Mindscape 253 | David Deutsch on Science, Complexity, and Explanation
HTML-код
- Опубликовано: 15 окт 2023
- Patreon: / seanmcarroll
Blog post with audio player, show notes, and transcript: www.preposterousuniverse.com/...
David Deutsch is one of the most creative scientific thinkers working today, who has as a goal to understand and explain the natural world as best we can. He was a pioneer in quantum computing, and has long been an advocate of the Everett interpretation of quantum theory. He is also the inventor of constructor theory, a new way of conceptualizing physics and science more broadly. But he also has a strong interest in philosophy and epistemology, championing a Popperian explanation-based approach over a rival Bayesian epistemology. We talk about all of these things and more, including his recent work on the Popper-Miller theorem, which specifies limitations on inductive approaches to knowledge and probability.
David Deutsch received his Ph.D. in theoretical physics from the University of Oxford. He is currently a visiting professor in the Department of Atomic and Laser Physics at Oxford. He is a pioneer in quantum computation as well as initiating constructor theory. His books include The Fabric of Reality and The Beginning of Infinity. Among his awards including the Dirac Prize, the Dirac Medal, the Edge of Computation Science Prize, the Isaac Newton Medal, the Breakthrough Physics Prize, and a Royal Society Fellowship.
Mindscape Podcast playlist: • Mindscape Podcast
Sean Carroll channel: / seancarroll
#podcast #ideas #science #philosophy #culture 
Наука
“Knowledge is different from eucalyptus leaves.” Best quote ever
Lol I love it
As I sit down to play this, I'm beyond excited. These are far and away my two most admired scientists. I, like others, have been looking forward to this exchange for a long time. Now, nobody bother me for the next 1 hour, 42 minutes, and 6 seconds. 🙂
That was so satisfying! The one thing I was hoping for that didn't get discussed was that Deutsch thinks that there is such a thing as objective morality, and Carroll thinks the opposite. I think they each have compelling arguments, and I'd love to hear them explore that space together. Nevertheless, I'm thoroughly delighted by this episode!
AI generated timestamps
0:00 - Intro
0:29 - David Deutsch met Hugh Everett
1:10 - Deutsch pioneered quantum computing
1:58 - Explanatory knowledge is transforming the universe
2:41 - Monotony: long periods of stasis followed by rapid change
4:50 - Patreon shoutout
5:26 - Starting the interview with David Deutsch
5:45 - Deutsch recalls meeting Hugh Everett
8:33 - Everett understood the nuances of his own theory
12:06 - Wheeler invited Everett to give a seminar at UT Austin
14:16 - Deutsch was already sympathetic to Everettian QM before meeting Everett
18:27 - Newton may have understood more about energy/dynamics than he revealed
21:24 - Explanatory knowledge emerged relatively recently in the universe
23:03 - Knowledge allows "people" to transform the universe
25:15 - No beings with greater explanatory powers than humans
26:11 - Human brains are Turing complete
30:13 - Turing completeness is a hardware property
31:13 - Humans also have explanatory universality
33:50 - Mathematics and explanation are not fully formalizable
35:08 - Optimism: everything is possible with enough knowledge
36:17 - A simple program could lead to AGI, but we lack the right philosophy
39:05 - AGI may arise suddenly like Darwinian evolution
40:31 - Constructor Theory formalizes what transformations are possible
43:27 - Constructor Theory related to explanation and knowledge
48:12 - Still unclear on advantages of Constructor Theory perspective
49:25 - Constructor Theory provides overarching principles like conservation laws
51:47 - Newton may have understood energy concepts
53:43 - Constructor Theory connected to Popper's philosophy
55:05 - Deutsch rejects Bayesian epistemology
56:26 - Bayesian epistemology embeds credences that obey probability calculus
58:06 - Popper/Miller theorem: evidence only increases credence of propositions that closely match evidence
1:00:30 - Good explanations are key, not Bayesian credences
1:03:10 - Non-transitivity of support for explanations
1:07:36 - Using Bayesian reasoning in limited practical cases
1:08:54 - Looking for best explanations, not high probability theories
1:13:39 - Can't assign credences to unknown theories like simulation hypothesis
1:15:47 - False precision of credences, seek good explanations
1:18:21 - Galileo sought simplest explanation, not most probable
1:22:25 - Good explanations can't be formalized but can be judged philosophically
1:24:08 - Reject explanations that can explain anything
1:25:08 - Deutsch believes quantum theory is false
1:26:49 - Conflict between quantum theory and general relativity
1:28:22 - Theories are approximations, useful in different domains
1:29:03 - Problems with axioms of quantum field theory
1:31:09 - Considering alternative QFTs without problematic axioms
1:33:48 - Testing new QFT ideas like non-commuting particles
1:34:49 - New QFT compatible with Everettian view
1:35:57 - Preferring Heisenberg picture over Schrodinger
1:36:58 - Multiple worlds in quantum computation before decoherence
1:38:25 - Worlds split when computations branch and recombine
1:39:11 - Deutsch working on textbook and sci-fi book
1:40:06 - New approach to teaching QM focused on information
1:41:13 - Goal to change how quantum mechanics is taught
thanks for this!!
Thank you, you’re better than peanut butter
I’m so excited to listen to this one!! Carroll and Deutsch are heroes of mine.
I literally involuntarily murmured "Woooahh" when I saw that this episode exists. Wow - what a treat. Thank you so much, Sean (and David)!
Deutsch is the greatest thinker of our times.
For real
*you know of
and also, it takes one to know one, so I don't know if you're qualified to make this statement
@@opensocietyenjoyer on your first point, Deustsch would certainly agree with you. On the second point, credentialism is not good method of assessing hypotheses. Only creativity and criticism can.
@@opensocietyenjoyer and you are?
@@opensocietyenjoyer You don't need to have an elo of 3000+ to understand that stockfish is an excellent chess player
What a time to be alive!
More, David Deutsch, please!
Cant get enough either
Yaaayy we’re in the best branch of the universe where this exists🥳
fascinating way of thinking, thank goodness. stuff i've never encountered before.
He is one of my favourite contemporary thinkers.
Thanks Sean!!!! Great questions and beautiful runway you gave David without interruption
Been waiting for this one :)
for 4 years :)
Always learn a lot and have alot of new thoughts and ideas when listening to conversations with David Deutsche. Thanks!
Dr. Carroll - you're a very patient interviewer. This 7 second part summarizes everything succinctly: 1:10:45 until 1:10:52. Thank you for all the interviews so far. I follow your work and books. Looking forward to reading the upcoming one.
Deutsch! At last. This will be a treat. :)
WOW this is so damn amazing, and I am only at 44:50 where I have to pause damnit!
Fantastic job with the interview too Sean, as usual after all. But geee I am loving this!
Thank you both for the excellent discussion💜
Finally! What a great episode! I'm really looking forward to his new books and the paper on why Bayesian epistemology is false.
Fantastic interview
I've been waiting and hoping for this one!
Thank you, Gentlemen! Turing, incredible.
thank you very much, the very great conversation
Finally!!!
I've waited a long time for this one.
Deutsch is a Nobel mind, hopefully his year will come 🙏🏻
It's cool hearing things that I've wondered about independently mentioned here.
What a great episode. That was a really nice example of two minds working together and in competition in real time. Keeping each other honest but still feeding off the other.
I’ve been waiting!!!
What a kind and patient interview. Now inspired to read "The Logic of Experimental Tests, Particularly of Everettian Quantum Theory"
Not pretending to understand any of this but... I see why Deutsch wants to look at physics from a different angle: to create the kind of understanding breakthrough Darwin made. My problem is the examples of Constructor Theory I've heard just seem to state the obvious. The question of 'what is possible' already occurs in the course of current scientific thinking. Not sure what this theory is adding besides formalization of something rudimentary. What am I missing?
Nothing; his theories are hollow and redundant. Reminds me to the "emperor's new clothes” aimed at the undiscerning. I have follow his articles on this subject and come to the same conclusion not even D.D himself can explain his construct theories in a coherent an clear way. He also seems to ignore or be unaware of the implications of Chaos theory Complexity and its implications .
thank you for this!
HOLY MOLYYY its David Deutsch!!!! with Sean Carroll!!! * insert Jonah Hill screaming meme *
Loved the podcast! Didn’t quite get why David felt AGI is beyond current efforts using LLMs, neural nets, etc?
@desgreene2243 I think the idea is that they don't generate new explanatory knowledge. They collate existing knowledge.
I’m curious if Sean has read either of a David’s books The Beginning of Infinity or The Fabric of Reality. David has specific vocabulary that can take some getting used to.
You bet, those are pretty popular books, specially among supporters of the many-world interpretation of QM and philosophers of Physics in general. They certainly didn't pass unnoticed to someone as knowledgeable as Sean Carroll.
@@candidobertetti27 Right, I know Sean is aware of them and he is knowledgeable and knows about many worlds. I'm just asking if Sean has read David's books.
There were just a few points where if he had read his books I felt like he would have asked the questions in a different way.
Anyways, I enjoyed the conversation.
Hey Sean, ever thought of filming these? Loved your conversation with Brian Greene at WSF.
Deutsch’s preference for observables (Heisenberg over Schrödinger) reminds me of Rovelli’s relational QM.
Please explain the difference between the Heisenberg and Schrodinger picture in Everettian QM. That was the only part I didn't understand.
There is no such thing as Everettian QM. Everett made a trivial mistake. Having said that, the Heisenberg picture has severe ontological problems IMHO if you try to talk about non-unitary aspects of the theory in it.
I loved the talk, thank you Sean for being a courteous host. The bayesian epistemology stuff was a bit nuanced and incoherent for me. I felt like some important points were made or perhaps not quite made, and i'd love to know what those were - Id love to know what Sean's thoughts are on this. I guess I better go study Bayesian epistemology!
Bayesianism is bullshit. That doesn't mean that Bayesian methods are. They make perfect mathematical sense... as estimators of bias. It's just the heavy dose of amateur "philosophy" that comes with it that is complete nonsense.
0:00 Intro monologue
4:47 Patreon plug
5:22 Start of podcast
5:49 Everett
12:29 Enlightenment
18:04 Explanatory Knowledge
25:30 Universal Explainers
32:58 Explanation isn't Formal
35:29 Artificial General Intelligence
40:02 Constructor Theory
52:38 Popperian v. Bayesian Epistemology
1:07:40 The Linda Example
1:12:49 Credence vs Probability
1:19:53 Use of Bayes' Theorem
1:22:06 Good Explanations
1:24:52 Future of Quantum Theory
1:35:28 Multiworld Interpretations
1:38:37 Book Projects
Worth throwing down a few time stamps
Finally!
very cool
I've been waiting for this for a while... Thanks
35:25, David on AGI, AGI more into Philosophy?
Is there a seminal paper or reference I can cite that elaborates on the theory of humans as turing complete?
Creatures sitting around a campfire... full bellies and stars above... a time to share thoughts, think, and evolve!
It's interesting to consider the Law of excluded middle here. Explanatory Theories obviously do not follow the law, which in turn is required for Baysian reasoning
Amazing 👏 😊😅
You've introduced David Deutsch as a pioneer in quantum computing. What about Artur Ekert, from what I've heard his works were groundbreaking in quantum computing. I'm not an expert, it's from what I've heard from our theoret physicist Andrzej Dragan.
I googled it and Ekert was a student of Deutsch.
Science fiction wow!
One of those episodes that makes me want to write down some ideas.
Don't. There is already enough bullshit on the internet... like this episode. :-)
Damn. This is excellent
47:30 - Even the meta-narrative is underpinned with the continued use of dichotomy.
Link Please? At 45:00 you mention David's associate "Kiara Moletto"? I could not find them in your past videos or even with chatGPT with my attempt at spelling it. Can you help an idiot out?
I heard theory comes before evidence, evidence is to check the better theory among theories, can someone expand on this and give example please. Coming from beginning of infinity.
Ps: isn’t new theory based on the evidence we get which either contradict the existing theory. I might sound idk what I am talking but kindly explain me.
What if there is an experiment permitted by the laws of physics, but it requires more mass to be gathered in one place than exists in our causal sphere? It sounds like David Deutsch would tautologically call this experiment "against the laws of physics", which is a significantly different way of using the phrase to most people. Now imagine that 10 billion years ago, universe hadn't expanded sufficiently and there actually would have been enough mass that was available. David Deutsch would be committed to saying "this used to be permitted by laws of physics 10 billion years ago, but not any longer". I think most people would regard this as a practical limitation rather than a matter of laws of physics permitting or not permitting.
There are experiments (the things you do in the lab because you have money for them) and then there are Gedankenexperiments that you don't (because you don't have money for them or because they are impossible for reasons like the one you mentioned). To proper physicists only the first kind matters. The second kind is usually used by educators to explain science to students, sometimes with more, sometimes with less skill.
9:00 Everett made the big bucks $$
Looking forward to the "quantum information" book.
"Plato's Cave" ~ inside the BH...
At last!
Lots of sizzle, very little steak.
What would David mean with this pure state also being multiple worlds? (During the quantum computation, as he mentiones...)
Since any state can always be expressed as the sum of (basis) states it begs the question: what does a sum of states have to do in order NOT to be multiple worlds?! If it's not diagonalizing the density matrix, I mean...
(A bit like: what does an interaction have to do in order NOT to be a measurement? The Copenhagenists always fail to answer that question, but of course here we are among civilized people. So maybe this first question can also be answered?)
Decoherence also presupposes that we define an "environment", some coarse graining.
In either case ( Deutsches or Carroll 's), "talking about many worlds", is not a very well defined concept. Similarly to the Copenhagen -related interpretations, in these different Everettian QM versions there seems to be always that fuzziness between what we consider quantum and what semi-classical or macroscopic .
I haven't yet seen any compelling mathematical description of splitting geometry (and how that gives us GR as the classical approximation).
Am I asking too much? Perhaps yes, but untill then, all these debates about interpretations ( and second order interpretations within other ones...) seems to me very premature...
P.S. I'm not asking about a theory of Quantum Gravity here, only how that "splitting geometry" is related to the world we observe, that has as a classical approximation our best to date theory about spacetime (GR).
My problem here (at 1:36:12 in the clip) is the definition of "different worlds" (just like we would need a definition for "bulk matter" to judge whether that emerges from atomic properties).
Clearly, the block diagonal density matrix after decoherence can serve as a definition, but David wants to speak of different worlds in other cases too! Which cases? He doesn't specify, only gives an example.
NB: he says that you must be able to "recombine" the branches, but you *always* can recombine different terms as long as you still have a pure state, so that is meaningless...
Perhaps he means to restrict it to pointer basis decomposition, i.e. terms that *would* quickly become decoherent by environment interactions will *already* qualify as different worlds even if those interactions have not taken place... But that sounds like insisting that a single atom already has bulk matter properties, which is simply wrong.
at last
Deutsch is not only one of the best minds we have, he is my favorite person - feel like I could trust him with my life every time I hear him speak.
You trust someone who sounds like a drunk in an empty bar? ;-)
To borrow Nick Land's terms - drawing an interesting oblique Parallax, I think, in regard to this Great talk ... Enough with all that Transcendental Miserabilism! (which has been the tendency of "Progressive Materialist" worldviews, for more than 100 years already - imo. So much so that Optimism is the last protocol floating through intellectual discourses nowadays...while the Grand Unified Theory of Life Only remains the Meaninglessness of it All. Full period. Or?)
I actually have to say I'm slightly disappointed, it sounded to me as if Sean waa actually holding back and biting his tongue due to several potential flaws in his argumentation
Agreed
Care to elaborate on these flaws?
Agree.D.D his theories are hollow and at best redundant. Reminds me to the "emperor's new clothes” aimed at the undiscerning. I have followed his articles on this subject and came to the same conclusion not even D.D himself can explain his construct theories in a coherent and clear way. He also seems to ignore or be unaware of of theories of Chaos and Complexity and its implications .
Thumbs up before I even hear david
34:48
Mrs Richards: "I paid for a room with a view !"
Basil: (pointing to the lovely view) "That is Torquay, Madam ."
Mrs Richards: "It's not good enough!"
Basil: "May I ask what you were expecting to see out of a Torquay hotel bedroom window ? Sydney Opera House, perhaps? the Hanging Gardens of Babylon? Herds of wildebeest sweeping majestically past?..."
Mrs Richards: "Don't be silly! I expect to be able to see the sea!"
Basil: "You can see the sea, it's over there between the land and the sky."
Mrs Richards: "I'm not satisfied. But I shall stay. But I expect a reduction."
Basil: "Why?! Because Krakatoa's not erupting at the moment ?"
Well, I understood a small fraction of this. Either David Deutsch is a bad communicator or I am dumb. I'm sure it's the latter.
No, you're not dumb. He isn't a good communicator. What's communicated (within all that jumping and bouncing around) does not (let's say, "yet") provide any concrete explanation or prediction or have any empirical support. It's abstract and seemingly targeting a more fundamental level of analysis and understanding. But why stop at that level (of constructor theory)? What theory tells us when to stop digging? The current laws of physics are pretty awesome and tell us enough about what can and cannot be done.
@@johnrod7180Thank you for your reply. Yes, I think Sean also sides with the position you've outlined.
@@johnrod7180 that sounds like the sentiment of the typical person who is happy to remain within the boundaries of hard won knowledge. Fundamental physics is in dispute and requires resolution. It needs a deeper theory. It needs deeper or better explanations. In order to go beyond you have to make conjectures first. You have to dig. Constructor theory may well turn out to be useless but then again it may not. Its also not baseless. I can respect the attempt and the journey. If Constructor theory does yield fruit then I guess that would mean you would revise your limit on the accepted level of digging?
@@Wouldntyouliketoknow2 I wouldn’t say fundamental physics is in dispute. But yes, there are discussions, debates, and the pursuit of deeper insights. I wouldn’t say it requires resolution either, if by resolution you mean a sense of reaching a final, ultimate answer. But yes, it benefits from ongoing refinement and exploration. I would say that the pursuit of deeper knowledge and better explanations should always be grounded in rigorous scientific methodology. It seems to me that if I understand Constructor Theory (which I am not sure I do), it faces significant hurdles. It lacks experimental validation, which is crucial in science. Now you may say, well, it’s early to test it, we don’t even know how to test it, perhaps we haven’t developed the technology to test it. I may add it’s too abstract and not well defined at all to test in a meaningful way. I am not even sure it relates to relativity or quantum mechanics in any way. Like I said, it’s not making sense to me. Perhaps, this will change in the future. I hope it does because I don’t think of myself as someone who is happy to remain within the boundaries of hard won knowledge, as you put it. Give me more delicious stuff like relativity, quantum mechanics, etc. Give me also cutting-edge exploratory approaches but stay within the boundaries of the scientific method. Also show first that information is fundamental. I might be wrong but I think it’s an emergent concept (this is partly why I don’t understand how an emergent concept can be used at a more fundamental level than quantum mechanics and relativity). Anyway, thanks for the response. You made me think. Now I’m thirsty. :)
D.D theories are hollow and at best redundant. Reminds me to the "emperor's new clothes” aimed at the undiscerning. I have followed his articles and posts on this subject and came to the same conclusion: not even D.D himself can explain his construct theories in a coherent and clear way. He also seems to ignore or be unaware of of the implications of theories of Chaos and Complexity
Alright.......
David has spent more time in the past few years weeping about wokeness and postmodernism than developing his theory. Kind of like Bret Weinstein, except Deutsch has a legacy that will take a little bit longer to destroy with quackery.
Oh my, this has put me in a twiz (head gone, is that even a word), do I save and savour or just gobble it up now?
Had to edit to add AND it's 1 hour 45 minutes
It's hard for me to not be extremely skeptical of his claim that we have explanatory universality. For a man who readily, even gleefully admits how little we know...
It is admittedly in some sense a fairly weak claim that is impossible to think of a counter-example for. Perhaps I'm just detecting that it is philosophically incoherent, with no formal definition of explanation.
Respectfully disagree with most commenters. This episode mostly sounded like a physicist who hasn’t thought very hard about philosophy doing philosophy. He’s obviously a great physicist, but the epistemology here sounded extremely half-baked.
Agree; his theories are hollow and at best redundant. Reminds me of the "emperor's new clothes” aimed at the undiscerning. I have followed his articles on this subject and came to the same conclusion not even D.D himself can explain his construct theory in a coherent an clear way. He also seems to ignore or be unaware of the implications of Chaos theory Complexity .
This has been my impression as well. If David is available, I would like to use a Miro board and map out his reasoning around constructor theory and how he views criticaal rationality and epistemology. My skeptics senses are tingling…
Well it's easy to criticize without explanation, right bonehead?
Correction: he is not even a great physicist.
Return that dirty Templeton money.
Who?
@@CliffSedge-nu5fv Deutsch, whose research is funded by Templeton.
@@Paine137
Yikes.
Follow-up question: Why?
Shouldn't Deutsch know who they are?
@@CliffSedge-nu5fv It’s often a mix. Going along with Templeton’s nebulous philosophy that there is something “beyond science” - while not agreeing - in order to secure funding; using Templeton to reach more people and thereby generate interest in said research or books or projects; or, sadly, essentially to interview for the Templeton Prize, which was pathetically accepted by Martin Rees and Jane Goodall in years past.
"knowledge == static information", so I don't believe in knowledge I believe he's referring to understanding which is integration of information and its applicability.
Information == entanglement
Knowledge would imply static anything in our universe epistemologically makes no sense.
where can I find your definition of "Computation"?
His argument seems to be self-referential due to, by gaining the understanding of physics, you then have the ability to utilize those physics to "achieve" things beyond the capacity of the entity and environmental constraints.
I believe math is an oversimplification of the universe that is unreasonably effective. the universe, is quantum mechanical, and so blue, red and superposition is still 3 degrees of freedom, 3 degrees of "possible information" per bit. that is, according to Prof Doc Feynman, " a file clerk that can do three things instead of two!"
therefore, math which is binary, on off, fails to take into account the entire spectrum of shadow or on and off, or superposition, or red and blue simultaneously but not purple! i.e., it's a hologram. therefore, it mostly works but only 66.6666666666 repeating percent.
anyway, back to self-referential, I don't think it's a dq but I think we need to resolve it eventually.
The one rule for everyone is: Every ______ (pick your local social contract) has the right to stasis and the defense there of.
Stasis defined as: differential geometry of energy into system of critical points, defined by entanglement structure.
I'm not sure obedient and creative are a dichotomy.
my dog is obedient mostly, and she is pretty creative, infact creativity could also be thought of as adaption to constraints on the system. an organism like a plenary worm is given a boron environment and the "creativity" stems from adaptation to that environment.
the state of adaption is a "creative process"
entangling two things that have never been entangled before rendering novel information in the environment. enough novel infomation provides for the environment to manifest a solution due to repeated exposure to novel information in stasis based amounts
David supported Brexit. Logic is not his best trait.
First
This is wayyyy over my head 😅
😊
Omg finally!
Finally!