What this covers

This monthly Ask Me Anything episode features Sean Carroll fielding questions from Patreon supporters across physics, philosophy, and adjacent fields. The range spans foundational issues in cosmology and quantum mechanics—the status of the Big Bang, the nature of particles and fields, the emergence of spacetime itself—alongside questions about consciousness, personal identity, morality, and how to evaluate speculative hypotheses. Carroll's approach throughout treats the deepest puzzles as resolvable through rigorous analysis of what fundamental physics actually says, rather than appeal to intuition or cultural authority.

The conversation covers several distinct arenas. On quantum mechanics, Carroll defends the Many-Worlds Interpretation and explains how decoherence rather than measurement collapses the wave function, and clarifies what particles actually are within quantum field theory. On cosmology, he argues the Big Bang is not a temporal event but a boundary condition where classical general relativity breaks down, yet the smooth low-entropy early universe remains a genuine puzzle worth explaining rather than dismissing as brute fact. On emergence and locality, he develops the case that spatial locality and causality are emergent approximations from Hilbert space evolution, and that strong emergence—a system behaving differently in different contexts despite identical local physics—contradicts the well-tested core theory. He also addresses near-death experiences as a phenomenon of neural filtering coming offline, applies Bayesian Reasoning to reject the simulation hypothesis on empirical grounds rather than aesthetic ones, and argues personal identity does not require temporal continuity. Throughout, Carroll emphasizes the importance of explicit mechanisms and empirical grounding over mathematical elegance when evaluating claims about fundamental reality.

Sharpest takeaway

Carroll argues that the deepest questions in physics and philosophy are best addressed by treating concepts like emergence, locality, causality, and morality as constructed or emergent frameworks grounded in well-tested fundamental physics, rather than as mysterious primitives or objectively given truths.

  • Locality and causality are emergent approximations from an underlying wave function, not fundamental primitives
  • Morality is constructed from systematized intuitions, not discovered as objective fact, yet remains action-guiding
  • Speculative claims (strong emergence, simulation, octonions) must be evaluated by Bayesian reasoning and explicit mechanisms, not vibes

The argument · threads69 threads · 69 claims
01

Good definitions should apply to both simple and complicated situations so understanding simple cases helps illuminate complex ones, not exclude simple cases by counting emergence only when mysterious.

1 point
  • Many definitions of emergence are constructed to leave out all the examples we already understand, counting emergence only when it is mysterious; this is backwards, because a good definition should apply to both simple and complicated situations so that our understanding of the simple cases helps us gain insight into the complex ones we don't yet understand.

    “there are some people who have definitions of emergence that, by construction, leave out all of the examples that we understand, so they only want emergence to count when it's mysterious”

02

Democratic voting rights should extend to all responsible adults because they have interests, values, and desires deserving fair representation, not be restricted to those earning franchise through public service.

1 point
  • The real moral argument for democracy is not that voters are smarter or better educated or have done more service, but that people who live in a country have interests, values, and desires that deserve to be reflected fairly; therefore everyone above a certain age who is a responsible adult should have the right to vote, and earning the franchise through public service (as in Starship Troopers) is not a good system.

    “people who live in a country have an interest in it, they have desires, they have values, they have things that they want to be their own interest to be protected, and so therefore they have a right to vote. That's the argument for me, it's not that they're smarter or better educated”

03

Intelligence is not a single definable trait perfectly correlated with physics success; mathematical ability and physical insight are distinct, and different people excel at physics in different ways.

1 point
  • Intelligence is not a single definable thing and is not perfectly correlated with success in physics; mathematical ability and physical insight are distinct skills (Einstein had unmatched physical insight despite being a weaker mathematician than Poincaré), and there are many different ways to be good at physics, so the better question is whether you are good at doing physics, not how intelligent you are.

    “Albert Einstein was not as good at math as Henri Poincaré, but he did better at understanding relativity 'cause his physical insight was completely unmatched.”

04

Many-worlds constrains the Big Bang by requiring low branching and low entropy near it since branching is irreversible toward the future, and these conditions are probably identical.

1 point
  • Many-worlds does not say anything special about the Big Bang, but the Big Bang constrains many-worlds: because branching is irreversible and happens toward the future, the theory requires a quantum state near the Big Bang with a small number of branches and low entropy, which are related but not identical statements, and these low-branch and low-entropy conditions are probably exactly the same thing.

    “what you need to set up the whole theory and make it work is a quantum state near the Big Bang where there weren't that many branches, it was a small number of branches in the way that we would normally think of them, and branching happens toward the future.”

05

Weak emergence exists when a macro theory's coarse-graining commutes with time evolution, preserving reliable predictions while discarding microscopic information to capture real patterns.

1 point
  • Weak emergence exists when there is a useful macro theory in which coarse-graining commutes with time evolution: you can either evolve a microstate forward and then coarse-grain it, or coarse-grain it first and then evolve the resulting macrostate under the macro theory, and arrive at the same place; this captures a 'real pattern' that lets you discard almost all microscopic information while retaining reliable, possibly probabilistic, predictions.

    “the diagram commutes, you get to the same place by either one of those two routes, coarse-graining commutes with time evolution. Okay? I think that's the most important aspect of emergence.”

06

A universe with no life forms can exist without contradiction, just as uninhabited places like the sun's center exist; existence requires no observation or measurement.

1 point
  • A universe with no life forms can exist without contradiction, just as plenty of places in our own universe exist that no life will ever witness, such as the center of the sun; existence requires no observation or measurement to obtain.

    “the center of the sun will never be looked at up close and personal, but it still exists.”

07

In quantum field theory a particle is not an interaction location but a field state; a single field describes zero, one, two, or more particles via different energy levels.

1 point
  • In quantum field theory a particle is not an interaction location but a state of the field: the lowest-energy state is the vacuum, a set of higher-energy states with a fixed rest-frame energy constitute one-particle states (that fixed energy being the particle's mass), the next set constitute two-particle states, and so a single quantum field simultaneously describes zero, one, two, or more particles.

    “The single quantum field describes simultaneously the possibility of zero particles, one particle, two particles all the way up”

08

The simulation hypothesis can be dismissed by logical analysis rather than dismissing it as sounding like God; treated as a Bayesian hypothesis, the absence of simulation-like features in our world counts against it.

1 point
  • The simulation hypothesis can be dismissed only by counter-arguments or logical flaws, not by saying it 'sounds like God'; treated as a proper Bayesian hypothesis, if we did live in a simulation the world would likely look different, so the absence of any simulation-like features in our world counts against it, just as theism predicts a world that would look different from the naturalistic one we observe.

    “I find no evidence that there's anything gained by imagining that we're living in a simulation, and in fact, I think I could easily argue that if we did live in a simulation, the world would look different, and so as a good Bayesian, I count that against the argument”

09

Spatial locality emerges from the wave function as abstract high-dimensional mathematics becomes equivalent to local description; locality requires chopping the wave function into subsystems interacting only with nearest neighbors.

1 point
  • Spatial locality emerges from the wave function the way three-dimensional billiard-ball positions emerge from a single point in a 6N-dimensional phase space: the abstract high-dimensional description and the local description are mathematically equivalent, but the local one provides intuition and counterfactual reasoning, so locality means you can chop the wave function into subsystems that interact only with nearest neighbors, and this requires the Hamiltonian/Schrödinger evolution, not the static wave function alone.

    “if there is a spatial locality, then you're gonna be able to chop up your Wave Function into different subsets that interact with their nearest neighbours, as you interpret it after the fact, and not interact with regions far away.”

10

Requiring undergraduate students to address professors formally as 'doctor' or 'professor' better supports those facing power differentials from age, appearance, race, and gender than using first names.

1 point
  • Asking undergraduate students to use a professor's first name has unanticipated consequences because power differentials and respect are distributed unequally across professors by age, looks, race, and gender, so aging white male professors face the least difficulty being respected; it is therefore better for everyone to address undergraduate professors as 'doctor' or 'professor' to support those who must work harder to establish authority.

    “some people need to work harder to get the respect they deserve as professors than others do. And aging, white dudes like myself, need to work the least hard”

11

Branching of the wave function occurs only during decoherence, meaning entanglement with the macroscopic environment; measurement is just decoherence in many-worlds, not something mysterious.

1 point
  • Two fields entangling, such as a single electron and photon interacting, does not constitute branching of the wave function; branching happens only when decoherence occurs, meaning entanglement with the macroscopic environment, so it is only when we measure with a many-part detector that decoherence and hence branching actually take place, and measurement is just a fancy word for decoherence in many-worlds.

    “a single electron and a single photon interacting with each other and entangling doesn't count as branching of the wave function. If we then measured the position... then decoherence would happen because measuring in that sense means I've detected it in a macroscopic detector that has a lot of moving parts.”

12

Intuitions can be developed and trained through practice in both physical skills and science, contrary to Kantian views treating intuition as innate from birth.

1 point
  • Contrary to a Kantian view that treats intuition as fixed from birth, intuitions can be developed and trained, both in physical skills like hitting a baseball (system one rather than the conscious system two) and in science, where learning relativity, quantum mechanics, or biology lets you develop a genuine intuition for the subject.

    “I don't believe that intuition is just something that appears instantly when we're born and never changes, I think that intuitions can be developed.”

13

Labeled nothingness is contradictory, and whether something could exist rather than nothing is probably an ill-formed question arising from linguistic construction failure rather than deep reality.

1 point
  • Something labeled as nothingness cannot exist (a contradiction in terms), and the deeper question of whether there could be nothing rather than something is probably ill-formed because the word 'can' implicitly compares possible actual worlds, and a world in which nothing exists is not a world at all; this is likely tomfoolery from a failure of linguistic construction rather than deep insight into reality.

    “I think this is just Tomfoolery caused by a failure of our ability to linguistically construct the question accurately rather than some deep insight into nature of reality.”

14

Complexity is usefully defined negatively as the regime where neither physics' simplicity forms apply: too many individual moving parts that still matter, yielding unpredictability and non-trivial distributions.

1 point
  • Complexity can be usefully defined negatively as the regime where neither of physics' two forms of simplicity applies: you neither have a small number of moving parts you can track individually, nor a large number whose individual idiosyncrasies smooth out and can be averaged into collective variables; in complex systems many things go on and the individual things still matter, giving rise to unpredictability, power-law behavior, and non-trivial spread distributions.

    “it's when neither of those two things happen, when you neither have just a small number of particles nor do you have a large number for which all of the individual idiosyncrasies smooth out and can be averaged over”

15

Evolution finds good-enough solutions rather than truly optimal ones since biological space is combinatorially immense and environments constantly change; convergent evolution shows different trajectories can reach similar morphologies.

1 point
  • Because biological space is combinatorially immense, evolution will never find a truly optimal solution and the notion of one may not even be sensible given constantly changing environments and competition; what matters is good-enough solutions, and convergent evolution means very different trajectories through genome and environment space can arrive at more or less the same morphology, not exactly the same and not perfectly optimal.

    “it's not really a matter of the super-duper exact optimal solution, it's a matter of a pretty good solution, a good enough solution, which is a better sort of characterization than optimal.”

16

Modern science already works well by claiming priority through public results, incentivizing maximal sharing; no special need exists for NFTs or blockchain in science and publishing.

1 point
  • Carroll sees no special need for NFTs or blockchain in science or publishing because modern science already works well by claiming results through making them public, where whoever puts a result out first gets priority, which incentivizes maximal sharing; this is the opposite of older practices of hiding results in codes or anagrams, though other economically or politically sensitive fields might differ.

    “whoever puts the result out there first gets priority for it. I think it's a pretty good system actually, because it incentivizes sharing the scientific results as much as possible, which is good for science.”

17

Great individual effort at earning income would take roughly 400 years to become a billionaire even with substantial support, showing great fortunes require systematic mechanisms, not just individual effort.

1 point
  • Even with 100,000 Patreon supporters at a dollar per episode, after taxes and Patreon's cut yielding about $50,000 per week, it would take roughly 400 years to become a billionaire, illustrating that great fortunes are not made by individual effort alone but require a system in place to funnel money to single people.

    “it would take me about 400 years to become a billionaire. So this is not the way to become a billionaire, to have a podcast here and support it by Patreon”

18

There is no metaphysical 'you-ness' persisting through time; you are a pattern in atoms, so a reconstructed clone with all memories is the same person and equally accountable despite a gap in lifespan.

1 point
  • There is no metaphysical 'you-ness' that persists through time; you are a pattern in a collection of atoms (most of which differ from your infancy), so a reconstructed clone with all of Alice's memories, indistinguishable in every way, is genuinely the same person despite a gap in her lifespan, and is just as accountable for the original Alice's crimes, because there is no rule requiring temporal continuity for personal identity.

    “there is no such thing metaphysically as the you-ness of your identity that persists through time and follows you from one moment to another. You, what we think of as you, is you are a pattern in the collection of atoms and molecules”

19

The early 2000s new atheist movement had both good repercussions like making atheism respectable and bad ones like turning off some people; Carroll guesses the overall effect was good without firm evidence.

1 point
  • The new atheist movement of the early 2000s had both good repercussions (atheism became a respected position that must be confronted rather than dismissed as in the 1980s) and bad ones (some people were turned off by the derogatory, debate-driven attitude), and while the relative magnitudes cannot be measured, Carroll's guess is the overall effect was good though he lacks evidence.

    “atheism is much more on the table these days than it ever was as a point of view that needs to be taken seriously and respected... in the '80s, people just wouldn't take atheism seriously at all.”

20

The Bekenstein bound limits the number of basis states in Hilbert space, not the continuous information specifying which linear combination a system is in; finite measurement resolution prevents access to continuous information.

1 point
  • The apparent conflict between the finite Bekenstein bound and real numbers requiring infinite information is resolved by recognizing the bound limits the number of basis states (the dimensionality of the Hilbert space), not the continuous information needed to specify which linear combination of those basis states a system is in; we never actually have that continuum of information because all measurements have finite resolution.

    “the thing that is being limited by the Bekenstein bound is the number of basis states, not the amount of information required to fully specify the linear combination of such basis states that would make you an arbitrary quantum state”

21

Treating something as a brute fact correctly recognizes it as one proposition with non-zero credence; the universe's initial conditions are probably less plausibly brute facts than alternatives like wave function of the universe.

1 point
  • Treating something as a brute fact is not giving up but correctly recognizing it as one proposition among many with non-zero credence; for the universe's initial conditions, a brute-fact explanation is probably less plausible than alternatives like the wave function of the universe or inflation, but for the existence of the universe itself, attributing it to a brute fact is the most plausible option, and good Bayesians weigh all such options as investigation proceeds.

    “it's not giving up. It's just correctly recognizing that sometimes that is the most plausible looking thing.”

22

Our understanding of nature's laws in tested regimes is better than it deserves to be; Carroll sees zero obstacle to understanding all laws of nature and is unmoved by vague arguments about fundamental comprehension limits.

1 point
  • Our understanding of nature's laws in tested regimes is better than it has any right to be, which is good news for human achievement but bad news for physicists seeking improvement; while this is not proof that we can understand all laws of nature, Carroll sees absolutely zero obstacle to doing so and is unmoved by vague arguments that we may be fundamentally unable to grasp them.

    “our understanding of the laws of nature in the regimes where we really tested things is honestly way better than has any right to be.”

23

Physicists' tradition of falling in love with mathematical constructions and forcing them into physics never works; actual advances run the opposite direction, starting from physics and forcing it into higher mathematics.

1 point
  • There is a long tradition of physicists falling in love with mathematical constructions and trying to shoehorn them into physics, and as far as Carroll knows this never works; the important advances run the other way, starting from physics and being forced into higher mathematics (Einstein into differential geometry, Heisenberg into matrices, gauge theory into homotopy), so hoping octonions or quaternions will yield deep insight into particle physics is backwards relative to historical progress.

    “there's a long tradition in physics of people falling in love with mathematical constructions and operations and trying to shoehorn them into physics. And as far as I know, it literally never works.”

24

Asking where a human's energy goes after death is like asking where a candle's energy goes when extinguished; life is a process converting energy between forms, so energy remains constant across death.

1 point
  • Asking what happens to a human's energy after death is like asking where a candle's energy goes when it stops burning: nothing goes anywhere, life is a process (like combustion) that converts energy from one form to another, and when the process stops the energy remains, so you have exactly the same amount of energy the moment after death as the moment before.

    “it's still there, you have exactly the same amount of energy the moment after you die as the moment before.”

25

Placing the northern hemisphere at the top of maps has minor justified basis since far more people live there; this is a minor implicit bias compared to others worthy of attention given finite resources.

1 point
  • Putting the northern hemisphere at the top of maps may privilege it in a minor way, but it is not a completely arbitrary choice because far more people live in the northern hemisphere; among the many implicit biases people hold, this is one of the more minor ones, and we must pick our battles as finite creatures.

    “a lot more people live in northern hemisphere of the Earth than the southern hemisphere of the Earth, so it's not a completely arbitrary choice.”

26

Over-emphasizing perfection and exactness distorts understanding; the world is not perfect or exact, and relaxing demands for perfection changes impressions of what is possible and what happens.

1 point
  • Many opinions in both philosophy and science suffer from over-emphasizing perfection and exactness; the world is not perfect or exact, and when you relax the demand for perfection a little, your impression of what is possible and what actually happens can change by a lot.

    “I think a lot of questions in both philosophy and science suffer a lot of opinions or ideas, suffer from over-emphasising perfection and exactness”

27

James Webb Space Telescope observes galaxies roughly a billion years after the Big Bang, not close to it; the cosmic microwave background gives evidence from 380,000 years after and nucleosynthesis from minutes after.

1 point
  • Contrary to popular framing, the James Webb Space Telescope is not looking very close to the Big Bang; it observes galaxies roughly a billion years (at best hundreds of millions of years) after the Big Bang, while the cosmic microwave background gives detailed evidence from 380,000 years after and nucleosynthesis from minutes after, with the intervening 'dark ages' being very hard to observe because little light was given off.

    “JWST is looking at things maybe that were happening a billion years after the Big Bang. Okay. Maybe we can go to hundreds of millions of years if we're really, really lucky, but anyway, nowhere close to the Big Bang”

28

Weak effective altruism is sensible common sense; the strong version of optimizing all life choices for world utility is unrealistic, and Joshua Greene's mixed strategy respecting both personal and impact-optimized causes is better.

1 point
  • The weak version of effective altruism (deploying altruism more effectively) is sensible common sense, but the strong version of running your life to optimize world utility is unrealistic and never actually practiced; the best approach is Joshua Greene's mixed strategy allowing people to give both to a personally meaningful cause (the local cat shelter) and to an impact-optimized cause, which respects real human motivations.

    “I kinda like the version that was promulgated by Joshua Green when we had him on the podcast, where there's sort of a mixed strategy, where you allow people or you encourage people to both give money to their favourite charity to save their local kittens at the cat shelter, and also something that has been optimised for saving the most lives”

29

Local quantum field theory underlies everyday life, making strong emergence incompatible with locality; strong emergentists must show exactly how to modify well-tested equations rather than merely asserting locality could be broken.

1 point
  • The core theory underlying everyday life is a local quantum field theory, meaning quantum fields interact only at the same point in space and an electron responds only to other fields at its own location, so strong emergence (an electron behaving differently in a brain versus interstellar space despite identical local conditions) is incompatible with locality; while locality may be only emergent and broken by gravity, those gravitational violations are invisible in everyday life, so a strong emergentist must show exactly how to modify the well-tested equations rather than merely asserting locality could be broken.

    “all of the hopes of the strong emergentists that it basically come down to the idea that an electron would behave differently if it were in a rock or in a brain or in interstellar space, even if it's local conditions of the other fields at the same point in space were the same... that's just incompatible with locality full stop.”

30

Fundamental ontology as a Hilbert space vector with Schrödinger equation evolution defines many-worlds; Zurek's quantum Darwinism falls within many-worlds even if he avoids other-worlds terminology.

1 point
  • As long as the fundamental ontology of your theory is a vector in Hilbert space and the fundamental dynamical equation is the Schrödinger equation, you are doing many-worlds, regardless of sub-versions; Zurek's quantum Darwinism and his account of pointer states fall within many-worlds even if he prefers not to talk about other worlds, and this is a healthy research program of people trying different ideas under the same core definition.

    “as long as the fundamental ontology of your theory is a vector in Hilbert space and the fundamental dynamical equation is the Schrödinger equation, You're doing Many-worlds.”

31

Black holes' characterization by mass, spin, and charge refers to their external appearance; interiors formed from different infalling matter must differ, but they are described by exterior properties since interior observations cannot return.

1 point
  • The statement that black holes are characterized only by mass, spin, and electric charge refers to what they look like from outside the event horizon; the interiors formed from different infalling matter must differ across spacetime, but we describe black holes by their exterior because once inside you cannot communicate observations back out.

    “when we say the black holes are characterized by mass spin and electric charge, we do mean by as far as what they look like outside is concerned.”

32

The block universe is not correct because physics is quantum; in many-worlds a block wave function evolves deterministically but the single-branch experience is irrelevant to libertarian free will questions.

1 point
  • The classical block universe is not actually correct because physics is quantum; in many-worlds it becomes a block wave function describing multiple worlds whose whole evolves deterministically but is irrelevant to human experience on a single branch, and in stochastic versions it is not fixed by present conditions; thus the block universe is irrelevant to libertarian free will, which is already ruled out by determinism within a single universe prior to any block-universe consideration.

    “the Block Universe is kind of irrelevant to the question of free will, determinism or non-determinism or the laws of physics.”

33

Scientific theory attention in public versus professional domains need not match; scientists should honestly convey how respectable and popular ideas are within professional communities when discussing them publicly.

1 point
  • The amount of attention a scientific theory receives in the public domain need not equal the attention it receives among professional scientists, as seen with Constructor theory and loop quantum gravity (both more prominent publicly than professionally) versus string theory; what matters is honesty, so when discussing an idea publicly scientists should clearly convey how respectable or popular it is within the professional community.

    “the amount of attention being paid to Constructor theory in the public domain is quite a bit larger than the amount that is paid to it by professional scientists.”

34

Einstein never saw de Broglie-Bohm, Everett, or QBism; Einstein was right that Copenhagen is incomplete, but Bohmian and Everettian theories are well-posed scientific theories Einstein would still have disliked.

1 point
  • It is misleading to contrast Einstein with de Broglie-Bohm, Everett, or QBism because Einstein never saw them; Einstein was right that the Copenhagen interpretation is incomplete and terrible, but Bohmian mechanics and Everettian many-worlds are well-posed scientific theories rather than the slapdash incoherent rules of Copenhagen, and Einstein would have disliked both (Bohm for its non-locality, Everett for denying the primacy of spacetime).

    “the other things you're mentioning, especially de Broglie-Bohm and Everett, those are well-posed scientific theories, they're different than the sort of slap dash set of incoherent rules that the Copenhagen interpretation gives you.”

35

Doing the work matters more than winning prizes; writing down general relativity matters more than vote-based awards, giving credit that no one can take away.

1 point
  • Doing the work is more important than winning the prize; if you were the person who wrote down general relativity (which never won a Nobel) that matters more than winning a vote-based prize, something no one can take away from you, so Carroll does not care who wins Nobel prizes.

    “if you are the person who wrote down the theory of general relativity, that matters more than winning some prize that people vote on.”

36

P versus NP asks whether problems easy to check (NP) equal problems easy to solve (P); sensible people believe P does not equal NP since P equals NP would make implausible things like axiomatizing creativity easy.

1 point
  • P versus NP asks whether problems whose solutions can be checked quickly (NP) are the same class as problems that can be solved quickly (P); to say P does not equal NP is to say some problems are hard to solve but easy to check, which all sensible people believe, because a world where P equals NP would make implausible things like axiomatizing creativity easy, so Carroll sees lots of reasons to think P is not equal to NP and none to think otherwise, though no proof yet exists and there is no deadline.

    “to think that P is not equal to NP is to say, I think there are some problems that are hard to solve, but it's easy to check the solution for... to say that they are equal as complexity classes is to say that there are no problems that are hard to solve, for which we can also check the solution easily”

37

Modern physics including much quantum mechanics requires roughly a year and a half of college-level math covering complex analysis, vector calculus, and differential equations; advanced tools are typically taught within physics courses.

1 point
  • The vast majority of modern physics, including much of quantum mechanics, requires only enough calculus to handle complex analysis, vector calculus, and solving differential equations (especially partial differential equations); roughly a year and a half of solid college-level math gives a strong grounding, and more advanced tools like tensors are typically taught within the physics courses that need them rather than assumed beforehand.

    “Once you understand complex analysis, vector calculus, and solving differential equations, especially partial differential equations, then you can do the vast majority of modern physics, including much of quantum mechanics.”

38

The early universe was hot yet low-entropy because gravitational effects matter; higher entropy means lumpier configurations so smooth early universe was low-entropy regardless of temperature.

1 point
  • Unlike a fixed-volume liquid where higher temperature means higher entropy, the early universe was hot yet low in entropy because gravitational self-interactions matter: in the presence of strong gravitational fields, higher entropy means lumpier configurations (up to black holes), so the smooth early universe was much lower entropy than a lumpy one regardless of temperature, and the casual phrase that 'gravitational degrees of freedom' were low entropy is ill-posed because it actually refers to the configuration of matter fields under gravity.

    “In the presence of these gravitational self-interactions, higher entropy means lumpier, generally.”

39

Black hole information measured by surface area does not logically imply storage at the horizon; holography and complementarity hold no fundamental answer exists because different observers see different things.

1 point
  • The fact that a black hole's information is measured by its surface area does not logically imply the information is stored at the horizon; the modern view of holography and complementarity holds there is no fundamental answer to where the information is stored, because from a far-away observer's perspective it is as if stored at the horizon, while from an infalling observer's perspective it is as if spread through the volume inside, and different observers see different things.

    “there's no fundamental answer to the question where the information is stored. From the perspective of a far away observer outside the black hole it is as if the information is stored at the horizon. From the perspective of someone falling into the black hole, it is as if the information is spread across the volume inside”

40

Both arguments about alien intelligence communication are reasonable: similar development might yield similar strategies, or vast differences might make thinking unimaginably different; no persuasive resolution exists.

1 point
  • Both arguments about communicating with alien intelligence are reasonable: one says alien minds developed in response to similar problems with similar resources and so would use recognizably similar strategies, the other says we have only one data point for advanced intelligence and should be humble that aliens could think in unimaginably different ways; Carroll finds no persuasive resolution because people with strong opinions on aliens always seem biased rather than coldly rational.

    “We only know one way to be advanced technological, highly intelligent creatures, mainly the way that we do it here on earth... we've only literally have one data point. So we should be humble”

41

Space emerging from entanglement does not mean entangled particles are spatially close; empty-space degrees of freedom vastly outnumber particles and determine geometry in this program.

1 point
  • In the program where space emerges from entanglement, it does not follow that two entangled particles are spatially close, because the relevant entangled systems are the vacuum states of empty space (which has vastly more quantum degrees of freedom than particles), so it is empty-space degrees of freedom that determine geometry; the ER=EPR conjecture works clearly for large entanglement between regions of empty space but is unclear for single entangled particles.

    “the entangled systems that we're talking about are empty space, not particles... there are far, far more quantum degrees of freedom in empty space in their vacuum states, than there are particles even at the center of a neutron star”

42

Nobel committee's three-person limit fails to credit thousands who built the LHC and discovered the Higgs; changing rules to honor big collaborations and builders like Lyn Evans would be better.

1 point
  • The Nobel committee's rule limiting a prize to three people fails to credit the thousands who built the LHC and discovered the Higgs boson, since the spokespeople who gave the discovery talks were not necessarily the most important contributors but the collaboration representatives at that moment; Carroll favors changing the rules so big collaborations can win and would support honoring builders like Lyn Evans.

    “the Nobel committee has made up a rule where it can only be given to three people at a time. And thousands of people worked on finding the Higgs boson.”

43

Online education increases accessibility but cannot replace in-person college socialization during ages 18-22 and credentialing for employment, so physical four-year attendance will persist.

1 point
  • Online education will continue growing in importance and has made high-level knowledge far more accessible, but it is not a replacement for in-person education because college serves crucial socialization roles during ages 18-22 (meeting diverse people, informal conversations, group work, dating) and provides credentialing that matters for employment; Carroll therefore sees no end to the system where people largely attend college for four years at a physical location.

    “it's not a replacement. There are other things that ordinary in-person education does that online education doesn't. Many of the interactions, many of the important roles played by going to college are socialisation roles”

44

Temperature has a quantum equivalent; quantum systems can be macroscopic, and quantum thermodynamics provides explicit formulas for defining temperature in quantum systems at equilibrium.

1 point
  • Temperature does have a quantum equivalent because the relevant distinction is between quantum versus classical and micro versus macro, which are two different axes; quantum systems can be macroscopic, and there is no problem defining the temperature of a quantum system in equilibrium with explicit formulas, as in the field of quantum thermodynamics.

    “The difference is between quantum and classical, and micro and macro, but those are two different sets of distinction. Quantum systems can be macroscopic.”

45

Gene-editing ethical concerns often rest on false naturalness distinctions; editing will happen because countries and individuals will do it, obvious goods exist like preventing disease, and probable net benefit despite real concerns.

1 point
  • Many ethical concerns about gene editing arise from a feeling of ickiness that edited genes are less natural than randomly recombined ones, which is a false distinction; gene editing is going to happen because it is not that hard and other countries and even individuals will do it, there are obvious goods like preventing disease and enhancing health, and it is probably less bad than many fear, though real concerns exist (editing for harmful goals, eliminating certain kinds of human beings) that current discourse treats too primitively.

    “I'm not someone who thinks like, there seems to be a bunch of people who think we can just write down rules and stop gene editing from happening in human beings. I think that's just hopelessly unrealistic. It's gonna happen.”

46

Wave functions should be thought of evolving in Hilbert space rather than events in spacetime; the spooky feature is that distance and locality are useful emergent concepts at all.

1 point
  • We should think of wave functions evolving in Hilbert space rather than things happening inside spacetime; on this view the spooky thing is not action at a distance but that the very idea of distance and locality is a useful emergent concept at all, so the question to be answered is why space and approximate locality exist, completely reversing the usual framing.

    “the weird thing, the thing that needs to be explained, is that the idea of distance and locality makes sense at all, is a useful emerging concept”

47

Effective altruism may appeal less to organizers because it attracts those preferring rational allocation over emotionality; emotions and passions are important drivers getting people to take difficult world-improving actions.

1 point
  • The effective altruism movement may focus less on effective organizing because EA appeals to people drawn to crisp rational allocation of charitable dollars and turned off by the emotionality of protest and organizing; but emotions and passions are not mistakes, they are important drivers that get people up off their butts to take difficult actions to improve the world, even if such actions are hard to analyze at a purely rational level.

    “emotions and passions are really important drivers of getting us up off our butts to do more difficult actions.”

48

The Higgs is the only neutral spinless boson found; its mass is puzzlingly low compared to Planck scale because it lacks a known symmetry protecting it, creating the hierarchy problem.

1 point
  • It is easy to write down neutral spinless bosons, but the Higgs is the only one we have found, and the puzzle is why its mass is so low compared to the Planck scale; for spin-one gauge bosons and spin-half fermions, gauge invariance and chiral symmetries respectively keep masses low, but the scalar Higgs has no known symmetry protecting it, which is the hierarchy problem, and the expectation that something like supersymmetry would explain it predicted new particles at the LHC that have not yet appeared.

    “The Higgs boson, since it's a scalar particle does not have any known symmetry that keeps its mass low. So there is a mystery. Why is the mass of the Higgs boson so low?”

49

Near-death and psychedelic experiences resemble each other because barriers between conscious perception and background visual processing break down; geometric hallucinations reproduce from visual cortex vibration equations.

1 point
  • Near-death experiences resemble psychedelic experiences because in both cases our guard is down and barriers between conscious perception and the visual processing constantly churning in the background of our brains come down; geometric hallucinations can be reproduced by solving equations for vibrations in the visual cortex, so the question is not why they appear under psychedelics or near death but why they do not appear all the time, the answer being that they are normally filtered out.

    “the reasons why you hallucinate when you're on psychedelics and so forth, and probably also when you're having a near-death experience is not because there is some new visual thing that pops up, but that the barriers between your conscious perception and the existing visual things churning around in the background of your brain come down.”

50

Abstract ideas have causal effects at the emergent level where causality applies via irreversible dynamics; ideas like patriotism are crucial degrees of freedom for understanding human behavior.

1 point
  • Abstract ideas can have causal effects on the material world, but only at the higher emergent level of description where causality itself applies (irreversible dynamics, arrow of time, unpredictability), not at the micro-physical reversible level; when describing actual human behavior, abstract ideas like patriotism or moral obligations are crucial degrees of freedom you must include for the best understanding, which is the same argument Carroll gives for compatibilist free will.

    “those abstract ideas are gonna have a hugely important causal effect. They help understand what's going on. You know, like this person sacrificed himself because of a feeling of patriotism.”

51

Asking about the internet's peak era is premature like asking about automobiles' peak in 1925; the obvious answer is now despite terrible aspects because it is genuinely useful and good.

1 point
  • Asking what the peak era of the internet was is probably premature, like asking the peak era of the automobile in 1925 when it was still new and far from equilibrium; with that caveat, the obvious answer is right now, because despite the internet's many terrible aspects it is also genuinely useful and good, and consumers bear some responsibility for finding the good rather than dwelling on the bad.

    “I don't think we're anywhere near an equilibrium state for the internet right now. So asking what was its peak era is probably wildly premature.”

52

Strong gravitational waves especially at intersections can form black holes; resulting black holes are not temporary because they persist through evaporation in classical general relativity.

1 point
  • It is possible to make a gravitational wave so strong it forms a black hole, especially when two strong-amplitude waves intersect and produce enough stretching of spacetime, but the resulting black hole is not temporary because in classical general relativity black holes are forever (modulo eventual evaporation); conversely no crest can be high enough to let light escape a black hole, because light cannot escape black holes.

    “it is possible to make a gravity wave so strong that it forms a black hole, but it's not going to be temporary. Black holes are not temporary.”

53

Polling physicists on quantum interpretations is useful for understanding beliefs and countering pluralistic ignorance; such polling does not decide physics any more than election polls decide elections.

1 point
  • Polling physicists about quantum interpretations does not decide the laws of physics any more than election polls decide elections, so opposition to such polls is misguided; polling is useful for understanding what people believe and can counter pluralistic ignorance, the Emperor's-New-Clothes phenomenon where people withhold their true beliefs because they wrongly assume everyone else disagrees, which can cause real harm.

    “pluralistic ignorance is when you are reluctant to say what you actually believe, because you believe everyone else believes something different and will think badly about you, but it's not true. They actually all think the same thing”

54

Attaching names to buildings honors people, so renaming after discovering terrible historical conduct is sensible; but one should not over-correct by assuming everyone in the 1950s-60s was a bigot.

1 point
  • Attaching someone's name to a telescope or building is entirely to honor and reward them, not to record history, so if a person turns out to have done terrible things it is perfectly sensible to change the name, as Caltech did with its Millikan-named library after recognizing Millikan's prominent eugenicist and racist activity; but one should not over-correct by assuming everyone in the 1950s-60s was a bigot, since many campaigned for LGBT rights at the time, and there is a distinction between private feelings and active efforts to harm.

    “the choice of attaching someone's name to a telescope is entirely to honour them, that's what it's for, it's not to remember something historically or whatever, it is a reward in some sense”

55

The relationship between microscopic and macroscopic theories is one-way information loss; multiple realizability means asking whether quantum world exists because of human agency is asymmetrically less legitimate than vice versa.

1 point
  • The relationship between microscopic and macroscopic theories is very much one-way, a relationship of throwing away information: from all particle positions and velocities you can compute the center of mass, but not vice versa, and there is multiple realizability whereby any set of macroscopic phenomena could arise from many different microscopic arrangements; therefore, unlike the symmetric causal relationship between past and present under reversible dynamics, it is not equally legitimate to say the quantum world is the way it is because of human agency, since we discussed agency for millennia without knowing quantum mechanics.

    “the relationship between the microscopic theories we're talking about and the macroscopic theories is very much one way. It's a relationship of throwing away information.”

56

Distinctions between private and public virtue do not map cleanly onto ethics and morality; philosophers admit these words are usually interchangeable with at most slight connotational differences.

1 point
  • While one can draw many distinctions between private and public virtue, those distinctions do not map cleanly and universally onto the words 'ethics' and 'morality'; philosophers who specialize in the field will admit these words are usually used interchangeably, with at most a slight connotational difference (ethics leaning toward rules, morality toward the private), so Carroll's interchangeable usage in the MacAskill conversation was simply not being careful rather than ignoring a real distinction.

    “If you talk to people who are philosophers specialising in ethics and morality... they will admit that usually these words are used interchangeably.”

57

Moral constructivism differs from nihilism and relativism by constructing moral rules from intuitions and justifying intervention without requiring objective realism, being more honest about moral feelings' origins.

1 point
  • Moral constructivism differs from both nihilism (which denies any moral truths) and relativism (which confines morality to one's society); the constructivist constructs moral rules from intuitions, believes them subjectively rather than objectively right, and has every justification for intervening to stop others from violating them, since the supposed requirement that intervention needs objective moral realism is itself an unexplained appeal to 'rights' that realism cannot ground any better, making the constructivist more honest about where moral feelings come from.

    “I as a moral constructivist, think that murdering innocent babies for no good reason is wrong, and if I see a person murdering innocent babies for no good reason, I will stop them. I will intervene in their affairs. I see no problem with that.”

58

The Big Bang's smooth low-entropy appearance is a precious clue for cosmologists rather than dismissible brute fact; it is natural to seek general context explaining this initial state.

1 point
  • Because we do not know what happened at the Big Bang and it very easily could have been highly inhomogeneous and high-entropy, the fact that it instead looks smooth and low-entropy is a precious clue for theoretical cosmologists rather than something to dismiss as a brute fact; while you are allowed to say the Big Bang just happened a certain way, since we lack understanding it is natural to seek a more general context in which this smooth low-entropy state is explained.

    “There's nothing that we know currently in our understanding of early universe, cosmology, or the Big Bang that would've prevented the Big Bang from being highly inhomogeneous and much higher entropy other than the fact that it doesn't look that way.”

59

The Big Bang is a moment in classical general relativity but does not actually exist as an event; it marks where classical GR breaks down, and 'time since Big Bang' means time since spacetime became classically describable.

1 point
  • The Big Bang is a moment in time in classical general relativity but does not actually exist as an event; it is a feature of classical GR in a regime where classical GR does not apply, so it functions merely as a warning sign that the theory breaks down, and 'time since the Big Bang' is really the time since spacetime became classical enough to describe, which makes 'one second after the Big Bang' sensible even though the Big Bang itself is not.

    “the Big Bang doesn't exist. It's a feature of classical general relativity in a regime where classical general relativity doesn't apply. So it's just a warning sign that says, Don't believe classical general relativity here”

60

Long-term predictions are difficult because society and ecosystems are highly nonlinear; you cannot adopt know-nothing attitudes but must form credences and weigh consequences even though accuracy is limited.

1 point
  • It is difficult to have high confidence in long-term predictions because society, life, and the ecosystem are highly non-linear systems with many possibilities for unanticipated consequences, yet you cannot adopt a know-nothing attitude; you must form credences and do your best even if your best is not very good, and while historical decisions had both good and bad consequences (the industrial revolution, religion), going forward you must weigh net positive against net negative because you either do a thing or you don't.

    “we need to have opinions about what is likely, we need to have credences. Okay? You can't just have a no nothing attitude that says, 'Well, it's hard. Therefore I don't care.' You have to do your best”

61

Utilitarianism fails systematically on moral intuitions like the repugnant conclusion and organ harvesting, and fundamentally the summing of utility across different people makes no sense.

1 point
  • As a moral constructivist, Carroll objects to utilitarianism on two levels: it fails to systematize his actual moral intuitions (as the repugnant conclusion and organ-harvesting cases show), and more fundamentally the very idea of utility as a sum over the sub-utilities of individuals makes no sense because there is no good way to compare utilities across different people; while any rule can formally be cast as maximizing some number, the specific feature of adding utility across persons is an unjustified starting point.

    “the specific feature of utility that worries me is the idea that we must add it up among people. That the idea that somehow people have different amounts of utility and we add them up, and that's what we maximise... That just seems to make no sense whatsoever to me.”

62

Human vision limits to 400-700 nanometers because expanding wavelength sensitivity would require costly chemical superstructure in eyes that do not pay off in human ecological niches.

1 point
  • Why human vision is limited to 400-700 nanometers cannot be answered by asking why we lack other abilities, but only by asking what those abilities would cost and whether they are worth it; all biological design is a set of compromises balancing many constraints, as seen in how the human brain consumes a large energy budget that only pays off in ecological niches where thinking beats running and jumping, and adding sensitivity to other wavelengths would require costly chemical superstructure in the eye.

    “All of evolution and all of biological design is a set of compromises. So you can't just say, why don't we have this ability? You have to say, what would it cost to get that ability? And is it really worth it?”

63

Intuitive physics explanations like time-dilation as falling's cause risk illusory understanding; negative pressure explaining dark energy is bad because dark energy's persistence without dilution explains better.

1 point
  • Purported intuitive explanations in physics (like time-dilation as the cause of falling, or negative pressure causing cosmic acceleration) carry a huge danger of giving the illusion of understanding by stringing together individually correct words that yield the right answer while a miracle step is smuggled in; the negative-pressure explanation of dark energy is bad because it relies on the unexplained quantity energy-density-plus-three-times-pressure, when a better explanation is that dark energy is persistent and does not dilute as the universe expands.

    “there's also a huge worry in purported explanations of this form, which is that they give the illusion of understanding, that they put together a bunch of words that sound kind of right and give you the right answer, and therefore you accept them.”

64

Recent JWST images do not contradict Big Bang or cosmic expansion; early compact galaxy formation prompts updating galaxy-formation theory, not abandoning the Big Bang or cosmic inflation.

1 point
  • Recent JWST images do not contradict the expanding universe or the Big Bang model; the claim is that galaxies are forming too early and too compact for the conventional model, but forming a galaxy is a messy complex-system problem, so data that does not fit our current best theory of galaxy formation should prompt updating that theory, not throwing out the Big Bang, while cosmic inflation (an addition to the standard model describing the first fraction of a second) is far less settled and could in principle conflict with the data.

    “if you find data that is not an immediate fit to your current best theory of galaxy formation, you don't throw out the big bang, you throw out your current best theory of galaxy formation or you even maybe probably just update it a little bit.”

65

Deutsch's objection that refuting experiments leave you unsure whether theory or experiment failed poses no problem for Bayesians; all possibilities are weighted using priors in rational updating.

1 point
  • Deutsch's objection that a refuting experiment leaves you unsure whether to lower credence in the theory or the experiment is no problem for a good Bayesian, because all possibilities (theory wrong, experiment wrong, both wrong, something missing) are coupled and weighed using priors; if a theory's prior is very strong and the prior that experimenters erred is non-trivial, then a faster-than-light-neutrino result barely moves credence in the theory, and you go get more data.

    “we have the credence that the theory is wrong, we have the credence that the experiment is wrong, we have the credence that both of them are wrong or that we're missing something entirely, and we're allowed to take all of our information and our other priors into consideration.”

66

Preferring popular vote over Electoral College reflects finding better ways to aggregate people's will into representative choices; changing this would be one step forward though not the most important political fix.

1 point
  • The point of preferring the popular vote over the Electoral College is not to change which candidate wins a past election (Trump genuinely won under the existing rules) but to find the best way to aggregate the will of the people into choosing representatives; even if a campaign would have won under either system, the current Electoral College reflects antiquated reasons that no longer apply, so popular vote would be one step forward though not the most important political fix.

    “My goal here is not to change the rules so that Clinton wins the game. My goal here is to get the best political system to pick our representatives and the reason why we have the Electoral College in its current form are very antiquated reasons that just don't apply anymore”

67

The speaker has no expertise in octonions and does not know the relevant details.

1 point
  • I have no idea, and I'm not an octonion expert.

    “the very short answer is I have no idea. I'm not an octonion expert.”

68

Obnoxious people are not worth having around even if brilliant since pleasant brilliant people exist; admirable are good people who disagree strongly, and academics should develop cordial disagreement capacity the broader world lacks.

1 point
  • Obnoxious people are not worth having around even if they are brilliant, because there exist other brilliant people who are pleasant to work with; the more interesting and admirable case is good people who strongly disagree with you, and academics often manage cordial, respectful relations alongside fierce intellectual disagreement, a capacity central to being a good scholar that the broader world should develop, though it can slip into damaging personal attacks.

    “if you say you're trying to hire someone as a post-doc or faculty member... and you say, 'Well, they're a really terrible person and I don't wanna talk with them, but they're really brilliant and do good work,' I don't wanna work with that person”

69

Inconsistent theories like quantum mechanics and general relativity pose no Bayesian problem because theories have domains of applicability; both fit their experiments well despite not extending to comprehensive unified theory.

1 point
  • Deutsch's worry that the conjunction of two inconsistent theories like quantum theory and general relativity has zero probability misses that theories have domains of applicability: general relativity is a classical theory true in the classical limit and never claimed to be comprehensive, while quantum mechanics is true in its domain, so a good Bayesian has no problem saying both fit their respective experiments well even though neither can be extended to a fully comprehensive theory simultaneously.

    “theories have domains of applicability. Neither quantum mechanics as we currently understand it, nor general relativity, especially general relativity, are purported to be true in every domain.”