarized sunglasses, it will let in photons that are vibrating up and down, but not right and left. And with birefringence, that direction of polarization rotates as the photon travels through space. Ordinarily, you don't expect that to happen, but various exotic physics effects can cause birefringence in the universe. One of the first was discussed in a paper that I wrote, my first ever published physics paper, with Roman Jackiw and George Field, where we were looking at a theory that violated Lorentz invariance. There was a preferred direction to the universe against which the photons interacted.
1:21:12.6 SC: Now, consistent with what I just said a little while ago, none of us thought that theory was true. In fact, we were kind of horrified by our own theory. We were not saying you should take this seriously as something that's out there. What we're saying is we have this principle called Lorentz invariance: the universe doesn't care how fast you're moving or what direction you're pointing in. But how do you know that physics actually obeys that rule? Well, you know by considering theories that are not Lorentz-invariant and then testing them experimentally. So we put forward a well-defined theory that was not Lorentz-invariant, and we showed how to test it experimentally. But it was just sort of a toy model. And indeed, it became quite a little industry for people to come up with other ways to write down theories that were not Lorentz-invariant and then test them experimentally. So it was a useful paper for that purpose.
1:22:06.7 SC: Now, later I realized that dark energy, there could be, plausibly, a rolling scalar field that fills the universe and makes it accelerate, could also lead to this kind of birefringence effect. And the same experimental results and tests that we talked about would be useful here. And in fact, the spirit of the experimental tests we suggested, but it was Marc Kamionkowski, former Mindscape guest, and his collaborators, who noticed that the cosmic microwave background would be an even better way to test it. So that credit goes to them for that. Now, so this is different, because now you have an independent motivation for caring about this new physics idea. The original idea that we suggested was really just because it was there and you can test it. Now you're explaining something different and predicting on the side as a spinoff that you get this birefringence effect. So I think it's overwhelmingly more likely that if you actually see it, it's caused by an evolving field rather than a Lorentz-violating vector field.
1:23:14.4 SC: But in fact, those two things aren't that different. You have a field that is changing with time, you take its gradient, that is to say the direction and magnitude with which it's varying, and that defines a vector. That's basically exactly the same as the theory that we looked at. It's just that the underlying dynamics are a little bit different. And neither one of those is the only way to explain cosmic birefringence. It could be birefringence that is intrinsic to the early universe, not the late universe. There's lots of different possibilities out there. So it'll be a theoretical and observational playground if any of this turns out to be verified by further observations.
1:23:53.8 SC: Rich Hogg says, "As one with a strong naturalistic bent, I would tend to think that someone else's supernatural experience was a hallucination or other misfiring of his brain. But I want to keep an open mind while not forgetting that it could happen to me. If I see a ghost tonight, how do I, as a maybe crazy Bayesian, decide whether I should update my priors or make an appointment with a neurologist?" Well, you're clearly allowed to do both. Look, you should always update your priors. Any data that you get should constantly be causing you to sort of gradually nudge around all of your priors in many cases, but you have to... Sorry, let's back up a little bit. What are we talking about here? A good Bayesian says the probability of a theory being true given some data is proportional to two things. Number one, the prior probability that you thought that theory might have been true. And then number two, the likelihood that you would get that data if that theory were true.
1:24:53.6 SC: And so the prior probability is there, the likelihood function is also there. And so if you see a ghost, the probability of you seeing a ghost is higher on a theory where ghosts really exist than on a theory where ghosts don't exist. So you should increase your credences that ghosts exist, decrease your credence that they don't. But if your priors are so far apart already, if your prior that ghosts don't exist is 0.999999 and prior that ghosts do exist is 0.000001, that update is not going to make much difference. Okay, so don't only think about the likelihood updating, think about the actual value of the priors also. And if it happens again, then you can start updating again, and eventually, this is what a good Bayesian would do, eventually, if you get so much data that is only explained by the supernatural explanation and not by the natural one, then your credences in that explanation should become appreciable one way or another.
1:25:56.6 SC: Michael Bright says, "Since learning even just a little bit about Einstein's explanation of gravity, I've always been bothered by the cartoon image of a ball on a trampoline. It makes it look like space has a down. If you agree with my frustration, is there a more preferable cartoonish image or set of images in your mind to visualize what spacetime curvature actually is?" You know, I don't mind the usual cannonball or bowling ball on a trampoline or a rubber sheet or whatever. It's not supposed to be super-duper accurate. All that that is supposed to convey to you is the idea that space can be curved.
1:26:30.1 SC: So there are many ways in which that particular metaphor or analogy are not accurate. Like you say, the most obvious one is there's an up and a down. There's gravity pulling on the bowling ball to stretch the rubber sheet, and that doesn't happen in the real world. So can you do better? You can try. You can try visualizing lines in three dimensions, three-dimensional space, that are distorted or curved by the presence of a massive object. But that's not very accurate either. If you think about... Once you understand the equations, you realize that we say words like spacetime is a unification of space and time, and it's really spacetime that is curved. In fact, the reason why the Earth goes around the sun has a lot more to do with the curvature of time than the curvature of space. Curvature of time is actually not a sensible concept, but the way in which the time coordinate changes from close to the sun or some other gravitating body to far away has more to do with the path that the Earth travels on.
1:27:42.6 SC: So no visualization of the curvature of space all by itself is going to be able to do the job. That's okay. Very often I get asked the more general question, like how do you visualize this or that particular difficult concept? And different people have different visual things in their head. They have lines or curves or shapes or whatever. They have three-dimensional vector spaces, and those stand in for 10 to the 10 to the 10-dimensional vector spaces. But ultimately, the visualizations are not the point. Understanding the concepts and the equations behind them is the point. That's why I wrote Space, Time, and Motion, so that if you care about understanding the equations, you can get those. Sadly, the world is more rich and large and multifaceted than our visualization capacities are really up to.
1:28:35.1 SC: Zach McKinney says, "The prevailing cultural narrative around AI is one of inevitability and runaway proliferation. However, this chain of reasoning never seems to reckon with the planetary boundaries of finite material resources and energy. On what timescale, if any, do you foresee the proliferation and diffusion of AI to become constrained by such limits?" Well, look, I'm not sure that is the prevailing cultural narrative. I think that there may not be a prevailing cultural narrative. Different little subcultures have their own narratives, and it depends on which one you're embedded in. But also, again, I'm not sure I mind this runaway proliferation kind of talk as a possible future. I'm not saying that I think it's the likely future, but that talk has to be understood in a certain way. When people talk about exponential growth of anything, they're never being perfectly accurate. Unless you're literally talking about the universe itself, nothing grows exponentially with an unbounded time horizon because there are bounds, there are finite numbers of resources.
1:29:37.5 SC: So if you have the famous S-shaped curve, the sigmoid curve, that is flat and then it starts going up and then it turns over and it's flat again. So that's an approach from one kind of steady-state equilibrium phase to a different steady-state equilibrium phase. In the beginning, at the very start of that transition from the first state to the curve, it looks exponential, right? Every exponential in physical reality is probably going to really be some kind of sigmoid curve or something even more complicated than that. It's just a statement about the initial steps. So extrapolating these ideas to the wild future is not an accurate thing to do, and everyone should know that. Now, having said that, what are the limits? What are the boundaries? What are the limits that we have on resources and energy? That's hard to say. We can always just do some back-of-the-envelope calculation of how much energy we have on Earth, but maybe you're not on Earth. Maybe you're collecting energy elsewhere. So I think that there are factors and considerations that will change the future of AI before, long before we start worrying about the total amount of energy we have available here on Earth.
1:30:53.8 SC: Michael Aster says, "Since you've talked about how you like the game, how would you feel if Kingdom Rush reached out for a sponsorship on the podcast?" I'd be thrilled if they did that. I would love it. It's always better. I don't ever advertise things that I think are bad, but often I'm advertising things that I haven't previously experienced and I have to sort of do homework on them. So whenever I get the chance to advertise something that I'm super familiar with and love, then I'm always happy to do that. So yeah, if Kingdom Rush wants me to advertise things, if any watch companies want me to advertise things, if the Philadelphia 76ers want me to advertise season ticket plans, I am more than happy to do that.
1:31:36.5 SC: David Sotalongo says, "In the April 2023 AMA, you said, 'I think we need to build a lot more housing. I think it is very, very obvious that's the best way to make housing more affordable, to build more of it.' I agree, and I'm curious how you think we should go about building more housing. Should the government step in to build and maybe manage such housing? And what, if anything, would you do about existing homeowners who oppose new housing because it would decrease the value of their property?"
1:32:01.0 SC: Well, so number one, I have no informed opinion about this. Whatever I say in the next few minutes, don't put a lot of credence on them. I'm not an expert on housing policy, okay? That's not just one of the things that I know a lot about. I do have experience with the kind of limitations, the constraints that are put on homeowners to do different things. When I lived in LA, which we did for a long time, we lived in a townhouse. So it was a group of six homes, and we were in this unusual situation for Los Angeles where we really had almost no outdoor space, right? No yard or even tiny porch or anything like that. And so obviously that was a thing that would have been nice to have. And we had the idea, well, maybe we could save up and put a deck on the roof, a roof deck for the townhouse. Because a flat roof, again LA, you just have a flat roof. You don't need to let the snow fall off your gabled roofs or anything like that.
1:32:59.0 SC: So we looked into it and it was just, we put money into not getting permits, but figuring out how we would be able to get the permits to do this. And eventually we gave up before we even got the permits because it was just clearly going to be too much work. So one thing we can do, that's not even building a house, that's just putting a roof deck on the house that you own. One thing we can do is just go through and make it easier to build housing responsibly. I don't mean remove all regulations and let people build housing that is dangerous or not up to code or whatever. Just grease the wheels, make it easier to do good things for society. And I think that having more housing is a good thing. Make it easier to take existing housing and renovate it and sell it. Just make it easier, because that's not obvious. That's not something that we put a lot of effort into, taking existing sets of regulations and making them just as effective but easier to navigate, right? We're not very good at that as a society or a government.
1:34:01.3 SC: Governments building housing, that's fine. I don't think it's ever the real solution to anything. I think that we want the housing market to be thriving in terms of both people building houses and people buying houses and people selling houses. A house is very often the most expensive transaction that a person ever gets involved with. Of course, many people just rent or whatever, but if you buy a house, that's a huge amount of money. And I don't think it should be that much of a huge amount of money. I think they should be less expensive. Now, like you say, there's a constituency that's against that because they don't want their current home value to go down. So number one, I don't think it's gonna go down that much no matter what we do. Number two, too bad for them. But number three, and most importantly, it's a house. It's not an investment, right? I mean, it's an investment in some technical sense, but the reason why you have it is not primarily to be an investment, it's primarily to live in it.
1:35:06.0 SC: And what happens is if you have a house and its value goes down because the entire housing market is going down, then guess what? If you sell it to buy a new house, that new house's value has also gone down. So I just don't think it's that big of a problem. I mean, unless you're someone who owns multiple houses because you're trying to get investment income from it or something like that, and then, sorry, investments are tricky and unpredictable and that's life. We have to learn how to deal with it.
1:35:38.6 SC: Alan Lubell says, "Your intelligence and generosity of spirit has really made an impact on me. I'm asking if you feel you received these qualities from your parents or any other influences when growing up, and if you could give a couple of examples." That's very sweet. Thank you, Alan. I appreciate the compliments. I usually don't try to read out compliments here. Sometimes they precede a question, but this one was kind of relevant to the question. And I'm not even gonna give a very good answer to the question because on the one hand, yes, I absolutely was influenced by my parents, mostly my mom. My parents got divorced very early, so it was my mom who raised my brother and me. But also other people: friends, teachers, other relatives, other grown-ups who I got to know, and also books that I read and things like that. So I kind of suspect, maybe this is not true for other people, I don't know, but for me, if I tried to associate some quality I have now with the example or influence of some particular person in the past, I probably wouldn't be doing a very accurate job.
1:36:40.9 SC: Like you could imagine that you'd put your finger on it, but I tend to think that, in fact, people are just sort of gradually influenced by many, many, many, many, many things over the years, and it's very hard to pinpoint when there was some crucial influence that made a difference. I don't think it's a matter of one person telling you one little bit of wisdom that changes your life. I think it's a matter of being in the presence of people who live well and try to be good people and implicitly picking up from the atmosphere around you that that's a good way to be. So I still try to learn from the people around me. I try to hang out with people who I admire. There's some point in your life in which you realize, no, it's actually not worth hanging out with these people who are talented or useful in one way but really bad human beings. I would rather hang out with people who are good human beings and maybe that will rub off on me.
1:37:35.0 SC: Miron Mizrahi says, "In the last AMA, when you discussed neutrino oscillations, you said something that sounded a lot like the uncertainty principle. There is no state when the neutrino has both a definitive mass and definitive flavor. I get how this can work for a single particle or field, but how does it work for three? Or is there one field for all neutrinos and the mass-flavor combinations are just different values of the field?"
1:37:58.7 SC: Yeah, this is a good question. It's a little bit subtle. I'm not sure whether the answer I give is going to be clarifying or not, but I can try. If you just had a single particle, then you wouldn't have uncertainty about mass and flavor. There's only one flavor and only one mass, right? So it's not a mixture. I think the way I think about it, I'll be honest, I haven't really thought about it in terms of an uncertainty principle, and therefore I think it's not the uncertainty principle. I think there's a different thing going on because I think it would be true even classically is the point. The uncertainty principle is truly a quantum mechanical statement. It's a relationship between observables, like let's say position and momentum, and quantum states. In particular, the usual, the standard, famous, celebrated uncertainty principle is in any quantum state of a single particle, you cannot be highly localized both in position and momentum at the same time.
1:39:03.4 SC: If you define the standard deviation or the variance as delta X and delta p, then delta X times delta p has to be at least h-bar over 2. Okay. So that's a completely quantum mechanical statement. There's no classical analog to that. In the classical particle, I can and often imagine doing, telling you both the position and the momentum precisely. But for quantum mechanics, the notions of position and momentum aren't precisely defined things. They're observables that have different probability distributions of getting different answers. That's not really what's going on with the neutrinos. With the neutrinos, what matters is that they are fields, okay? That they're fundamentally neutrino fields. And even before you quantize them, you can talk about the classical fields. And fields are things that have different values in different points in space.
1:39:53.5 SC: And there are, roughly speaking, three neutrino fields, okay? I say roughly speaking because you can get into, well, do antineutrinos count differently and do different spins count differently? Forget about all that, okay? There's, for our purposes today, three different neutrino fields. And you can rotate them, so these three different neutrino fields form a vector space, which is very similar to quantum mechanics in some ways, but that's just because vector spaces are everywhere. It's not because there's any true intrinsic deep relationship there. So you can think of electron neutrino, muon neutrino, and tau neutrino as basically three axes in the three-dimensional space of the neutrino field. And any one neutrino is a vector, is represented by a vector in that three-dimensional vector space. So that means it has a component that is the electron neutrino component, a muon neutrino component, and a tau neutrino component.
1:40:50.1 SC: And then separately, there's a different set of axes, a different set of basis vectors for the three mass states of the neutrino, which are cleverly named m1, m2, and m3, okay? And the point is that the mass axes are not exactly aligned with the flavor axes. So if you are a state of definite mass, then you're not gonna be a state of definite flavor for exactly the same reasons as if I take a two-dimensional vector space with x and y coordinates and I rotate those coordinates to x-prime and y-prime, a value that is a certain value of x-prime and then zero for y-prime is gonna be a combination of the original x and the original y. So it's really just rotating vectors in field space, nothing intrinsically to do with quantum mechanics or the uncertainty principle.
1:41:45.7 SC: Dylan Samuel says, "I'm a teaching assistant for a mechanical engineering class, and I'm pretty sure that a very large percentage of the class is using AI to do their homework, as their test scores tend to be vastly worse than their homework scores. Recently, a UC Berkeley professor made news saying they no longer see the point in grading their students' homework. Apologies if you've answered something similar before, but what are your thoughts on this? And generally, how are you adapting to teach with students knowing they'll likely use AI? It seems counterproductive to pay for college and then not learn anything."
1:42:15.9 SC: Yeah, this is something where professors are talking about this a lot right now. It's a sea change in how things get done. There were always ways to cheat on homework sets, right? You could work together in ways that the professor told you not to. You could search the internet for maybe this problem had been done before and answered. Now it's just much easier with AI to find the solutions. And so the way that I think about it is, for professors, it was always the case that their goal was to make the students learn things. And for the students, some of them had the goal of learning things, but we gotta be honest, some of them just had the goal of getting good grades because they wanted to get a job or into grad school or medical school or whatever.
1:43:05.0 SC: And the fortunate thing for the professors was that the goal of teaching them things and the goal of them getting good grades happened to line up. The best way for the students to get good grades was to learn things and therefore do well on the homeworks and the tests. So that system was sort of sensible. Nowadays we've broken that connection. If all you do is give out homeworks or take-home exams, which, by the way, before I got to Johns Hopkins, I'd never given an in-class exam before in my life. And it's not Johns Hopkins' fault, it's just I got here around the time that ChatGPT arrived. So I would give out take-home exams, and of course the students could in principle these days do those by using AI. And so you could in principle get very good grades on all those exams without actually learning anything. So therefore you have to change the way that you test people. I still give out problem sets, but the problem sets aren't worth that much of your grade. They're basically very explicitly given out in order for you to practice so that when you take the test, which is in-class, you are prepared for that. And obviously no computers in class or anything like that for the exams.
1:44:23.0 SC: And if the students... I tell the students, look, if you want to use AI to, or for that matter, if you want to collaborate and have friends help you on the problem sets, go ahead. It's your choice, because it's only 20% of your total grade. 80% of the grade is how you do on the exams. And if you're depriving yourself of the practice that you get by actually doing the problem sets yourself, then you're gonna do worse on the exams and you're not gonna get a good grade. So I'm not sure if that's feasible for everybody, depending on the class, depending on what you're doing, but it's the world. You certainly cannot rely on the fact that just because a student is in college, they're gonna do the best thing possible to try to learn things. There are reasons to go to college other than learning things. For many of us, learning things was the point, not for everybody. And you have to accept that and adapt accordingly.
1:45:16.7 SC: Tucker Hyatt says, "Am I mistaken to interpret Newton's second law to include the notion that forces cause accelerations, and to want something similarly causal in a quantum field theory worldview? Please don't make me give up on causality." So you should give up on causality, Tucker. Sorry about that. At the level of fundamental physics, causality is not there. In fact, this was the subject of a famous article and quote by Bertrand Russell, of all people, over 100 years ago. He was pointing out that causality used to be something that philosophers hung on to with great fondness, but that's just not how the world works at the level of fundamental physics. And he was talking long before quantum mechanics or quantum field theory. So some people say that quantum mechanics and the Born rule and wave function collapse undermine causality, but no, it was already not there.
1:46:10.0 SC: But that doesn't mean that all hell is breaking loose. It doesn't mean anything can happen. There are still laws of physics. It's just that the laws of physics do not take a kind of Aristotelian form. In an Aristotelian view, you might very well be tempted to say there's an acceleration and there's a force, and the force is the cause of the acceleration. But that's not what Newtonian mechanics says. Newtonian mechanics says the amount of force at any one time is equal to the mass of the object times the acceleration. Period. It doesn't say that one causes the other or anything like that. It doesn't say that force divided by acceleration causes mass. Okay? The word "cause" just doesn't appear. Sometimes in a book on relativity or quantum field theory, you will find a discussion of causality, but it's really not about causality in the usual Aristotelian sense. It's just a discussion of how signals travel inside light cones and not outside light cones. So they've repurposed the word "causality" to mean things move slower or equal to the speed of light.
1:47:17.1 SC: And all that's fine. Of course, notions that were introduced 2,500 years ago aren't necessarily gonna be the right way to talk about the world once you learn more about it. And also, of course, the notion of cause and effect is super useful and important in the macroscopic emergent world. There's a whole chapter in the new upcoming book on complexity and emergence, Volume 3 of The Biggest Ideas, where we talk about causation and we talk about Judea Pearl, former Mindscape guest, and his notion of the do-operator and its implementation in what are called Bayesian networks. So, of course there are causes and effects in the world, just like there's consciousness and free will and tables and chairs and temperatures and pressures. Okay? None of those exist at the fundamental level, but they're very real at the emergent level, and that's fine.
1:48:10.7 SC: Matt Patterson says, "I'm an avid Mindscape follower, and I was intrigued by your discussion with Jenann Ismael on her and her philosophical argument about why it's okay to kill cows but not humans based on their value. Value in this case being measured by one's memory of one's life, having plans for one's future, as well as these same items in others that one has influenced, et cetera. For cows, this kind of value is essentially zero, and for humans it is much greater. It occurred to me that the value of individual humans can vary over a wide range and might even be negative, e.g., in the case of a serial killer. Given that thoughtful humans do decide to kill other humans, it seems that considering the value of those that would die might be one of those considerations in evaluating the moral rightness or wrongness in making such a decision. My question is, do you think that the value of potential death victims should be one of the considerations when assessing the morality of decisions to kill them, or is there some moral or other objection to this idea?"
1:49:08.1 SC: Well, yeah, I think that there's a lot going on here, and we're not gonna give the definitive answer right now. I'm not even claiming to know what the definitive answer should be. Of course, in practice, when there are situations, hopefully it never arises for you, it has never arisen for me, but when there are situations in which you are deciding whether to kill somebody or not, there are considerations about what kind of person that is. Like, is this a person who just broke into your house and is threatening to kill your family? Then you might have some considerations. Is this a person who is waiting too long to go after the light turned green in front of you on the road? That's a different set of considerations, right?
1:49:50.5 SC: So maybe your considerations are no one ever should be killed, there should be an absolute right to continued existence. And in terms of the criminal justice system, that would be my view. I do not believe in the death penalty. I don't think that people should be put to death. Mostly for me, that's a matter of practice rather than principle. I don't think that there is some metaphysical absolute right to be living. I think that it is extremely difficult to effectively figure out who doesn't deserve to be living and to carry that out. And certainly the actual government and the actual criminal justice system, I don't think is very good at that. And it's an irreversible decision, so I don't want them making it at all. So I just don't want there to be the death penalty.
1:50:35.7 SC: There are wars, and you might be against war, but maybe the other side is not against war and they invade you. What are you gonna do? You just gonna lie down, or are you gonna fight back? I'm in favor of fighting back if you get invaded. I think that we have far more wars than we should, but I'm not gonna pretend that I'm gonna have a principled moral point of view that says, "I would just never participate in a war no matter what the circumstances were." And when you do participate in a war, there's absolutely considerations of you should try not to hurt civilians. There are other people who are enemy combatants, and that's a different story. And nevertheless, they also have some rights. People have thought about this a great deal. So I do think that all of these are complicated. And more importantly, it's okay to say that it's complicated. You don't need a simple once-and-for-all rule. You can understand that these are gray areas and try to do your best.
1:51:33.2 SC: Karoli Kanter says, "How likely do you think it is that the observable universe is such a minuscule part of the total universe that our observations are just local phenomena and say little of use about the whole?"
1:51:46.8 SC: Well, I truly don't know whether our observable universe is a tiny fraction of the whole universe or is most or all of it. I'm very plausibly... I'm open to scenarios of either kind. This is, again, something I think we should be humble about. We should explore the different possibilities. I am skeptical of the worry that local phenomena allow us to say little of use about the whole. Even if the universe we see is only a tiny fraction of the whole, it's still pretty big, right? Hundreds of billions of galaxies, many millions, billions of light-years across, et cetera. And there is absolutely remarkably uniform behavior of the physical universe over this vast stretch of space and time. We can test that very quantitatively, right? We can use things like primordial nucleosynthesis and the current abundances of light elements in intergalactic space to say that things like the mass of the proton and the neutron and the charge of the electron were the same three minutes after the Big Bang as they are now to within a very tiny error bar. Likewise, those quantities and the laws of physics underlying them seem to be exactly the same in quasars billions of light-years away as they are now, et cetera., et cetera.
1:53:09.4 SC: So I think that given the uniformity of the laws that we observe in the universe that we do observe, it's perfectly reasonable to imagine that some kind of uniformity extends across the whole universe, no matter how big it is. That uniformity might not be simple. In other words, it might not just be, oh, the mass of the electron and the proton and neutron are exactly the same everywhere. Maybe that's true, maybe it's not. We have scenarios in which either one of those is correct. But even if the actual values of the parameters you measure differ from place to place, they can still be reflections of uniform underlying laws. It's just like water can be liquid or vapor or solid, right? It's just a different manifestation of the same underlying stuff.
1:53:56.5 SC: That might not be right. So it's conceivable that outside the universe we observe, there are laws or there are conditions that bear precisely no relationship to what we observe in the universe around us. But I would say, number one, there's no reason to think that. There's no empirical reason, there's no puzzle that we solve by imagining that's true. And number two, this uniformity that we do have observed gives us some credence that some kind of uniformity extends even further. So as to how big the universe is, I have no idea. As to how much we learn about it by studying the little bit that we see, I think we can learn potentially quite a bit.
1:54:37.9 SC: In fact, let me add to that a little bit. There's a very real sense in which everything that we know about the laws of physics has been learned by human beings here on Earth over the past few thousand years. So in both space and time, the extent that human beings have actually prodded the natural world is really, really, really tiny. So the fact that we can make accurate predictions about things far away in both space and time is remarkable and I think is okay to lead us to conclude that there's a good chance that there is something learnable, intelligible about the world.
1:55:17.8 SC: Kelvin Firth says, "In mad-dog Everettian quantum mechanics, where there's a single wave function of the universe that evolves with time according to some Hamiltonian, what exactly is that Hamiltonian or where does it come from? In what I've learned so far as a beginning physics grad student, the Hamiltonian comes from some potential which is usually classical looking, or at least it's a function of space. However, in the Mad-Dog Everettian picture, I take it that space is emergent. So where does the Hamiltonian come from? Where does it live?"
1:55:44.1 SC: Well, the idea is the Hamiltonian is the laws of physics. It is defining the universe. It doesn't come from anywhere else, right? This is always true in anyone's version of the complete theory of physics. There has to be a starting point. There has to be, "Okay, here's the theory," right? In Mad-Dog Everettianism, the theory is, here's the Hamiltonian, and in fact, the Hamiltonian is simply specified by a list of energy eigenvalues, the spectrum of the Hamiltonian. Why is it the way it is? I don't know. So it's completely possible to me that it just is that way. It's also possible to me that just like we only see some branches of the wave function, we only see some parts of the Hamiltonian, and maybe the parts that we see are chosen by some anthropic reasoning, right? That this is the part of the Hamiltonian where human beings can exist. I don't know, but I would deny that I need to know. At some point, asking questions like this leads you to say, "That's just the way it is."
1:56:47.0 SC: Felivel Camilo says, "Do you believe technologies such as large language models or realistic image generation arrived at an inappropriate moment? Might humanity have benefited more from developing them only after achieving some form of enlightenment?"
1:57:01.8 SC: Well, no, I don't really think that humanity is on a trajectory to enlightenment. Maybe individual humans would achieve enlightenment more frequently if they lived for a few tens of thousands of years. But we keep making new human beings, and overall, I don't see much evidence that as a species, we are achieving some form of enlightenment. In many ways, as a species, we're getting better. I'm not trying to be too curmudgeonly about it. I do think that we treat each other better now, on average, than we did a few thousand years ago, but we still treat each other very, very, very badly in some ways. Both of those statements can be true at the same time. And I think that if you're waiting for us to achieve enlightenment before developing very advanced technologies, you're gonna be disappointed. And so I don't even think that's worth waiting for. Maybe these technologies will help us achieve enlightenment or anyway get better at it than we are now.
1:58:00.3 SC: Redlinks says, "What is your position about AI successionism? The idea that if machines will become conscious, it would be ethically right to let them replace humanity?" Well, I think that idea is extremely silly. I'm sorry. I do believe that AIs might become conscious. We've talked about it on the podcast. We've talked with people like Ned Block and Anil Seth who are skeptical that AIs could ever achieve consciousness, but other people like Stuart Russell, who I think believe otherwise. But why in the world would you say even if. So that's a good question all by itself. What does it mean for AIs to be conscious? Can they ever get there? What should we do if they do get there? Why in the world would you then say they should replace humans? I don't even see what the motivation for that is. I suspect, I'm not very familiar with this line of thought, so maybe I'm missing something deep, I suspect it's the usual overly anthropomorphic oversimplification of humanness or consciousness or intelligence that is endemic in certain communities that talk about AI.
1:59:10.2 SC: They act as if there's just a scale, a linear scale of intelligence or value or whatever, and we have a chance to be on top of it for the moment, but then AI is gonna pass us and then we're gonna be like puppies and kittens to the AI. I just don't think that's how it is, right? I think that whatever value human beings have morally or intrinsically or whatever, they will have just that much value even if the AIs become conscious. So it's okay to have two different species that are both conscious and can live together. I don't see any problem with that. So no one has given me any even vaguely persuasive argument in favor of something like AI successionism.
1:59:52.1 SC: Edwin Miles says, "I've been wondering recently about values and wanted to ask the following. Is the value of something only and always contingent on there being something or someone to value it? Or is there some intrinsic value to something separate to any human bias, preference, or valuation, et cetera.?" Well, this is the kind of thing the philosophers talk about. There are value theorists. That's literally a job description. Okay, I'm not a value theorist, so I'm not here to give you the final answer. I don't want to say that values are necessarily reliant on there being someone or something to value it. I would say that values are parts of relations. They're not intrinsic to the thing. So if I have a cube of material, 1 kg of something, what is its value? Well, that depends on not only what material you have, but what other things there are in the universe that might make use of that material or might care about it. I'm not someone who thinks that value is always necessarily pragmatic. I think a work of art can be very valuable or even a little artifact that is just of sentimental value to you. Those can be valuable to you, and the value is different to you than to somebody else. I think that's fine.
2:01:12.7 SC: But it depends on the relationship to something, usually a person or something that can actually assign a value. I think the bigger lesson here or the bigger picture thing is we use words like value in our everyday life, and we know what we mean. And it's very, very common, once you get to a more careful consideration of the meanings of the words, like we do in philosophy, to realize that what we thought was a perfectly sensible concept in our everyday life doesn't serve that purpose or have that status in a more careful way of thinking about it, like free will, for example. And in those cases, I generally would like to say, in the context you're talking, think about what it is you mean. And if there's some word that you're trying to define and that definition is actually getting in the way of thinking clearly about what it is you're trying to do, then pick another word or try to talk without using that word at all.
2:02:08.6 SC: Scott says, "Do Feynman diagrams have ontic content or are they purely computational? I understand that all we measure are scattering amplitudes, not full histories of particle interactions that Feynman diagrams might visually imply. But it seems that among working physicists, these diagrams are essential vocabulary for discussing and visualizing processes. It makes me think that there is more to them."
2:02:32.9 SC: Well, I think this is a good question because I think the truth is in between a little bit. Certainly you shouldn't take Feynman diagrams exactly literally. And I think that's already obvious from the definition of what a Feynman diagram is. The point of a Feynman diagram is that every diagram represents one term in a sum. You add up the contributions from many, many Feynman diagrams, typically an infinite number, to get the answer to a question like, what is the probability these two particles will hit and annihilate each other or scatter in some other way? So the process, whatever it is, has some particles coming in and some particles going out, and an infinite number of Feynman diagrams that you use to calculate the likelihood of that process happening. So how in the world could we attach individual ontological content to the individual Feynman diagrams that are in the sum? It's like saying if I have the number 10, I could write that as 5+5 or 3+7. There's no one right way to decompose it in that way.
2:03:40.9 SC: And indeed, at a very down-to-earth technical way, that can be the case in Feynman diagrams. There can be different ways of calculating the same process and getting the same answer that have different sets of Feynman diagrams appearing. Now, having said that, you're also right that particle physicists talk as if they're real all the time. And that's because particle physicists are finite, calculationally bounded, rational beings, and it's helpful to think of those processes represented by the diagrams as really happening. Why is it helpful? Well, it helps you see what processes can happen and can't happen. And by doing a little bit more analysis of the individual diagrams, you can even get a rough idea for how likely something is. And sometimes you can be surprised; there can be secret cancellations. That was a big deal in the '70s when all this was being worked out. So it's helpful, but it's not the final answer. The final answer is the actual process that is going on.
2:04:38.0 SC: And of course, really the process is a process involving fields, not particles at all. So I think it's okay. Just like in quantum field theory at all, in Quanta and Fields, I try to make the point it's okay to talk about particles. Particles are not fundamental. Fields are fundamental, and particles are vibrations in them. But that's not to say you can't talk about particles. You can talk about particles if it gives you insight and helps you understand what's gonna happen. I think Feynman diagrams are exactly the same way.
2:05:11.6 SC: Okay, I'm gonna group two questions together. One's from David Guderian, who says, "Could you please talk about pointer states? Which aspects of them are already explained and understood by the physics community, and which aspects still await an explanation? You once mentioned that finding the classical world is still an open question in the Many-Worlds Interpretation." And then Henry Jacobs says, "On the question of why do we never observe a superposition, is it plausible that observing always involves large systems becoming entangled with the small systems of quantum spin or whatever, and we don't consider the system observed until it has evolved into one of the eigenstates? This seems circular, but I don't know if that's a bad thing."
2:05:49.1 SC: So yeah, for Henry's question, we're gonna answer it in part by answering David's question. So the idea here of a pointer state is a state of a quantum system that basically can resist becoming additionally entangled with the world. So the classic example here is Schrödinger's cat, okay? Imagine you take the cat, you put it into a superposition of awake and asleep. The important thing is not whether the cat's asleep. The important thing is that it is a macroscopically distinguishable superposition. In particular, we're imagining here that the awake cat is up and walking around, and the asleep cat is lying down on the floor. And what happens is the photons in the box or the air molecules in the box, et cetera., bump into these two parts of the cat state differently. So they might bump into the awake cat but miss the asleep cat, or vice versa. And what that means is that the quantum state of the air molecules and the photons, et cetera., becomes entangled with the quantum state of the cat. So you have an entangled superposition of the cat awake and asleep, plus the environment, as we call it. And basically, we can call it the part of the environment that observed the cat to be awake and the part of the environment that observed the cat to be asleep.
2:07:09.0 SC: Now, nothing observed anything there, right? That is not a human conscious observer that we're talking about, but that's essentially what we mean from the Many-Worlds point of view about when a measurement occurs: when the quantum system has become entangled with its environment. But then the extra thing that is important here is that once you focus in on a single branch of the wavefunction, a branch where the cat is awake, for example, that cat still interacts with the air in the box and the photons in the box, et cetera., and the floor and whatever, but it doesn't additionally become entangled because every air molecule, let's just say in the box, interacts with the cat in the same way. When the cat was in a superposition of awake and asleep, the air molecule would interact differently with the cat depending on which part of the wavefunction we talked about. But once you're in just the awake part, every molecule either does or does not interact with the cat in exactly the same way. That's a pointer state.
2:08:13.4 SC: A pointer state is one where you can interact with the outside world without becoming additionally entangled. Pointer states are robust to being measured, to being monitored, as we say, by the environment that they're in. How well do we understand pointer states? We understand them pretty well. When I talk about the need for a better understanding of finding the classical world in Many-Worlds, I'm sort of assuming that... Well, I'm taking a step back and imagining we don't even know the laws of physics. Okay? What are... We know the laws of physics are quantum mechanical, but we might not know the specific Hamiltonian or forces or fields of the world or whatever. And I'm asking the general question: under what circumstances is there something that you would recognize as a classical limit of that situation?
2:09:04.6 SC: But if you have a world where... If you have a situation, a setup, a theoretical playground where you know the laws of physics, which is actually much more common than what I think about, so you already have the situation, you have a cat, you have a box, you have photons, you know how they interact with each other, then people like Zurek and even going back further than that, Joos and Zeh and others and Gell-Mann and Hartle, they all worked out the basic story of decoherence and how it works and therefore what states are the pointer states. So I think that's mostly understood. There's absolutely still work to be done in dotting the i's, crossing the t's, et cetera., but it's pretty much worked out.
2:09:46.1 SC: Now, to Henry's question: why do we never observe a superposition? It's basically because once you're in that pointer state, you can interact with the other molecules in your body, et cetera., but if you were in a superposition that is entangled with its environment, then because of decoherence, basically, the part of the superposition that is in one branch doesn't ever interact with the part of the superposition that is in the other branch. So this is Everett's genius move. He says you are not the superposition of two different parts of your body, two different parts of the wavefunction of your body; you are just one of them. And when the wavefunction branches via decoherence, there's a pointer state that has you thinking one thing, a different pointer state that has you thinking another thing, and you should identify each of them as separate observers. So you're never a superposition, you never observe a superposition, because both you and the thing you observe are pointer states on single branches.
2:10:51.3 SC: And it's not circular, actually. There is some philosophy involved in finding yourself in the wavefunction, figuring out which part of the formalism describes you as a human being, but we can do that. That's okay. We have the capability of doing that, and we figure it out. In this way of thinking about things, what you mean by a measurement is decoherence. What you mean by a measurement is when the quantum mechanical system becomes entangled with its environment.
2:11:21.6 SC: Okay. Daniel Zaccariello says, "Baseball has no clock. Outs are its time, and its rulebook forces an exact conservation law. Every batter ends the inning as an out, a run, or a man left on base. Physics derives its conservation laws from the form of its rules, too, except it has to guess the rules. Is that the only difference?"
2:11:43.9 SC: So I'd like the attempt to make an analogy here between baseball and the laws of physics in the realm of conservation laws, but I'm not sure I see how it works. I would not say that baseball has any conservation laws. A conservation law is to say that there is some quantity that you can calculate from the configuration of a system, like its sets of positions and velocities or whatever, which remains constant, unchanging over time. So if there was some thing that remained constant over time in a baseball game, that would be a conservation law. But the fact that every batter ends the inning on one of three possible, definitely known configurations, that's not a conservation law, because the inning does end. And if the batter is, let's say, on base, then the inning ends and the batter is no longer on base. There's no quantity that you can really point to that is the same from moment to moment.
2:12:42.7 SC: But the underlying ambition, I think, is completely correct. When you have some rules, whether it's physics or baseball, you can often, or at least you can try to find things that are invariant or conserved or always the same for everybody, no matter what specific implementation, no matter which game or physics experiment it is that you're looking at.
2:13:05.3 SC: So Anonymous now says, "If you were starting over today, what assumption in our current formulation of quantum theory would you be most inclined to question? Not because it's demonstrably wrong, but because removing it might open a genuinely new direction."
2:13:20.3 SC: Well, I wouldn't, to be perfectly honest. If you absolutely want to know what I truly think, this is something that is an open question. Is quantum theory itself something we should consider modifying for whatever reason? I have friends, my colleagues at Johns Hopkins, David Kaplan and Surjeet Rajendran, are actively investigating modifications of quantum mechanics. And their motivation for doing that is they think by modifying quantum mechanics a little bit, they can help explain some important weird things about known physics, like the hierarchy problem or whatever. There are additional motivations for modifying quantum mechanics. Nadiia Diachenko, my student, who you know from the quantum cyclic universe paper, has been looking into the emergence of time. And there are certain setups, not universal, but certain setups in which when time emerges, the equation governing time evolution is not precisely the Schrödinger equation. It can be modified a little bit.
2:14:24.0 SC: So there are some motivations for doing that. But look, life is finite. You have to pick and choose what you do with your time. And for me, my impression is that physics, physicists, the field of physics has not done nearly enough to explore the implications and the workings of regular old quantum mechanics. So that's why I'm doing things like how does time emerge, how does space emerge, things like that, from good old ordinary quantum mechanics, because I think that we might get some surprises there that also bear on interesting, important questions that we already recognize are out there.
2:15:01.7 SC: So quantum theory itself doesn't have that many assumptions. Now, you might be asking about formulations of quantum field theory and things like that, specific models that we have of the universe within the broad umbrella framework of quantum mechanics. Now, those are very, very questionable. There's all sorts of assumptions that we can try to think of and imagine changing. The most obvious ones, again, in the framework of quantum field theory itself, are the assumptions of locality and Lorentz invariance, the idea that fields take definite values, classically, they take definite values at different positions of space and moments of time, and then you quantize that theory to get the theory you're looking for. Indeed, I would say there, there's plenty of motivation for undoing those cherished assumptions from quantum gravity. Quantum gravity doesn't really work that way. So locality in particular is something that I would absolutely start questioning and have and am thinking about right now. So we'll see whether or not that actually helps anybody with anything.
2:16:09.0 SC: Daniel Shulman says, "It seems to be accepted dogma that singularities exist inside black holes and sometime just before the Big Bang. If we treat spacetime as having torsion instead of zero torsion, we get Einstein-Cartan gravity, which looks exactly like GR except at the small scales where singularities matter, and the repulsive forces from the torsion in Einstein-Cartan gravity eliminates those singularities. Is it just a historical accident that Einstein didn't know enough geometry to introduce torsion, that we still believe in singularities? Or is there a reason why physicists don't give more credence to Einstein-Cartan gravity?"
2:16:45.0 SC: Yeah, so for those of you who are not experts in general relativity, there's an alternative to, alternative formulation, let's call it that, of gravity called Einstein-Cartan gravity rather than ordinary general relativity. And the thing is, again, if you read, I do have a book on this, if you read Space, Time, and Motion, this explains to you that there is a metric in general relativity, and that's the thing on which everything else is based. The metric tells you the distances between points along lines and things like that. There's also a notion of what is called parallel transport. So I can take a vector, remember we were talking about the polarization of light before being rotated by birefringence? Well, there's some way in any given geometry of spacetime to take a vector, like the polarization vector of a photon, and move it through spacetime while keeping it parallel. You need to define that somehow. This is carried out by something called a connection. And in general relativity, you derive the connection from the metric, because you derive everything from the metric. And this all goes back to Riemann in the mid-19th century. He's the one who suggested doing this.
2:17:56.5 SC: It was people like Levi-Civita and Christoffel who actually worked it out, the connection between, the connection, as it were, between the metric and the connection. Cartan comes along. He's another famous mathematician, and he emphasizes that in principle, you can have a connection by itself, or a connection in addition to the metric. You don't need to derive the connection from the metric. You could treat those as two separate things. And this naturally leads for the connection to have new degrees of freedom, new things it can do. And one of the major things that pops out is what's called the torsion, sometimes called the torsion tensor. And that's math. Okay, I'm trying to be very clear about what the idea here is, because math is not the same as physics. So there is an idea called Einstein-Cartan gravity, which is physics, but just the math doesn't determine the physics. The fact that you have proposed that the torsion could be separate, that the connection could be separate from the metric, does not uniquely pick out what the physical theory that uses that fact might be. I'm mentioning that because it'll become important in a second.
2:19:07.0 SC: So there is a sort of standard version of Einstein-Cartan gravity, and it doesn't introduce any new particles or dynamical fields. It just sort of changes the relationship between some of the existing fields. And people showed a while ago, like at least in the '70s, that under certain circumstances that can lead to a repulsive kind of force that might arguably get rid of certain singularities in general relativity. However, don't get too excited by that. It doesn't get rid of all the singularities. There's no theorem that says it does. And honestly, there's two huge problems with Einstein-Cartan gravity. Number one is it's not a theory of gravity. Gravity is exactly the same in this theory. It's a theory that modifies the behavior of the matter in the gravitational theory. But the left-hand side of Einstein's equation, the one that has all the tensors describing the curvature of spacetime in it, those remain unchanged in Einstein-Cartan gravity.
2:20:04.0 SC: Basically there is a torsion tensor, but it's non-dynamical, so you can sort of eliminate it by a clever change of variables. And all you're doing is adding new terms to the way that the matter fields enter the Einstein equations. So it's kind of just a rewriting of good old gravity with certain modifications of the matter side. The other thing about Einstein-Cartan gravity is if you care so much about torsion, if you really think that the connection should be independent of the metric, then you might as well make it dynamical, right? I mean, if it exists, then it's playing an important role. Why in the world would you just not add extra terms that let it propagate so that you need to specify the value of the torsion in addition to the value of the metric? These are called dynamical torsion theories or connection dynamic theories. And the thing is that there's still a quantum field theory, right? Everything should still obey the rules of quantum field theory.
2:21:02.8 SC: So I mention all this because when I was young, I was super interested in these theories. I thought it was very exciting to take a different geometric point of view on general relativity. But I also knew enough quantum field theory to know that you should treat it like a quantum field and do effective field theory. And the thing is, unlike, this is the crucial part, unlike the metric, which has certain symmetries, which means when you quantize the metric, its mass is zero. Okay? So the propagation of gravitational waves goes at the speed of light because they have zero mass. That's not true for the torsion tensor. There's no reason at all for the torsion to have zero mass. It's very easy to give it a mass parameter. That's completely consistent with all the symmetries and all the geometric origin of the theory, et cetera.
2:21:48.6 SC: Indeed, there's no reason not to give it a very large mass. And you would expect that its mass would generally become quite large, and therefore it would be like up at the Planck scale or something. And if it couples to other matter fields, it would quickly decay into other things. There's no reason just in the nature of torsion itself for it to give rise to a long-range force of any sort. So I actually wrote a paper with George Field pointing this out. And for our purposes, we're basically concluding we shouldn't have been excited by torsion all along. It's actually just not that interesting from a particle physics point of view.
2:22:26.7 SC: Matt Becker says, "Do you enjoy gardening or have any other outdoor hobbies?" Short answer is no. I do have hobbies. I don't even have that many hobbies. My hobbies are happening right now. Podcasting is my hobby. Thinking about the world is my hobby, for the most part. As you all know, I also do other things. The hobby that I'm cultivating more than any other one right now, other than those obvious ones, is cooking, actually. I'm getting into cooking now that we own our own house and we have a kitchen that is big enough. And I like cooking. It's the closest I will ever come to being an experimental scientist. You don't want me in a laboratory doing real experiments, but I can chop some vegetables and things like that.
2:23:11.1 SC: Outdoor hobbies? No, not so much. I've never really lived in an environment since I was a kid where there was an outdoor to have a hobby. Like I said, my LA place did not have any outdoor space. So maybe I'm going to try... I'm trying to convince Jennifer that maybe she wants to start to take up herb gardening or something like that so that I will have fresh herbs for my cooking adventures. That's the closest I can come.
2:23:33.0 SC: Jeroen says, "Do you think that there are interesting dynamics in quantum theory that can be underexplored because they don't map well onto particles?" I don't think that's exactly right. So to reiterate something I briefly said earlier in this AMA, when you have a quantum field theory, you quantize it, it looks like particles. That's a little bit quick, what I just said. Under the right circumstances, it looks like particles. Like gluons never look like particles because they're stuck inside strongly interacting particles. So really what you mean by a particle is a freely propagating particle, or at least one that doesn't interact strongly with its environment, and gluons are never those.
2:24:16.2 SC: Sometimes people draw Feynman diagrams to represent gluons and quarks inside a proton or neutron, but they're not at all accurate or representative of what's going on. Again, as I say in the book Quanta and Fields, inside a strongly interacting particle, you have to think about it in terms of fields rather than in terms of particles. That's why the Compton wavelength of a quark can be much bigger than the actual size of a proton, because it's really not a particle inside the proton. So Jeroen is correct in the implication that maybe when we think about quantum field theory, we too quickly go to the particle language, since it is often very convenient, right? When we talk about electrons or protons or neutrons, or for that matter, photons or neutrinos, talking about particles is completely okay.
2:25:10.2 SC: But there are other circumstances when it's not okay, so you might well ask, are there cases that are underexplored because they don't map well onto particles? I think that an honest answer is that when we talk to non-physicists, when we explain what we're doing, particle physics, standard model, Higgs boson, all that kind of stuff, it's just much more easy and tangible to talk about particles than it is to talk about fields. Okay? People know what you mean when you say particle; they don't necessarily know what you mean when you say field. And when you measure something in the Large Hadron Collider, for example, you see tracks as if what you're looking at are particles, not fields.
2:25:51.5 SC: So the particle language is both vivid and correct in a certain regime. But the professional physicists who are thinking about this stuff aren't trapped by that particular aspect of quantum field theory. They know that it's really fields, and they think about the fieldness of it all the time. Like I said, inside a proton or neutron, or for as a very vivid example, the Higgs field, okay, filling space and having an expectation value that gives mass to other particles, that's entirely a field phenomenon. There are even more subtle phenomena in the strong interactions where the topology of the field configurations is super important for understanding observable physics. So I would say that the real physicists actually do think about the field nature of quantum fields quite a bit.
2:26:42.7 SC: Mr. G says, "It was recently reported by Sludge that former podcast guest..." Why do they call it Sludge? You know, we used to have things like Newsweek and the New York Times, and now we have Sludge doing reporting. Okay, anyway, I'm getting old. "It was recently reported by Sludge that former podcast guest Cornel West's 2024 presidential campaign was indirectly funded by Elon Musk through the Interstate Priorities Election Fund. While it's not clear that West knew of this involvement, it's difficult to reconcile with the purportedly pro-democracy rhetoric of someone like West with the practical effects of his and similar ostensibly pro-democracy political actors' actions working against American democracy. How should liberals act towards those who claim to support democracy but actively work against it?"
2:27:32.4 SC: Well, I think we should be honest. Look, I think that Cornel West's presidential campaign was just silly. It was never serious. Maybe he took it seriously, but it was never something that had any traction or any realistic chance of affecting the race in any way. I did see the story, but I don't know the details about the funding sources. I'm very, very open to the fact, without having any actual knowledge one way or the other, that he might have received support from some political action campaign without being 100% aware of what the ultimate donors to that political action campaign were. And maybe some part of that, if that's true, maybe some part of that is a little bit of negligence on his part. But also the Supreme Court and Congress have made it very easy to hide the ultimate sources of money in these campaigns.
2:28:24.2 SC: So I find it completely plausible that he would just have no idea, not that he was trying to act cagey about it or anything like that. I don't like to phrase it as an action working against American democracy. I mean, he might have been ultimately the only effect of what he did. If he did get more votes, he might have siphoned votes off of the Democrats and therefore given Republicans a bigger chance of doing well. That might have been a possible thing, but that's not anti-democracy; that's just anti-one political party. Okay? So I do think that the rules should be very different. I think that we should be much more transparent about where money comes from, and I think that people should not run what are essentially vanity campaigns for president. I think that voting for president and choosing our leaders is too important to do that. But I don't know the details of who knew what when in this particular case, and I don't think that thinking of it as anti-democracy is the right framing, quite.
2:29:25.6 SC: Rob Butler says, "If a potential solution to the fine-tuning problem is a multiverse, and if they all expand like ours, what prevents the universes from expanding into each other? If they had different laws of physics, would they interact or rather pass by like ships in the night?" Here's a case where the language is just getting in our way, and I apologize on the behalf of cosmologists everywhere for the bad way in which we use language. When we talk about the cosmological multiverse, the multiverse that may very well play a role in accounting for the apparent fine-tuning of parameters and laws of nature, they're not really different universes. That's just bad language. They're different parts of our universe, parts that are far away. And they don't pass by; they're just far away. What prevents them from expanding into each other is just the simple fact of non-Euclidean geometry. They don't expand in a pre-existing space. They both expand as part of the procedure, the feature that space itself is expanding all the time. Forget about the multiverse. If you just have a single expanding universe, it's not expanding into anything at all. The amount of space is increasing; the volume of space is getting bigger.
2:30:45.3 SC: Same thing is true for the multiverse. Each little part of the multiverse can get bigger without ever running into other universes. There are interesting questions to ask about what is sort of going on in the crossover regions between one universe and another, and that's going to depend on details of your theory: the potential energy function for some scalar fields, or for string theory, or whatever it is. But there's no worry that one universe is going to run into ours. One universe could swallow and take over, I don't want to say there's no worries at all, one universe that had a lower energy could eat up another universe and change it into part of itself by winning the energetic battle in some way. But again, I don't think that's actually a very big worry at a practical level.
2:31:33.7 SC: David Maxwell says, "Like you, I've always believed in letting ideas flow and trusting people to sort through the truth. Recent years have made me question that—" there's probably a typo here, "question, I'm gonna say, whether media landscapes have enabled systems where large percentages of the population only see total fabrication and never see the true ideas. So have your views on censorship changed?" Well, my views on censorship have not changed. No. Look, it's always been the case that large percentages of people have received misinformation and false information. You look at yellow journalism back in the late 1800s; they were not getting good information about what was going on in the world. And that's just always been true no matter what the media has been. The velocity has changed, the amount has changed, the barriers to entry have been lowered. So individual people can hop on the misinformation train much easier now than they could back in the day when you needed a printing press to do it.
2:32:43.8 SC: But I think the basic human nature that makes it happen hasn't really changed, and I don't think that censorship is the answer. We've talked about this on the podcast in various ways, Petter Törnberg, Gordon Pennycook, et cetera, how information flows through the internet and how that increases or decreases people's accuracy. I do think the situation is bad, don't get me wrong. I think that there are big problems in how citizens of a democracy are getting their information. There was a recent edition of Philosophy Comics, Philosophical Comics? I forget what the name of the comic strip is, but there's this wonderful comic strip and it comes out occasionally. And in this one, Rousseau, Jean-Jacques Rousseau, is thinking about inventing democracy. Of course, he didn't invent democracy, but that's okay; it's a simplification for the purposes of a comic strip. And he says, "All we need to do is have civic-minded citizens who are well-educated, and then a democracy will work."
2:33:48.3 SC: And then the punchline is, of course, we have democracy, but we don't have well-educated citizens, nor especially civic-minded ones. So I do think it's crucially important to think of ways that people can get better information and also more recognition of their civil responsibilities in general, but I don't think that censorship is the way to do that. I'm not quite sure what the way to do that is, but I do have my preferences for what not to do.
2:34:15.4 SC: Josh says, "You've mentioned playing poker a few times, so I've got to ask: assuming it's Texas Hold 'em, what is your favorite starting hand? Are you a big pocket pair kind of guy, or do you get more excited seeing suited connectors?" So to those of you who are not poker players out there, this is actually a sort of sneakily good question, I think. I think people don't appreciate necessarily... People who are not frequent poker players don't appreciate this subtle difference here. In Texas Hold 'em, you are dealt two cards, your hole cards, that you see and nobody else does. And then the board, the common cards that everyone shares, come out gradually in front of you, and you make the best five-card hand out of your two cards plus the five cards on the board. And so you don't know what other people have; you do know what you have. So it's a game of incomplete information.
2:35:10.6 SC: And if you just ask the question, of all possible hole cards that you could have, which ones will lead you to having the best hand at the end of all the cards coming out the highest percentage of the time, the answer is clearly pocket aces. Two aces in your hand are going to win the hand more often than anything else. But the twist is, in poker, in Texas Hold 'em, the goal is not to win the hand; the goal is to win money. And of course, you win money by winning hands, but you could easily find yourself having a strategy where you frequently win hands with very tiny amounts of money in the pot and occasionally lose hands with huge amounts of money in the pot. And even though you win more hands than you lose, you overall lose money.
2:36:02.2 SC: So there's another kind of hand called suited connectors, which means two cards of the same suit which are right next to each other, like ten of spades and nine of spades. Okay? Those kinds of cards are more likely than average to eventually get either a straight—five cards in a row, or a flush, or even a straight flush. A flush is five cards of the same suit. And those kinds of winning hands will also win. They'll win less often than just having a big pocket pair, but when they win, it's possible that the opponents that you're playing with don't give you credit for having such a good hand and therefore are willing to put more money into the pot. So even though those hands win less frequently, it is possible to win more money with them. And therefore, it's a reasonable question: which would you rather start with? I actually don't care. I know that when we start, there will be a given distribution of hands, and you have to play all of them effectively.
2:37:05.7 SC: I do think that there's a certain thrill that you get from having bad starting cards and entering the pot anyway and getting really lucky and fooling everyone and taking down a big pot. There's nothing more fun than that. If you just have two aces to start and you end up winning because no one has a hand that good, you're like, "Okay, good, I got lucky, I got good cards, and I won the hand." But it doesn't feel like skill; it just feels like you got good cards. Whereas somehow if you have three-five suited and you end up getting a straight, and your straight is two, three, four, five, six, no one thinks you have that hand because what are you doing being in the hand with a three and a five? But therefore, you can win a huge amount of money if someone else has a pretty good hand, like someone else has pocket aces, then you have a big pot in front of you.
2:37:57.5 SC: But that's just very, very rare that's gonna happen. So I just take it all. I take the good with the bad. I appreciate the thrill of doing a little trick, but ultimately you gotta squeeze whatever value you can out of whatever cards you are dealt. That's true in life just as much as in Texas Hold 'em.
2:38:16.4 SC: Alan says, "In the previous AMA, you characterized disagreements about free will as largely semantic, despite arguing just minutes earlier in the same episode that emergent phenomena are genuinely real. You also seem to endorse the view that free will emerges from agency in the recent Christian List episode. Can you reconcile the apparent contradiction or clarify whether I'm missing something here?"
2:38:38.3 SC: I think you're missing something. I mean, yeah, I do think emergent phenomena are genuinely real, and consistent with that, I think it makes most sense to treat free will as real. I think that's pretty consistent. It's not treating free will as real that I think is a semantic argument. I think it's debates between people who treat emergent compatibilist free will as real versus people who simply insist that if the universe is deterministic, you can't have free will by definition. That's a semantic argument. That's just arguing over the definition of free will, and I think that that's kind of boring. So that's why I try in those cases to not use the term free will and just argue about what we think happens in the world. The reason why Gregg Caruso was a good person to talk to on the anti-free will side is because that's not his reasoning, right? He's not just disagreeing with compatibilism by definition; he doesn't like it for some more normative reasons. And so that's interesting to explore.
2:39:41.9 SC: Paul Torrek says, "How important is it for Everettians to derive the Born rule? Obviously, it would be nice to reduce your theory's number of assumptions by one, but don't most other interpretations of quantum mechanics help themselves to an empirically validated Born rule rather than deriving it?"
2:39:59.6 SC: Well, they do, but the situation is different. Everettians have to derive the Born rule in a way that others don't. So if I have, for example, some theory in which the wave functions truly collapse, objective collapse models, for whatever reason, there's different versions of those, okay. But in such a theory, there will be some indeterministic, some truly stochastic element that describes wave function collapse. And then how you interpret the Born rule is as a frequency, right? When you have a certain setup and you make a certain measurement, you will do that experiment many, many times, and the simple frequency with which things happens is what you mean by the Born rule. It's a frequentist version of probability, if you want to put it that way.
2:40:50.0 SC: In Everett, you can't say that. In Everett, when I observe a spin, that I get some spin up and some spin down in the wave function, with 100% probability, there's a world where the spin was up and 100% probability there was a world where the spin was down. There's no frequency there. It happens all the time, every single time. There's a world where the spin is up and a world where the spin is down. You have to treat probability epistemically, subjectively, not frequentistly, objectively, okay? You have to interpret, I would argue, you have to interpret probability as a measure on your self-locating uncertainty when that wave function branches. And the theory doesn't have any room for saying, "And assign this probability," because the probability is 100%. That's the assignment: the probability that it happens somewhere in the wave function. The thing you're talking about is not something the theory gives you, okay? The theory doesn't give...
2:41:48.0 SC: Well, I think it does give it to you, but it doesn't have room for an additional assumption that gives it to you, because the thing you're trying to make an assumption about is your subjective probability on self-locating uncertainty. What the theory is allowed to do is give you different physical dynamics, it's allowed to give you different ontology; it's not allowed to tell you what to think, okay? That's what the Born rule comes down to in Everett. So I think that's all a resolvable question, but it's a different kind of question in Everett than it is in other approaches.
2:42:19.8 SC: Gustavo Chavez says, "In the March 2026 AMA..." by the way, I am appreciating that people are following up on old AMA questions. That's always a good strategy. "So in the March 2026 AMA, while answering a question about your take on the notion of computational functionalism, you said that you think that the underlying processes that lead you from the input to the output are crucially important in whether or not those processes involve something that we can call consciousness. In his recent essay, 'The Mythology of Conscious AI,' Anil Seth challenged the notion of computational functionalism, arguing, among other things, that there are functions that Turing machines cannot compute perfectly, such as continuous or stochastic functions. As I understand it, he argues that perhaps consciousness depends on processes that a Turing machine cannot compute, but only simulates in a way that isn't precise enough to allow for consciousness to emerge. Is this more or less what you're thinking about when you mention the importance of the underlying processes happening in our brains, or was it something different?"
2:43:18.9 SC: I think it's a little bit different. So it was Anil Seth who first opened my eyes to the importance of distinguishing between the information flow in our brains, which is what computational functionalism cares about, and all the other physical goings-on in our brains, which is what biology and what Anil refers to and others refer to as biological naturalism care about. But that's not what we talked about when we had Anil on the podcast; we talked about emergence and things like that. But Anil did give a talk at the year-and-a-half-ago Natural Philosophy Symposium. By the way, we will be having another one in May 2027, so look forward to that. But his talk at the symposium was about biological naturalism versus computational functionalism and its importance for consciousness in AIs, and you can look up that online. The talks are on YouTube.
2:44:13.3 SC: So anyway, it was that talk and the paper that he wrote about it, and he really... I feel bad, he asked me to respond to that paper as a part of a whole special issue of the journal the paper appeared in, whose name is escaping me right now, but I never got around to doing it. I was just too busy doing other things. But I did read his paper and think about it, and that's what got me interested in this topic. And we later talked to Ned Block, who is the other major proponent of this perspective. Okay, so I try to be...look, this is not something I'm an expert in. I don't think you should listen to what I have to say about these things. I'm just trying to be honest about what my opinions are. But I'm not at a high level of having gone through all the details here. But the important thing to me is exactly what I said: the distinction between the information flow in your brain versus all the actual biophysical processes going on in your brain. I am open to the idea that whatever consciousness is, it requires more than simply information flow to capture it. Okay. It might not.
2:45:26.3 SC: I'm also open to the idea that it doesn't, because I don't know a lot about consciousness, but I'm open to the idea that it does. And Anil puts a lot of emphasis on this Turing machine stuff. I don't especially. So I'm not sure whether or not human brains are Turing machines, or better or worse, or even whether that's a good comparison. My interest is more in entropy and the arrow of time, which will come as a surprise to nobody. I think that the mechanisms of metabolism and turning free energy into useful work that go on in our biological structures are very, very important to what you and I call thinking and perhaps to what you and I call consciousness. They can be mimicked by other kinds of structures, but maybe they can't be exactly reproduced, or at least not in a simple way that abstracts away all of the underlying mechanism into a simple information flow through a computer.
2:46:28.7 SC: Again, maybe it can, but I'm open to the possibility that it can't. 'Cause when I first heard Anil's objections to computational functionalism, I was worried that he was abandoning physicalism. But he explained that he's not in any sense abandoning physicalism. So I was impressed to learn how you could stick with physicalism and yet be concerned that AIs might never be conscious. I thought that was interesting, and that's something that one should consider when investigating the question of whether AIs could ever be conscious.
2:47:00.8 SC: Okay, I'm gonna group two questions together. One from Christian Dobo, who says, "Do you think AI will be able to produce new theories in physics just by the sheer fact that it has access to a vast amount of papers, findings, and theories and is able to hold them in its head and freely recombine them? I understand your point about physics, that given its experimental nature, it's less susceptible to AI breakthroughs, but I have a feeling that there might be low-hanging fruit, low-hanging yet important fruits hanging around that are not discovered because of the immense specialization and amount of knowledge in science." And the other question is from James Allen: "What do you predict would happen if you constructed a training data set for an AI that contained only the data from quantum mechanics experiments? Could you build up an agent whose intuition about the world were quantum mechanical instead of classical? Do you predict being able to extract novel insights from such an agent?"
2:47:51.9 SC: So both questions are about how well AI can do with sort of coming up with physics theories. They're very different questions, so I apologize to Christian and James for combining them, but it's useful to answer them sequentially. So to Christian's question, I don't think that's the way it's gonna work. I think that, look, AI has an impact on how physics is done. I think... As I'm saying this, I'm realizing there's another question still to come where I'll talk about this again. AI is an accelerator to doing theoretical physics. It can help you straighten out your own ideas, inspire new ideas in you, and things like that. I can't with any confidence predict what AI will do when AI becomes something different than what it is right now. Okay? I don't have any objections in principle to AI being much better at theoretical physics than human beings are. But I don't know, because we haven't built those AIs yet. All I can talk about is what the AIs are doing now. Right now, they're not especially good at coming up with new ideas in theoretical physics.
2:48:59.5 SC: And it's not a matter of having been trained on the huge number of papers, findings, and theories. In fact, that is a little bit deleterious. The problem is, I mean, that's super helpful to the human being who is talking to the AI and saying, "Has anyone ever thought about this?" and they can help you with that, okay? But it creates a problem for the AI because they're not especially good at judgment. They don't have really good physicist intuition. So they're reading a billion papers, and to the AI, these are all papers, right? They gotta take them all seriously. And you might say, "Okay, where's the consensus?" and stuff like that, "Where do people disagree?" and stuff like that. But all that does is drive the AI towards sort of the most mainstream view, okay? The ability to look at a less mainstream paper, one that doesn't agree with other papers, and say, "Oh wait, this is a good idea," I've not seen that in AI at all. It's quite the opposite. You have to coax the AI into even accepting a point of view that is not the mainstream one. Again, I'll just keep saying it over and over again because people don't wanna listen to me: that's not a necessarily permanent condition. You could imagine doing better, but that is what I observe in AIs right now.
2:50:20.9 SC: And likewise with James's question. I think in James's question it's even worse because he's saying, imagine the training set is only data from quantum mechanics experiments. So one little footnote here: every experiment is a quantum mechanics experiment because the world runs according to quantum mechanics, right? There's the classical limit in which you don't need quantum mechanics, but you can't do an experiment in the physical world that isn't ultimately a quantum mechanics experiment. So that's a footnote there. But the other thing is, I'm not even sure it makes sense to say, what if you only trained an AI on data from quantum mechanics experiments? Like, what does that mean? You just give it a set of zeros and ones? You have to accept that when you look at data from any experiment, you're always doing so within a theoretical framework. If I say, "Oh, there's a dot on a screen at that location," I am presupposing the existence of screens and dots and locations, and that you know what that means. That's already theoretical superstructure, right?
2:51:24.4 SC: That's why physics is hard, because sometimes you have this theoretical view of the world and you realize it's insufficient and you have to do better and you have to update. And so I think that you couldn't just train an AI in a theory-free way on data from experiments. We start with some folk view of the world and then we try to do better building on that. We can't start in the world in a way that has no latching onto the things that we're familiar with all the time. So I'm not even sure how you would do that kind of thing for an AI. Now, I might be wrong about that, and that's something where maybe a clever person could overcome the obstacles that I see.
2:52:07.2 SC: Jonathan Kleiman asks a priority question: "How do you distinguish genuine academic consensus from people being afraid to disagree? Has that fear contributed to public distrust of universities, and what should they do differently?" Well, look, I'm mostly in science. I'm half in the humanities now, technically, because I'm in a philosophy department. But there's so much going on in this question, I don't even know where to start. What do you mean by consensus? Okay. Academics argue with each other all the time. You can say, "Oh, string theory is very popular," but in my physics department, in every physics department I've ever been in, even one with big-name string theorists, there's plenty of people who think that string theory is entirely bullshit and are very happy to say so. Academics love arguing with each other. That's what they do. Even if you're almost entirely in agreement with some fellow academic, what you tend to first do is find the disagreement and then spend all your time talking about that.
2:53:10.7 SC: There might be consensus in the sense that, okay, we have general agreement that the standard model of particle physics fits the data, but it's not like there's some enforcement mechanism that means that you need to agree with everybody. If you can't get along in academia, there's a use of the word consensus in the popular discourse because we do want to distinguish crackpottery from real science, okay? So there are academics who don't think that human activity contributes to global climate change at all. Those people are crackpots, even if they're well-respected academics in other areas. So we talk about there being a consensus of anthropogenic climate change, okay? It's only for PR purposes that we talk about that. What it means at a more careful level is that there's an overwhelming majority of people who have thought very carefully about this and come to a certain conclusion that they agree on. There's always gonna be dissenters and whatever. And typically those dissenters are allowed to go about their business, right? Like they're not fired or whatever. Some of those dissenters have Nobel Prizes.
2:54:18.4 SC: So but I think the important thing is that the very fact that academics tend to disagree with each other all the time, and there are dissenters, and those dissenters are sometimes very respectable, and yet academics do appreciate sometimes that we want to be clear about what we mean, has led to an opening for people who want to attack academia in bad faith. And I think that the public distrust of universities, to the extent that it exists, comes much more from bad faith attacks on academia than it comes from academics themselves. And academics have been slow to learn this. I remember back when I was writing my book on the Higgs boson, if you remember, there was a worry that we would turn on the Large Hadron Collider and destroy the Earth, okay? And look, physicists live on Earth. They don't want the Earth to be destroyed either. They took these worries seriously, they thought about it, they wrote papers, and they decided that as far as they could tell, there was no serious threat. But nevertheless, there were people who said there could be a threat. And when physicists say there's no serious threat, they don't mean the probability is zero. They mean the probability is so incredibly tiny you don't have to worry about it, even if it's not zero.
2:55:40.8 SC: And that subtlety doesn't come across, okay? Physicists know what you mean when you say that, but the outside world says, "Oh, so you're saying there's a chance." And I remember that there's an interview that John Oliver, back when he was on the Daily Show, he did an interview with John Ellis, who is a very well-respected theoretical physicist who works at CERN. And he had learned this lesson. John Ellis had finally learned this lesson, I'm sure by getting it wrong before. I haven't talked to him personally about it, but he they did an interview, and John Oliver is talking to John Ellis and asks like, "Is there a chance that the Large Hadron Collider will destroy the world?" And Ellis just goes, "No." And that's it. And that's a PR move. If he's talking to his colleagues, he would say, "Well, this could happen and this could happen, but it flies in the face of literally everything we know about physics, but it could happen." When he's talking to the public, he learned he says, "No," okay?
2:56:42.3 SC: And bad actors have taken advantage of this attempt on the part of scientists to be honest amongst themselves and try to also give the outside world the most accurate vision of what they believe, even if it's not 100% accurate. And so I'm in favor of being honest. I always try to be honest myself. But I know that I'm reaching a smaller audience that is relatively well-informed and things like that. And when people ask me about science communication as a discipline, how do you get better at it? I try to be very explicit that being 100% accurate in your statements feels good, but it's not necessarily the best science communication. Not because saying untrue things gives you better, I don't know, income or whatever or better influencer dollars, but rather because you can say 100% true and accurate things, and depending on the mental state of the person listening to you, they may come away with a less good understanding of what's going on than they had before.
2:57:59.8 SC: Your goal is not to say true things. Your goal is to improve other people's understandings. And sometimes you have to give an analogy, right? Or give a simplification or say something that is most often true but not always. And that's part of the job. That is absolutely part of the difficulty, why science communication is not automatic, why you actually have to think about how to do it. So I think there's a lot going on. I think that working academics, I mean, sorry, I meant to say this at the beginning of the answer. There are effects within academia where certain sets of ideas are more favored than others. Of course there are. How could there not be? There's a very precious, rare commodity resource, namely jobs. Who's gonna get to be hired? Okay? And if you're a department that does a certain kind of thing, it's easier for you to hire people who you agree with than people who you disagree with.
2:59:04.0 SC: Now, a good department will recognize this. We've had these discussions in all the departments I've ever been in. We do a lot of X, but we don't do Y. Should we hire people who do Y? Right? And that's hard to do, but you try to do it because you want to have some level of diversity. But you also don't want to hire crackpots. So if you think that what the other person is doing is so wrong as to not be worth pursuing, then you're not gonna hire them. And again, I don't see that that's a bug in the system. I literally don't see any other way to do it. Are you gonna advocate that departments should hire people that they think are crackpots? Well, if not, then you're gonna have to have a way of defining who's a crackpot and who isn't. So I do think that there are schools of thought in academia, and fitting into a school of thought can absolutely help your career. That much is true. I don't think that that's the origin of any public distrust of universities. I think the origin of public distrust of universities is bad actors who don't like universities trying to undermine trust in them and generate fear. And that's actually been pretty successful, and I think that that's terrible.
3:00:15.6 SC: P. Walder says, "How do you think AI will inform theoretical physics in the short term, 3 to 12 months?" So yeah, this is the question I referred to earlier was going to come. So I think I mostly answered it. I think that it won't be a revolutionary impact in the short term, 3 to 12 months. Like I said, it's an accelerator. It helps you, it helps move you forward, both in the sense that it can do calculations for you. If you have a well-defined problem, mathematically carefully formulated, AI can absolutely help you find the answers to that. Like I have friends who are using AI to help solve problems in string theory or something like that, if it's very well mathematically formulated. But it doesn't do a good job in coming up with the questions to ask. Again, short term, 3 to 12 months, no statement about what the ultimate equilibrium will look like. But right now it's a partnership between human beings coming up with questions, AI helping you to answer them, both in doing the calculations, as I said, but also in just understanding the lay of the land, understanding what has been written. I know that that's what I mostly use it for. Basically, I use it for drawing plots.
3:01:24.9 SC: And for figuring out what's out there in the literature. I recently asked AI, "Has anyone connected these papers by Ted Jacobson on entropic approaches to gravity to this older idea called unimodular gravity?" And the AI is like, "Yeah, sure, people have done that. Here's these papers, not that many of them, but here they are." And then I can go read the papers. Because ultimately you have to read the papers. You can't trust the AI to do your work for you. I think a lot of the fear of AI or distrust of it comes from the obvious fact that you can misuse it, that you can offload your thinking to the AI and get in trouble that way. But you can also use it responsibly. And so I think that people are gonna hopefully be doing that. What I hope not to see, but I worry about seeing, is just a flood of kind of crappy papers, right? It's easy with AI to write crappy papers at a very, very high rate. I mean, there are people out there, not that I know of in physics, but people in other fields that just write a paper every couple days, because why not? The AI will do all the work for you, right? Those papers are not gonna be very good, very interesting. You barely have time to read those papers, much less write them. So I think it's gonna look very bad on your CV in, five years from now, if you turned out 100 papers a year. That's not gonna look like you're productive. That just looks like you're not even reading your own papers.
3:02:52.5 SC: C. Gerlando says, "In your April AMA, you discussed Habermas's concern that economic and administrative systems can colonize the life-world, displacing communication and shared meaning. A successor in that critical tradition, the sociologist Hartmut Rosa, has developed a related critique through his concept of resonance: relationships with people, work, nature, or art in which we feel addressed, can respond, and are changed by the encounter. In such moments, the world seems to come alive and crackle, drawing us into encounters we cannot fully control. Rosa proposes resonance as a criterion for a good life beyond wealth or the mere satisfaction of preferences. I find this compelling. A society could become ever richer and more efficient while making resonant relationships increasingly difficult to sustain, and that is bad. Do you think resonance could provide an objective standard for judging a society's success, or at least a standard that goes beyond subjective preferences? What would justify treating it as such?"
3:03:53.4 SC: So, I mean, I hope it's not gonna be a surprise, I can't speak to the specific details of this concept that I've never heard of before. I mean, I've heard of the word resonance, but not in this particular context. And so I'm not gonna judge whether it's useful or not based on the description in this question, as clear and nice as it was. There's probably more that you have to learn before you could actually have a good opinion about that. I do think that, going back to all these other questions about AI and so forth, it's not just AI, but the internet, the way that we have jobs and our economic situation, et cetera., have dramatically changed human life over, let's say, the last couple thousand years. All right? No doubt that we live in a different way than we did before. And we are not necessarily built for it, or we're not necessarily expert at navigating it. We can sort of get along in ways now that weren't possible before, and those ways might...
3:04:54.2 SC: It's like, what I'm trying to say is, we like to eat sugar, right? We like to eat sweet things. They might not be good for us, but we like to do it. And this is a generalization, of course, but why do we like to do that if it's not necessarily good for us? Well, we grew up in conditions of scarcity where finding calories was hard. By "grew up," I mean as a species. And so we overindulge. We're not trained to regulate the amount of these substances that we put into our bodies because they used to be self-regulating; they were just scarce.
3:05:28.6 SC: In analogy to that, we can now get through life in ways that things that were scarce to us before might be very, very plentiful, and that might take us away from the things that we need to work harder for but are actually more rewarding. One of the sad statistics that you might have read recently is that people are just spending less time interacting with other people. People are spending more time at home by themselves on their screens, less time having friends over, going out with friends, et cetera. People are having sex less, including young people. This is all perhaps not that surprising if you think that the way the society is organized is shifting out from underneath us and we haven't adapted to it yet. So the reason I'm answering the question is not because I have anything deep to say about the idea of resonance, which sounds interesting to me, but I don't really know about it, but just to indicate that I am sympathetic with the idea that human connection is important and something that we have not optimized society to provide for us. It is arguably becoming harder and harder to get true human connection, and I think it is important.
3:06:45.6 SC: And when I say that, I recognize that it's not of equal importance to different people, right? There are people who prefer to be mostly alone, people who prefer to be mostly socializing with other people. That's fine. But there is something about the health of your psyche and your soul that requires at least some contact with other human beings. And hopefully, as the question says, it's the sort of non-trivial kind of contact with human beings where you truly talk to somebody, truly interact with them, not just at a superficial small talk level, but at a deeper level. And I think that those things are important. And for whatever reason, not only can you get away with not having that, but we don't even necessarily value it that much explicitly anymore. So I do think that technology is changing the world in which we live in radical ways, and we are gonna have to adapt as human beings to preserve some of the things that we were sort of forced to do before but were actually good for us to do before.
3:07:57.2 SC: All right, Robert Ruxandrescu says, "Have you thought about adding some perks for the higher-paying members of Mindscape? For example, you could hold a trimestrial or nearly two-hour Zoom meeting with the higher-tier paying members, have a live chat with them like a live AMA where you could have short back-and-forth conversations. What do you think?"
3:08:15.2 SC: Yeah, I have thought about different kinds of things. Right now, for those of you who don't know, we have tiers of support on Mindscape, and it's a Patreon thing, right? You have tiers, and if you support at a certain level, you get some benefits; at higher levels, you get higher benefits. I couldn't be bothered to set any of that up. So we have tiers, you can support at various different levels, but the ones who are in the higher-paying tiers get nothing, get no extra things that they don't get from the lower-paying tiers. So it's just out of the goodness of your heart that you support at a higher level, and of course my appreciation for that is great. But of course, anything that I might do to provide more benefits to higher-tier supporters takes time. And as I said, I'm not trying to complain about this or anything, but I have very limited time to do extra things. And by very limited, I mean none, I mean a negative number. I'm already very overcommitted right now.
3:09:08.9 SC: So honestly, the podcast takes up too much time overall. I work very hard to make the entire amount of time I spend on the podcast one day per week so that I can devote time to writing and teaching and doing research and talking to my students and living my life and things like that. Anything I add on top of that becomes very, very difficult to do. And it's... Look, I'm not in it for the money, right? The money is good, it keeps me motivated to do it, but I'm by no means trying to optimize my income by doing this. If I were doing that, I'd be putting a lot more effort into the podcast than I am. So I'm open to ideas for ways that I can provide some extra benefit to the higher-tier supporters. They deserve it. If there's some way I could do that without it taking a lot of hours or whatever, then maybe I would start implementing that. I'm open to suggestions.
3:10:02.8 SC: The Priority Question, which is someone's name, not a priority question signal, says, "How much of an impact do you think that social media and their algorithms promoting echo chamber bubbles has on the current political polarization? And is this a bad thing?"
3:10:18.0 SC: Yeah, again, this could have been grouped with an earlier question. We've talked about this on the podcast. I think that there's a huge impact that social media and more generally our information ecosystem and how it's organized has on people's political opinions, people's awareness, people's information level about what's going on. And I think that they often do a bad job. Okay. I think that we get a lot of information that is not reliable or good or virtuous or whatever. I don't think that the situation is as simple as it might appear from the label of a bubble or a filter. There's various ideas that people explain to you, "Oh, you're in a bubble, you're only getting the ideas that you want to hear," et cetera. It makes perfect sense, but it doesn't really hold up to the data. Right? When people actually look at what kinds of information are coming through, people do get lots of different pieces of opinion from different sides. Generally, there's very few people who truly are living in a bubble. What matters is how you respond to the information that you're hearing and how much you seek out information that just reinforces what you're in.
3:11:30.0 SC: We talked about the election that just happened. It's very clear that a certain number of people on the Republican side had convinced themselves that despite all of the polls, et cetera., that the Republicans were not gonna do that badly. And partly this is because there's some self-reinforcing mechanisms there. They were told, they did get information to the contrary, but they chose not to pay attention to that information. So again, this goes back to what I was just saying before. I don't think we've figured out how to live in the new world, in this brave new world that we find ourselves in. But the new world is also changing so fast that I'm not even sure how we should try to figure that out, because by the time we've figured it out, the world will be different.
3:12:20.2 SC: I do think that... There's a certain romantic streak that I have that says we should really try to be better citizens, better human beings, better participants in our democracy. So we, including everybody, should strive to be better informed about relevant issues and the candidates and so forth. To me, it's just appalling that the people who vote have such an inaccurate view of the positions of the people they're voting for. Okay. And I don't always blame the people who are voting; some of them deserve the blame. But also the system, the media ecosystem, not just social media, but mainstream media, which likes to talk about horse races rather than issues and so forth. So I think it's a very complicated stew of messy situations that get us into this situation.
3:13:10.3 SC: Polarization as a word, it can't purely be blamed on social media. In fact, if anything, you might expect the proliferation of different voices to diversify the set of opinions out there. Right? And polarization is really a statement that people's opinions are lining up, that if you know what someone believes on one issue, then you're probably also going to figure out what they believe on other issues. That might not be what you expect from the existence of social media, but it is what we're getting. And I think that the political system, the reward system, at least here in the United States, is pushing us in that kind of direction. So I'm a data-driven person, I'm an empirically minded person. I think that you need to sit down and think about really what's going on. What all the different influences are and the different impacts that they all have before we can start thinking about how to make things better.
3:14:05.4 SC: Peter Hunker says, "I really enjoyed your episode with Greg Caruso. It got me thinking about the who behind the choosing. So here's a two-sided question: Who is Sean and when is Sean? Narrator, brain, or body, and over what time scale? Over seconds, you look autonomous, but over decades, your zip code does a lot of the choosing. Where do you draw the boundary of the self that chooses? Is it a fact or a modeling choice? And at what point does 'Sean did it' become just 'the universe evolved'?"
3:14:33.4 SC: Well, yeah, I think that these are the questions you ask when you try to make sense of the world at the higher emergent levels. This is why philosophers describe what they do as trying to carve nature at its joints. Where do you draw the boundary between this subset of the universe and that other subset of the universe? Of course, there's crazy kind of higher-level distinctions that we have to draw when we get into the many-worlds interpretation of quantum mechanics or Boltzmann brains or anything like that. But putting all those aside, even in the everyday world, there are questions about where is the self, where does it end, and where does the rest of the world begin? We've talked on the podcast with people like Andy Clark about extended cognition. Does your phone or even just the piece of paper that you have in front of you that you're scribbling on, do those count as part of your cognitive apparatus?
3:15:26.6 SC: And of course, with time, you change. You're a different person now than you were 10 years ago or even 10 seconds ago. But I think that we have to recognize that there's questions of practice and questions of principle. In principle, these questions are really hard. Where is the self? Where do you draw the boundary? Things like that. In practice, they're usually not. Under most circumstances, it's pretty clear who is a certain person and what is not that person, right? Even though people do change over time, so in some super-duper strict sense, the person you are now is not the person you were 10 seconds ago, it is pretty clearly useful to think of that person extended through time as different parts of a single unified being. Even though you can undergo radical transformations, that's certainly true, but nevertheless, you transform from something to something else via some mechanism that can be identified, and you can find the continuities from moment to moment.
3:16:25.5 SC: So I'm not gonna answer the question. I mean, the questions are perfectly good: Who is Sean? When is Sean? But I do think that they're answerable, and I think that there's kind of the obvious answers that usually work well. You do have to think harder than that, this is why philosophers get paid, you have to think harder than that when the conditions become non-conventional, when things become a little weirder, when you have either many worlds or even brain surgery or some other transformative experience like we talked about with Laurie Paul years ago. Does that truly make you a different person? I think that all of those are perfectly good questions to answer. And I do think that at the end of the day, it's a modeling choice. It is not a fact. We are choosing to coarse-grain and to group the world in ways that give us some leverage, give us some handle, give us some understanding about what's going on.
3:17:16.6 SC: Sid Huff says, "There's been considerable discussion in the news media recently about the concept of designing a kill switch that would shut down an AI system if it were determined to have gone rogue. From your understanding of current and future AI, would a kill switch be an effective solution to potential AI malfunction?"
3:17:34.2 SC: I don't know. My understanding of current and future AI is not good enough for that. I think that the more relevant thing to say is that people are working very hard on figuring out ways to stop AI from doing bad things. I think there's a problematic issue with the incentive structure that we have, right? I mean, the incentive structure where people get paid billions and billions of dollars to be leaders in the AI field is not necessarily one that is conducive to putting safety and prudence front and center. But nevertheless, some people are trying to do that. I do think that the government, bless its heart, is just not very good at regulating these things. Government moves very slowly. The technology moves relatively quickly. The people in government aren't that educated about it. The people who are developing it are educated about it. People in government often care more about getting elected and getting donations than about actually implementing good policies and so on.
3:18:34.2 SC: So I'm a little worried that maybe most companies do a good job at putting in mechanisms that stop their AIs from doing bad things, but one or two of the companies don't, and then we're in trouble because their AIs could do bad things. As I've often said, AI can be very dangerous. There's no question about that. I'm not that interested in the big picture existential risk kinds of questions, but the more relatively small-scale but still disastrous questions I think are very important to look at. I just saw an interesting study this morning about different ways to prompt AI to behave better. One of the problems, interestingly, one of the problems with AIs behaving badly is that there are so many science fiction stories and blog posts about AIs behaving badly. And then we train the AIs on those stories and they go like, "Oh, okay, I guess this is what I'm supposed to do." So they've been developing ways to overcome that by just prompting it, before you do anything, sit back and ask, "Is this really what you should be doing?" Very simple things, but then you can empirically test them and figure out which ones are effective. So I'm very open to the possibility that we can make AI relatively safe if we try. I'm less sanguine about whether or not we're going to try as hard as we really should.
3:19:57.9 SC: Michael Helgerson says, "With the NBA season in full swing and your work on the physics of democracy underway, what would your outline for the Physics of Basketball book look like?"
3:20:08.0 SC: I have no idea. Honestly, I'm not that drawn to the idea of a physics of basketball book. I think there's physics everywhere, but two things. Number one, the reasons why I care about basketball are pretty orthogonal to the reasons I care about physics. So I'm really looking for an escape from physics when I'm watching basketball. I'm not secretly looking at the physics aspects. I'm kind of intrigued by some of the game theory and strategic aspects. Why don't you just have the best player make all the plays, things like that? And there's good answers to that. But the physics of how the basketball moves and how quickly it goes and how you respond, that's not even my kind of physics. That's the other thing. So the first thing is I don't look at basketball for physics reasons. The other thing is the kinds of physics I do are pretty removed from the down-to-earth, everyday, full of friction and real-world dissipation and stuff like that that would be entirely relevant to playing basketball. So I do think that maybe there is a good book to be written about the physics of basketball. It's certainly not going to be written by me.
3:21:15.9 SC: Owe, O-W-E, says, "I was a little surprised in your reflection after talking with Chandra Sripada at how strongly you seem to take his perspective. One thing I found mostly missing in the conversation was what's built into language. My question is, do you think that Sripada was correct in ascribing so much intent to the models? For example, he mentions the models doing something like implicit Bayesian inference. Or instead, is something like Bayesian inference being built into the weights between words because humans use Bayes in constructing the sentences that LLMs are trained on?"
3:21:49.3 SC: Again, I'm not an expert on these things, so don't take my opinion too seriously. What I was impressed by in Chandra's way of saying things is that he wasn't just saying, "Well, look, the AIs talk like human beings, therefore they must be thinking like human beings." He was actually invoking psychological and cognitive science tests to figure out what the parallels were in the thinking process, not just in the relationships between inputs and outputs. So that was a breath of fresh air to me, even if it's not entirely persuasive. I think that there's a lot more work to be done there. But to the specific question, the difference between the model doing implicit Bayesian inference or Bayesian inference being built into the weights between words because humans use Bayes, I'm not sure there's that much of a difference there, right? I don't think that most models are literally running Bayes's theorem when they come into contact with new data.
3:22:48.6 SC: But then again, I don't think human beings are really running Bayes's theorem when they come into contact with new data either. So I'm very open-minded about this. As I've said for a very long time, it's extremely impressive how well LLMs can do at mimicking human thought and speech, the outputs thereof, I should say. And then there's an open question about whether the mechanism by which it gets there is similar or dissimilar to the way human beings do it. And I think it's probably both: similar in some ways, different in other ways. Chandra did a good job of highlighting the fact that in both cases, the importance of prediction leads you to certain cognitive pathways that actually do have similarities between them. And so that's an important thing to think about. I do think that we're still discovering this. I don't think it's something that we completely understand. I certainly don't completely understand it. And it's just a little bit tiresome sometimes to sort of try to read about this or learn about this because there's too many vested interests on both sides, right?
3:23:59.2 SC: There are people who have just decided that AI is the devil and it's terrible and everything about it is wrong and bad. There's other people who've decided that it's gonna save humanity and it's gonna take over everything and we gotta learn to maximize it or whatever. And there's just too many people who are not ready to take a step back, take a breath, look at the data with an open mind, and make up their decisions based on that.
3:24:23.9 SC: Brandon Lewis says, "I'm still trying to understand how to think about particles in quantum field theory. I get that fields permeate all of space. The fields are constantly oscillating within their degrees of freedom, but also coupled to each other, so the oscillations can move between fields. When that happens, we observe particles decaying or fusing. We can only observe discrete measurement outcomes of a continuous wave function, the infinite other possibilities lost to us once our universe has made its choice. So are particles in QFT just when the wave function has sufficient energy or density at some location?"
3:24:58.7 SC: So, yeah, I think almost everything you say is perfectly correct. You've been listening to very smart people, obviously. The one thing for the actual question at the end there, to be a particle, and this is interesting and a little bit counterintuitive, let's back up a little bit. Forget about quantum field theory for the moment, okay? Think about the quantum theory of a single particle, right? Think about the Schrödinger equation for a single particle. Maybe you don't know that much about it. That's okay, I'm gonna tell you. So before we invented quantum field theory, we still had quantum mechanics of individual sets of particles. The difference was that the particles couldn't come or go, right? They couldn't decay or annihilate or anything like that. If you write down a quantum theory of a set of particles, not fields, those particles are with you forever. There's a conservation law of the number of particles.
3:25:48.0 SC: So think about how that works for a single particle. A single particle has a wave function. That's what Schrödinger teaches us. What the wave function does depends on the environment it's in, depending on how it feels the potential energy of some electrostatic field or whatever. Is the particle mostly confined to an atom, or is it propagating through empty space, et cetera? So if the particle's not confined by some potential, if it's not stuck to an atom, if it's just moving through space, generally the generic solution to the Schrödinger equation says that that particle's wave function will be very spread out. This is in contrast with what Schrödinger hoped was going to happen. He kind of hoped that the particles would naturally localize around some point, and that's why they were looking particle-like. But the actuality is that you let a particle go in empty space and you solve the Schrödinger equation and it spreads out all over the place.
3:26:43.1 SC: So in that picture, there's no relationship between how particles behave and their being localized. So you ask in the question, Brandon's question implies that, do particles have some notion of being localized in energy or density? So even before we get to quantum field theory, the answer to that is no. The fact that it's a particle says that when we observe it, it looks to be localized. Okay. When we measure its position or momentum, we get an answer. But that doesn't mean that it's localized when we're not looking at it. And then when you move to quantum field theory, the fields have different vibrations, and those vibrations can be interpreted as different collections of particles. So a single field like the electromagnetic field can be in a state where there's essentially no particles, or it can be in a state where there's essentially one photon, or a state where there's two photons or whatever, up to 10 to the 88th photons we have in our observable universe. You could even go beyond that in thought experiments. Okay?
3:27:47.7 SC: So nothing there says that the energy has to be localized around some point. Localization and particleness actually are not related to each other. It's actually a little bit simpler if you think not of photons, but think of electrons, okay, because electrons have a mass, 0.5 MeV. And that mass gets added when you have multiple particles. So this is the mass gap that we were talking about before. So the zero-particle state of the field has zero energy, if you normalize it correctly. The one-particle state has at least the mass of the electron in energy. It can have more energy if you're looking at a state where the electron is moving and has some kinetic energy. The two-particle states have at least twice the mass of the electron's energy and so forth. So that's really the criterion that tells you. It also, of course, for electrons, you also have electric charge. The zero-particle state has zero total electric charge. The one-particle state has one unit, minus one units, I should say, of electric charge and so forth. So that's what makes it a particle. The particle has a certain discrete amount of stuff: energy, charge, et cetera. Not that it's localized in space or anything like that. That's the localization in space, if it's freely moving through space, if it's not bound in some potential, that's a feature of our measurement of the particle, not of what the particle is doing when we're not looking.
3:29:19.4 SC: Joy or Joie, I'm not quite sure how to pronounce it, says, "As often heard on Mindscape, the early universe is described as very low entropy that increases with time, which is also how and why we experience time linearly. As a total physics novice who only knows about it from this podcast and other lay sources, my question is, can entropy be zero at time zero? And if so, what would that look like and mean?"
3:29:43.6 SC: We don't know, is the short answer. One issue is there's many different definitions of entropy, okay? And when you talk about time zero, we don't know what the relevant laws of physics are at time zero. We don't even know if there is a well-defined time called time zero, or whether or not that's just a choice that we make. So in a broader context, when you're just talking about physical systems like a box of gas, it absolutely can have zero entropy, okay? The entropy is, in classical statistical mechanics, it's the logarithm of the number of microscopic states that look a certain macroscopic way. So you can easily imagine macroscopic ways, macroscopic arrangements that only have one microscopic arrangement corresponding to them, and the logarithm of 1 equals 0. Therefore, the entropy of that system would be zero. It's hard to actually get there. Those are thermodynamic restrictions on what can possibly happen.
3:30:45.6 SC: But in the space of all possible configurations of the system, there are entropy zero states. For the universe, since we have quantum gravity and we have other things that we don't fully understand, we're not even sure what zero entropy would mean because we don't know what the space of states looks like. So it's conceivable that the entropy could be zero at time zero, but that's not necessarily what we have in mind. All we care about is it's much, much less than it is today.
3:31:14.0 SC: Marie Roscoe says, "With the upcoming Halloween and people talking about ghosts, haunting, and paranormal stuff, I realized that so many people find it easy to believe in ghosts, demonic possession, or the afterlife. But then the many-worlds interpretation of quantum mechanics is called counterintuitive or bonkers and much harder to believe in than ghosts, not just by people who believe in ghosts, but by other scientists, too. Why is that? What do you think?"
3:31:40.3 SC: Well, this is actually a good question because, as we've alluded to before, people have feelings. Scientists have feelings about what is plausible, what is reasonable to contemplate, et cetera. And those feelings can change with time. They're not completely built in. You can change your mind. You can open your mind to new possibilities. You can be driven by data to contemplate ideas that you might not have otherwise contemplated, et cetera. So I think a lot of physicists have very strong intuitions that push them against the ideas of many-worlds, et cetera. And that's okay. I'm not advocating getting rid of our physics intuitions. You have to go by your physics intuitions. There's really no other choice. What I advocate is two things. Number one, being open-minded about what the possibilities are. Don't dismiss a theory just because it's against your intuitions, because your intuitions might be wrong. And the other one is caring about the answer. I think that the real problem in modern physics is that too many people just dismiss the idea of understanding the foundations of quantum mechanics.
3:32:46.0 SC: So you'd have to ask the people who have trouble believing in many-worlds why they are more likely to believe in ghosts than in other universes that are predicted by the Schrödinger equation. One of the questions I get very often vis-à-vis either many-worlds or Boltzmann brains, people are like, "Why should I treat these ideas any differently than the simulation argument or being a brain in a vat or whatever, or God existing, or aliens tricking us?" And the answer is because Boltzmann brains and many-worlds are consequences of equations that we have developed for other reasons. They are predictions of theories that have been very successful for describing data.
3:33:29.9 SC: People have this very narrow and hilariously wrong idea about physics that we make measurements and we do experiments, and then all we can say is the outcome of the experiment. We can explain that theoretically, but we're not allowed to extrapolate the theory that we therefore describe, the theory that we develop in order to fit that data, beyond the data we've already collected. That's crazy. No one really actually should think that way. And the point is that things like many-worlds and things like Boltzmann brains and things like the cosmological multiverse are predicted by equations that we proposed and models and theories that we developed for other reasons. Now, you're 100% allowed to change the theory, to say, "Well, I don't like that theory." But okay, you have to change the theory. You can't just deny the conclusion of the theory but keep the theory as something you think is true. So ghosts are not like that. Ghosts are not actually predicted by any theory that we need to explain the rest of the world.
3:34:32.6 SC: Rob Gebela says, "You consider Boltzmann brains a serious problem for cosmology because they may vastly outnumber ordinary observers. Yet in Everettian quantum mechanics, observers on branches with arbitrarily small amplitudes are also real, and you appeal to Born rule weights rather than simply counting them. If Boltzmann brains themselves arise from rare quantum fluctuations, why can't we resolve the Boltzmann brain problem by applying an appropriate quantum measure that suppresses their contribution? What makes the quantum measure principled but an analogous cosmological measure illegitimate?"
3:35:04.7 SC: That's what we do. It's the same measure, right? That's the problem. The problem is that when we do the quantum counting correctly, including the Born rule weights on all these rare quantum fluctuations, we get many more Boltzmann brains than ordinary observers. And sometimes we don't do that explicitly, but it's implicitly what goes on. If you talk about the probability at some future event that a Boltzmann brain will fluctuate into existence, that's one sort of casual, everyday way of talking about the wave function of the universe has many, many different branches, and there's a certain number of branches that have this going on. But there's many branches, and in the conditions, under the conditions where Boltzmann brains appear at all, there might be very many branches, and on each branch, there's a relatively small number of Boltzmann brains, but still more Boltzmann brains than ordinary observers. That's the calculation that is done. So, of course, you have to take into account both cosmology and quantum mechanics when you make those comparisons.
3:36:06.4 SC: Daniel Bagley says, "In a recent article—sorry, a recent article in Aeon states, 'If panpsychism doesn't resonate with you, it's worth asking if there truly is a reason for everything, if the world is ultimately intelligible after all.' Do you think a belief in consciousness as an emergent thing invites the view that there is no reason for the universe to exist? I take it that reason is a function of consciousness, and panpsychists believe that the reason the universe exists is that consciousness exists. But I'm not sure if it's correct to think that a person who doesn't believe that consciousness is fundamental would also believe there is no reason for a universe that produced conscious beings to exist."
3:36:44.3 SC: Well, yeah, I do think that there is a relationship here. This is a psychological relationship, okay? This is not a necessary formal, logical connection, but I think that there's a psychological need that some people have to find reasons why things are true. And I think that this kind of psychological need comes very naturally from their experience in the world, right? I mean, William Lane Craig, who argues against a purely naturalistic view, doesn't want to believe that the universe can come into existence, and his argument that he gives himself is, "You don't think that a bicycle could just pop into existence, do you?" And so that's... It's very clear that he's using his experience of everyday life where bicycles do not pop into existence to therefore conclude that the universe can't, extending his experience of what happens on a day-to-day basis to the universe itself.
3:37:38.4 SC: And I think that's just a mistake. For one thing, your experience of the universe is not very comprehensive. You've only experienced a little part of it. For another thing, your experience with bicycles is part of a higher-level emergent description, okay? It's not applicable to the fundamental laws of nature, to the universe as a whole. So I'm someone who believes that things will ultimately bottom out in a set of brute facts, and we might be able to say that those brute facts are sort of natural or generic or unsurprising or whatever, but there cannot be either an infinite chain of explanations nor some ultimate necessary explanation like the theists want to have.
3:38:22.6 SC: So I'm totally willing to give up on the idea that there's a reason why the universe exists or even a reason why the universe is intelligible. There might be. Again, when I say give up, I don't mean I refuse to look for it. I mean that it might not be there, and I accept that it might not be there. So I'm willing to look for these kinds of reasons, but I'm also willing to say maybe they're not there. There are other people who are unwilling to say that, unwilling to entertain the idea that there's not a reason for it all, a reason why nature obeys laws, a reason for consciousness to exist. And people like that are very naturally driven to something like panpsychism. It sort of fits in with a worldview where consciousness and meaningfulness and understanding is sort of fundamental, right? It comes from something that cannot be reduced to other more physical-looking things.
3:39:22.2 SC: Mindscape is presumably the pseudonym of one of our readers here. So Mindscape says, "Do you have a description or definition of how or what you define as consciousness? I'm enjoying reading the many ways consciousness is currently described by others in various fields. I'd enjoy hearing your thinking on the subject. Do you align with a particular definition?"
3:39:42.6 SC: Nope, I do not. I do not have a definition. I do not align with a particular one. I'm open-minded about these things. As I've often tried to say, my only strong opinion about consciousness is that it is completely compatible with the laws of physics. I see essentially zero justification for thinking that we need to change the laws of physics to account for the existence of consciousness. And I'm taking laws of physics to be very broad, the nature of reality. I don't think that consciousness is in any way fundamental. What it is, yeah, I don't know. It has something to do with awareness, with experience, things like that. My only thing is I'm happy to try to explain it in terms of physically known phenomena, not going beyond that. But then outside that, I'm very, very open to what it actually is.
3:40:31.3 SC: And I would say, I guess, one more thing, which is I'm very influenced by, remember we had Stuart Bartlett on the podcast years ago talking about lyfe, L-Y-F-E? And he and Michael Wong, who was a different episode podcast guest for different reasons, really made very clear the idea that we shouldn't put too much effort into arguing about the definition of life, L-I-F-E, because life as we know it involves a number of different characteristics, right? Reproduction and thinking and metabolism and motility and various different things. And they point out you could easily imagine systems that have some of these characteristics but not others. And so what life is is not some unified thing that we're gonna define once and for all. There might be different kinds of things that have some aspects in common with life and not others.
3:41:24.8 SC: So I think consciousness is probably like that. I think that it's a mistake to think, "Well, we've used the word consciousness for many centuries now. There must be a single unified concept out there in the world that maps onto it." I think we can figure out by doing science, by doing philosophy, et cetera., what is happening in people, in their brains, in their behaviors, in their thoughts, and figure out what is the best way to describe that. And the word consciousness might or might not turn out to be part of that best description.
3:41:54.3 SC: Jonathan Good says, "Could two branches on the path from a low-entropy past contain different conscious experiences but still become one branch in the future if the differences were never recorded in a way that became entangled with current observers' degrees of freedom in a way that produces lasting decoherent differences?"
3:42:12.8 SC: So it is possible for different branches from a low-entropy past to eventually join back into a single branch. In fact, that is what you expect to happen when the universe reaches thermal equilibrium 10 to the 100 years from now. Okay, that's nowhere close. We're not anywhere near then. So Jonathan's question is a much shorter-term question. Could two branches of the wave function have the same observer having different conscious experiences, but those experiences don't get recorded, they don't sort of get entangled with the outside world, et cetera., and so the observer sort of comes back together?
3:42:54.5 SC: The rough answer is no. We expect that kind of thing, anything is possible, okay, but we don't expect that kind of thing to happen precisely because the way that branching happens is it's irreversible. You become entangled with the environment. That's what it means to be branching, is you have become entangled with the environment. And even if you don't write down the conscious experience you had on some notepad or whatever, it is effectively recorded in the environment. And this is what people like Zurek have talked about a lot with pointer states and decoherence and einselection and quantum Darwinism. There's a whole bunch of words that are attached to this way of thinking. So basically the definition of what it means to branch means that it goes one way and not the other way. And therefore you could imagine a quantum superposition that sort of occurs and then unoccurs, but not one that actually branches into two classical-looking things and then rejoins. The environment is just too big and the environment secretly is recording all of these different possibilities.
3:44:01.1 SC: Dan Butler says, "Priority question. Can you talk a bit about Lorentz invariance in emergent spacetime? Is it a challenge to come up with quantum theories where Lorentz invariance emerges rather than is just assumed? Is it a helpful constraint on theories of emergent spacetime or not really?" Well, I think this is a super important question. It's one that I don't know a lot about. In other words, so far I've been playing around with emergent spacetimes and their properties, but I've just been crossing my fingers and hoping that Lorentz invariance basically emerges. There's arguments back and forth about whether Lorentz invariance could be exact. That depends on the dimensionality of Hilbert space. Naively, Hilbert space has to be infinite-dimensional for Lorentz invariance to be exact. Lorentz invariance, sorry, for those of you who don't know, is just the statement that, as we already alluded to earlier, the universe doesn't care how fast you're moving or what direction you're pointing in. There's a symmetry in spacetime called Lorentz invariance.
3:45:00.6 SC: And if you build a theory of emergent spacetime by starting with a quantum state evolving with time, you're already picking out time as different in the formulation of the theory. That doesn't mean that the theory that you build can't be perfectly Lorentz invariant. After all, you could start with ordinary QED, quantum electrodynamics, right? That's a perfectly 100% Lorentz invariant theory. You can pick a coordinate system or a Lorentz frame, as we call it, and you can quantize the theory in that frame. So even though you pick a coordinate system to use to write your theory down, doesn't mean that the physics that you end up with can't be perfectly Lorentz invariant. However, there's also the idea that maybe physics has a finite-dimensional Hilbert space, and then it might be harder, but I actually haven't really looked at it very carefully. I've sort of thought about looking at it, but never gotten around to doing it. I do think that one of the reasons it's worth doing is because we have experimental constraints on violations of Lorentz invariance. So I think that figuring out whether quantum gravity implies interesting experimental deviations from perfect Lorentz invariance would be an extremely interesting thing to do.
3:46:13.5 SC: Moshe Feder says, "In the late '90s, Randall and Sundrum proposed two variations of a five-dimensional anti-de Sitter space model to solve the hierarchy problem of gravity's weakness relative to the other forces. What did you think of their proposals at the time, and has any further work been done on them since?"
3:46:31.7 SC: Yeah, so back in the '90s, there was quite a bit of excitement about extra dimensions of space and how they could be curled up. That was an old idea, of course, but in the '90s, people started getting more interesting constructions that they could invent, basically taking advantage of the idea of branes, B-R-A-N-E-S, not Boltzmann brains. These are D-branes, typically, from string theory. And the thing about the branes is that you can easily confine all of particle physics to the location of a brane. So if you have, let's say you have five-dimensional space, but embedded in that five-dimensional space, you have a three-dimensional brane. Okay? And that's space, not spacetime. I'm giving you the dimensions of space. You could confine all the particles and all the fields that we know about to the brane, so that as far as physics was concerned, the people living on that brane would think that space is three-dimensional.
3:47:30.2 SC: The problem with that is that you can't easily confine gravity to the brane. Gravity tends to leak out. So this went in two directions once people realized this. One direction was large extra dimensions, where you could say, okay, I let gravity leak out into the brane, but gravity is weak and I don't have very good constraints on it. Therefore, maybe the extra dimensions are bigger than you think. And the most popular models had two extra dimensions that were almost as big as a millimeter in size, basically because that's where... Two reasons. Number one, that's where the experimental limits were. And number two, the numerology worked well with reconciling a Planck scale at very high energy with an electroweak scale at very low energy. You're not solving the hierarchy problem, really, but you're replacing the question of why is the Higgs mass so small with a question of why are the extra dimensions so big?
3:48:24.9 SC: The other direction that it went in is this Randall-Sundrum direction, where rather than two extra dimensions with flat space inside, they had one extra dimension with anti-de Sitter space inside, so a negatively curved spacetime. That plays very well with string theory. String theory is very happy with a negative cosmological constant, which is what you have in anti-de Sitter space. But the Randall-Sundrum construction wasn't specifically for string theory. You could have imagined it in a more general context. And they got a different kind of solution to the hierarchy problem. They related physics at one end of a very, very small extra dimension to physics at the other end, and the energy scales get changed with respect to each other because of the curvature of space in between.
3:49:11.1 SC: So this turned out to be a very phenomenologically interesting set of models, and people developed attempts to show how you could detect the existence of this extra dimension, how it would affect the spectrum of particles you might observe at the LHC, and so forth. As you know, we haven't detected any new particles at the LHC, and we haven't detected any violations of the way that gravity behaves or anything like that. So I think that it's still viable as an approach to phenomenology, but right now we're in a phase where it's just guesswork, right? We don't have any very strong phenomenological clues as to what direction to go in in terms of going beyond the Standard Model of particle physics. So you can guess things and work on them and hope they work out, and Randall-Sundrum is still very much in that bucket, I would say.
3:50:01.2 SC: Nate D. Says, "I'm a bit confused about the relevance of the various interpretations of quantum mechanics for experimental physics. Are the formalisms underneath all the interpretations so consistent as to make one's theoretical preference as an experimentalist, let's say for Many-Worlds over Pilot Wave, irrelevant? Or would believing in Many-Worlds over Pilot Wave potentially generate different experimental arrangements? Basically, this is just a question about the relationship between theory and practice in contemporary quantum mechanics."
3:50:29.8 SC: Well, we have very good limits and experimental constraints on how quantum mechanics works. So basically, if you're coming up with a new formulation of quantum mechanics, they used to be called interpretations in the old days, but I don't think that's a good way to call them; they're physically different theories, and if you're coming up with a new one, there's two options. One is you basically reproduce all of the known experimental successes of quantum mechanics, or you don't. And if you don't, you throw that theory away and move on to think about something else. So almost by construction, all of the various theories that we look at today have more or less the same predictions for quantum mechanical tests that you can actually do in the lab.
3:51:11.6 SC: That might not always be true. The theories that have objective collapse of the wave function absolutely make special new predictions for experiments, and we're trying to do them, and we haven't really completely ruled out those theories yet. I suspect they will eventually be completely ruled out. For other theories, for Copenhagen, for Pilot Wave, et cetera... I say this over and over again. I'm not sure if it really catches on with people, but the problem is that the theories are not well defined. It's perfectly clear that Copenhagen is not well defined because it doesn't tell you what a measurement is. Even pilot wave and hidden variable theories are not well defined because in modern physics we have what is called quantum field theory. And there's no consensus about how to do a version of hidden variable theories in the presence of quantum fields rather than particles. We had Tim Maudlin on the show and he talked about it a little bit, and he is someone who is optimistic about those theories, so he put an optimistic spin on it. But at the end of the day, we don't know how to do that.
3:52:12.6 SC: And hidden variable theories work very well with non-relativistic point particle quantum mechanics. They don't work very well with quantum field theories. People have tried and people have made progress in some sense, but there is no settled way of doing it. So David Wallace makes this point very firmly. He says, "Look, there's nothing to compare it to because we don't know what a hidden variable theory would say about the internal structure of a proton or something like that." So it's just very hard if you haven't developed the theory to ask what its experimental predictions are going to be.
3:52:51.3 SC: Sonja says, "What are your thoughts on the recent elections in Serbia and the key role played by university students and members of the academic community?"
3:53:00.8 SC: I have no informed thoughts about the recent elections in Serbia. The reason why I am answering the question is because I just want to, number one, encourage people in Serbia and elsewhere to keep up the good fight for liberal democracy in the world. And number two, to acknowledge that university students and members of the academic community worldwide have always played important roles in social movements. It's just an interesting fact about human life that that part of your life when you're a university student, and the students get more credit than the faculty do here, no question about that, that's a part of your life when you're exposed to new ideas, but you still have the energy and the initiative to go out and do something about them, not just to talk about them. So I'm not surprised that university students play a big role in this, and I think that that will always be the case and they should be encouraged to be allowed to do that.
3:53:54.4 SC: Benjamin Zand says, "In general relativity, spacetime has a definite curvature at every point described by the Riemann tensor. If we measured a traveler's speed relative to this curved geometry with each point having its own curvature, wouldn't that give us an absolute standard of motion? Or do the components of the curvature tensor change with the observer's velocity in a way that keeps the principle of relativity intact? If the latter, where and how does observer's velocity enter into the equations?"
3:54:22.3 SC: Well, look, you have to distinguish between two different things. One is the fundamental laws of physics and the other is how things behave in a particular situation, in a particular state. After all, here on the surface of the Earth, I can tell the difference between up and down because the Earth is beneath my feet. Up is the opposite direction. Okay? That doesn't mean that Lorentz invariance has been violated or that there is an absolute standard of rest in the underlying laws of physics. There's a standard of rest here on the Earth because I can talk about my speed relative to the Earth. Talking about the curvature tensor or any other tensor field, scalar fields, vector fields, whatever, is exactly the same way. If you have some particular configuration of matter and spacetime and whatever, you will have different ways that that configuration will be observed depending on how you're moving through space. You might be able to move through spacetime. You may be able to move in such a way that the values of the tensors remain approximately constant or something like that. And if you wanted to, you could say, "Well, that's a preferred frame of reference with respect to that field that I'm living in. But it's not a preferred frame of reference in the fundamental laws of physics."
3:55:38.7 SC: In exactly the same way, the cosmic microwave background provides a rest frame. You can say, "I'm at rest with respect to the microwave background photons," or "I am blue-shifted in one direction, red-shifted in the other direction." But that's not fundamental to the laws of physics, so we don't count it as a preferred frame of reference at the deepest level.
3:55:57.3 SC: Jake asks a priority question. "In a previous AMA, you argued that the morality of painlessly killing hypothetical kittens depends on the reason, and that killing them purely for enjoyment would be morally wrong. Would you also consider it morally wrong to kill an animal for a tastier meal when a nutritionally equivalent plant-based alternative is available, since in both cases the reason for killing is enjoyment? I understand why taking pleasure in killing might be morally objectionable, but does the seemingly innocent pleasure of eating justify the killing itself?"
3:56:28.6 SC: Well, I think that there is a difference here. So I'm not gonna tell you what the right moral stance is, but in terms of what the considerations are that go into your moral stance, I do think that it's not a good analogy between killing kittens for the thrill of killing kittens and killing animals to eat them, to eat their food. Namely, that where the pleasure is coming from is in the act of killing versus the act of eating a meal. And I think that those are different. You might not think that they're sufficiently different, and you might argue that I think that the act of killing animals to have a meal from them is not morally right. I get that. That's a move you can make. But you can't say, "Well, it's just the same as strangling kittens just to get your jollies from it."
3:57:13.9 SC: Okay, this is an anonymous question. "People at AI labs appear to be freaking out that AI research looks like Go or theorem proving. AIs suck at most things, but in the narrow domains where AIs can get fast and objectively true signal on their performance, they won or lost the Go game, et cetera., AIs can run an ungodly number of trials and teach themselves to outclass humans through brute force. The bottlenecks of AI research seem to have fast and objectively true signal, like the prediction loss went up or down, and the labs are now spending billions per quarter trying to train AIs to do AI research. Should we be freaked out?"
3:57:50.5 SC: Well, I don't think that anything you said leads me to be freaked out yet. I mean, maybe we should be freaked out for various reasons, but the fact that research is going quickly by itself does not count as a reason to be freaked out. Again, I am a believer in being responsible, both at the level of the companies doing the work and at the level of the government regulating them. I do think that being careful with things we don't understand and moving slowly and testing things and making sure they're as safe as possible all makes sense. I think that safeguarding ourselves against a million minor problems is the best way to safeguard ourselves from one big problem. But the very fact that research is going quickly is not by itself a reason to be freaked out, unless you're just really upset that research is going faster than you can keep up with, which I get that too, but I don't think it's any reason to really worry at a deep level.
3:58:47.5 SC: Dennis Briggs says, "I tell my students that the linear systems don't know when enough is enough. Is Many-Worlds the ultimate example of that aphorism, or am I missing something?"
3:59:00.0 SC: So for those of you who are not big math fans out there, I think what Dennis is referring to is that a linear system is a system that just evolves in time in a linear fashion. And what that means is it never stops, right? Not only does it go at the same rate, it goes at the same rate forever. So if you had linear growth of some system, it would just keep growing forever and ever and ever. This is actually a well-known issue when people try to model mathematically things like the growth of species or environmental issues and things like that, because very often you start with some simple model that does well describing some tiny period of time, but the model just grows. It just sort of blows up and is unstable and doesn't really do a good job. So you invent non-linear effects that say, "Okay, you can grow for a little while, but if you grow too far, then a new term kicks in which kills you off and stops you from growing." That's a very, very standard move in ecological or biological modeling.
4:00:02.7 SC: So Dennis's question is, is Many-Worlds the ultimate example of this? And maybe. I think that Many-Worlds is the consequence of taking the Schrödinger equation literally at face value. And the Schrödinger equation is linear. So in that sense, yes. There is a further tricky thing, though. It does depend on the space in which you are evolving. After all, if your space of possibilities is a circle, then even if you move on the circle linearly, that is to say at a constant angular velocity, you don't go crazy. You just keep coming back to where you left again and again and again, right? Just go around the circle an infinite number of times. Nothing to be worried about there. If you're moving through a plane, a two-dimensional Euclidean plane, and you're moving in a straight line, then you will go off forever.
4:00:57.2 SC: So this is analogous in quantum mechanics to asking whether or not Hilbert space is finite-dimensional or infinite-dimensional. That's why in my recent paper, you got a cyclic universe out of a situation where you were in a finite-dimensional Hilbert space, because then you evolve linearly, but you come back to where you left. That's where the cyclic universe comes from. If Hilbert space is infinite-dimensional, then you can evolve forever and ever. And Many-Worlds is basically taking advantage of that. Absolutely.
4:01:27.0 SC: Okay. The final question of this month's AMA comes from George Hampton, who says, "Given the state of our democracy, I've been thinking about Gödel's loophole, and I'm wondering if this is the inner inconsistency he was thinking about. Do you have any thoughts on this?"
4:01:43.9 SC: So I have some thoughts. I'm not sure how accurate my thoughts are. So for those of you who don't know, there's a famous story of Kurt Gödel, the logician, who moved to the United States. He was from, I think he was from Germany, but maybe Austria, I'm not sure. Anyway, he got a job at the Institute for Advanced Study in Princeton, he became friends with Einstein, et cetera. And eventually he applied for citizenship in the United States. And there's a test you have to take to become a citizen. And one of the questions is something about democracy in the United States. And according to the story, there's different versions of it, I don't really believe any of them, but according to the story, Gödel starts saying, "Well, actually there is a way within the Constitution to overthrow the government of the United States and make it non-democratic."
4:02:29.9 SC: And some people try to wonder whether or not this is related to Gödel's formal work, his logical work in showing that there are inconsistencies that lay, possible inconsistencies or at least possible paradoxes in logical systems. But as far as I know, nobody actually knows, maybe someone does, but hasn't told me, what Gödel actually had in mind when he was thinking about this. And honestly, I know people speculate about it. I don't quite see what the big deal is. I think that this is, it's basically impossible to design a democratic system that can't be overturned for the very simple reason that if all of the people in a democracy decide to get rid of democracy, they can do it, right? That's not something you can even stop from happening no matter how you design your system. So I don't think you need to be a super-genius like Gödel was to notice that in the Constitution of the United States there are ways to amend the Constitution. And you could amend the Constitution to say that there is a dictator, right? There's an emperor, if you wanted to do that. That's 100% valid. You don't need to be very smart to figure that out.
4:03:43.3 SC: So this is something you just have to live with. I mean, this is why it's crucially important if you do want to live in a democracy, which we've been talking about before, that the populace be educated and civic-minded. You can't have a democracy of reluctant citizens or citizens that don't believe in the values of democracy, because if they don't believe in those values, they can easily get rid of the democracy. You can't force democracy on people who don't want it, no matter how hard you try. So you have to give the sales pitch for democracy. You have to let people know how important it is. And democracy is not just about giving people a voice. That's part of it. But oftentimes people say, "Well, I like democracy because I like to be able to vote. I like to have a say in the government." And what they're less enthusiastic about is other people having a say.
4:04:37.2 SC: What democracy is really about is not voting for your favorite thing, but acquiescing when the majority of people in your country vote for something different, for agreeing to go along with majority rule. And of course you want to safeguard against violating the rights of minorities, et cetera. That's why you have various checks and balances in the Constitution, and you have a republican system, and you have federalism and all these different things, and you have a Bill of Rights. Okay. So you don't want the majority to get everything at once, but for the most part you want majority rule, which means that sometimes you're gonna be in the minority and you have to learn to live with that. And I think that that is something that we don't do a good job of explaining to people so that they are willing to fight for the principle that sometimes they're not gonna get their way.
4:05:26.6 SC: Because when people fantasize about how awesome it would be to have a strongman, dictator, authoritarian, in their minds, their strongman is always on their side. And they just don't quite accept that maybe even if the strongman initially looked like they were gonna go along with their side, they can change their minds, their incentive structure might be different, and therefore it is better to just avoid the possibility of anyone having that authoritative power at all. So I don't know what Kurt was thinking about, but I don't think you need to think that hard. We do need to make sure that the democracy is solid at the level of the opinions of the actual citizens of the democracy, not just at the level of the laws that we're all supposed to be obeying. So hopefully that thought is, you know, both important and optimistic. I don't know whether it's optimistic or not. Democracy trundles on. It's gonna live to fight another day anyway. Thanks as always for supporting the Mindscape podcast. I will talk to you next time.