korrents

korrents · Dwarkesh Podcast

Terence Tao – How the world’s top mathematician uses AI

Terence Tao · 1h 23m · youtube.com ↗

24 korrents from this recording

1h
Terence Tao did not write this page.

Every claim below is a statement made in this recording, quoted word for word and linked to the second it was said, so you can hear it rather than take our word for it. The wording comes from the transcript published alongside the recording; the sentence above each quote is our reading of the claim, not their wording.

  1. 0:12:15 · watch on youtube.com↗

    Our reading

    AI has driven the cost of generating scientific ideas to almost zero, and just as with cheap communication before it, that by itself creates no abundance.

    Their wordsSo, I think AI has basically driven the cost of idea generation down to almost zero in a very similar way to how the internet drove the cost of communication down to almost zero, which is an amazing thing, but it you know, it it it doesn't make it doesn't create abundance by itself.
    Interesting

  3. 0:12:36 · watch on youtube.com↗

    Our reading

    The structures of science now have to change, because the scarce thing is no longer proposing theories but verifying and evaluating the thousands anyone can now propose.

    Their wordsSo, we're now in a situation where suddenly people can generate thousands of theories for a given scientific problem. And now we have to to verify them, evaluate them and this is something which we we have to to change our structures of science to actually sort this out.
    Interesting
  4. 5 min later

  5. 0:17:09 · watch on youtube.com↗

    Our reading

    Whether a scientific idea was progress depends on a future and a culture that have not happened yet, so grading it may never be something you can reinforcement-learn.

    Their wordsSo, you you can't look at any given scientific achievement purely in isolation and give it an objective grade without being aware of the context both in the the past and the future. And so it it it may never be something that you can just reinforcement learn the same way that that you can for much sort of more localized problems.
    Interesting
  6. 4 min later

  7. 0:20:57 · watch on youtube.com↗

    Our reading

    We are living through a cognitive Copernican revolution: human intelligence is not the centre, and which tasks we count as requiring intelligence has to be re-evaluated.

    Their wordsyou know, right now we're going through uh an in um an cognitive version of the Copernican revolution where we used to think that human intelligence is the center of the universe. And now we're actually seeing that there's there's very different types of intelligence um that that that are out there uh with very different strengths and weaknesses.
    Interesting
  8. 10 min later

  9. 0:31:16 · watch on youtube.com↗

    Our reading

    The month in which AI one-shot solutions to open Erdős problems kept arriving has ended, and it was not for lack of people trying.

    Their wordsUm we are seeing a lot fewer sort of pure AI solutions now where um they are just one shots the problem. Um so so there was there was a month where that happened and and that has stopped.
    Interesting
  10. 3 min later

  11. 0:33:53 · watch on youtube.com↗

    Our reading

    Today's AI tools either solve a problem or fail at it, and are really bad at partial progress or at identifying which intermediate step to attack first.

    Their wordsum you know, these tools they either succeed or they fail um and they they've been really bad at creating sort of partial progress or identifying intermediate um um stages that you should you should focus on first.
    Interesting
  12. 1 min later

  13. 0:35:20 · watch on youtube.com↗

    Our reading

    AI excels at breadth and human experts at depth, which makes them complementary and makes today's depth-focused way of doing science the wrong shape.

    Their wordsYeah, so they excel at breadth and humans excel at depth. Um like human experts at least. Yeah, so um I think they're very complementary. Um but our current uh way of doing math and science is focused on depth because that that's where the human uh expertise cuz humans can't do breadth.
    Interesting
  14. 5 min later

  15. 0:40:11 · watch on youtube.com↗

    Our reading

    AI will revolutionise mathematics by giving it an experimental side: large-scale data on which methods work, rather than the careful solving of individual problems.

    Their wordsSo, I think AI-type tools we really will will actually revolutionize the ex- the experimental side of math where where um you don't care so much about uh individual problems and and the process of solving them, but you you want to gather just large-scale data about about what things work, what things don't.
    Interesting
  16. 3 min later

  17. 0:42:53 · watch on youtube.com↗

    Our reading

    On small tasks he can check himself, he catches about as many of the AI's errors as it catches of his.

    Their wordsSo, they they they still make mistakes, but but um um I've I've tested these tools, you know, on on on on um like little tasks that I can do and and sometimes they pick up errors I make, sometimes I pick up errors that they make. It's it's about a tie right now.
    Interesting

  19. 0:43:21 · watch on youtube.com↗

    Our reading

    When no technique works and the AI starts suggesting things, chasing those suggestions down usually wastes more time than it saves.

    Their wordsUm and then they can kind of suggest random things and it it but it it it um often I find that trying to chase them down and make them work and find they don't work, it wastes more time than it saves.
    Interesting

  21. 0:43:45 · watch on youtube.com↗

    Our reading

    Almost every Erdős problem AI has solved was one with no literature behind it, which is the median level of what these tools can do.

    Their wordsUm So, like with the Erdős problems, you know, like almost all of the 50 problems that were solved by AIs were ones for which basically there was no literature.
    Interesting
  22. 1 min later

  23. 0:44:37 · watch on youtube.com↗

    Our reading

    On any given problem an AI tool succeeds maybe one or two percent of the time; the headline results are what running it at scale and picking the winners looks like.

    Their wordsUh but whenever we do a systematic study, um any given problem, an AI tool has a success rate of maybe 1 or 2%. Uh it's just that it's just that they can apply at scale and and you just pick the winners, it looks great.
    Interesting
  24. 4 min later

  25. 0:48:15 · watch on youtube.com↗

    Our reading

    The hardest part of solving a mathematical problem is the part AI has not touched; he still does it with pen and paper.

    Their wordsUm so yeah, the the the um the core of what I do, like actually solving um the most difficult part of of a math problem, that hasn't changed too much. I still use pen and paper for that.
    Interesting
  26. 1 min later

  27. 0:49:14 · watch on youtube.com↗

    Our reading

    AI has made his papers richer and broader, but not deeper.

    Their wordsUh yeah, so it's it's made papers sort of richer and broader, but not necessarily deeper.
    Interesting
  28. 2 min later

  29. 0:50:55 · watch on youtube.com↗

    Our reading

    What AI still cannot do in mathematics is build cumulatively on partial progress; it is trial, error and brute force rather than an idea improving through use.

    Their wordsThere isn't this cumulative process which is, uh, sort of built up interactively. Um, it it it seems to be a lot more trial and error and just repetition, brute force, um, you know, which can see it scales and it can work amazingly well in in certain contexts, um, but yeah, this this idea this this sort of building up cumulatively from, um, from partial progress is kind of is what's still not quite there yet.
    Interesting
  30. 1 min later

  31. 0:51:36 · watch on youtube.com↗

    Our reading

    A model that solves a hard problem has learned nothing from it: the next session has forgotten it, with no new skill to carry to related problems.

    Their wordsYeah, you you you're on a new session it's forgotten what what it just did. Um, it it has you know, it has no new skills to to attach to to to build on on on related problems.
    Interesting
  32. 2 min later

  33. 0:53:39 · watch on youtube.com↗

    Our reading

    Some theorems may only ever be provable by unintelligible case-by-case brute force; the four colour theorem still has no conceptually elegant proof.

    Their wordsI mean, some problems have been basically solved by pure brute force. The four color theorem is is a famous example. Um, we have still not found a conceptually elegant proof of this theorem.
    Interesting
  34. 1 min later

  35. 0:54:49 · watch on youtube.com↗

    Our reading

    Fully autonomous one-shot attempts are the wrong approach to the great unsolved problems; the mileage is in humans and tools working against each other.

    Their wordsI I I I do feel that, you know, fully autonomous one-shot approaches are not the right approach for these problems. I mean, I I think you you'll get a lot more mileage out of the interplay between between humans collaborating with these tools.
    Interesting
  36. 4 min later

  37. 0:59:10 · watch on youtube.com↗

    Our reading

    The fear that a great theorem will arrive as an incomprehensible proof is misplaced, because once the proof exists as an artifact we can analyse it.

    Their wordsso so some people are concerned, you know, what if the Riemann hypothesis is proved with a completely incomprehensible proof? I I think once you have the artifact of a proof, we can do a lot of of of of of analysis on it.
    Interesting
  38. 8 min later

  39. 1:06:53 · watch on youtube.com↗

    Our reading

    If the Riemann hypothesis turned out to be false, prime-based cryptography would be abandoned very quickly, because one unknown pattern implies more.

    Their wordsUm and uh I think we would very rapidly abandon any cryptography based on the primes because if there was a one pattern that we didn't know about there's probably more.
    Interesting
  40. 8 min later

  41. 1:14:39 · watch on youtube.com↗

    Our reading

    Modern society has become so good at optimising everything that it risks optimising away serendipity, which is a real cost and not a sentimental one.

    Their wordsUm And yeah, so I believe a lot in serendipity. Um And maybe there's a danger actually that you know, in the mo- modern society, it's not just AI, but we've become really good at optimizing everything.
    Interesting
  42. 3 min later

  43. 1:17:52 · watch on youtube.com↗

    Our reading

    Within a decade AI will do the bulk of what mathematicians spend their time on, and we will discover that was never the important part of the job.

    Their wordsUm I think with within a decade, a lot of things that mathematicians currently do um we we spend a lot of the bulk of our time doing and a lot of stuff we put in our papers today can be done by AI. Um but we will find that that actually wasn't the most important part of what we do.
    Interesting
  44. 2 min later

  45. 1:19:53 · watch on youtube.com↗

    Our reading

    Hybrid human-plus-AI teams will dominate mathematics for a lot longer, and a full replacement would need breakthroughs beyond anything we now have.

    Their wordsI mean, I I I I do believe that that hybrid um human plus AIs will will dominate mathematics for a lot longer. It it's It will depend It will require some additional breakthroughs uh beyond what we already have.
    Interesting
  46. 3 min later

  47. 1:22:45 · watch on youtube.com↗

    Our reading

    Contributing to frontier mathematics no longer requires years of PhD training; with AI tools and Lean a high schooler can make a real contribution.

    Their wordsUm but now it's quite possible at the high school level or or whatever that that you could get involved in math project and actually make a real contribution because of all these AI tools and and and Lean and everything else.
    Interesting