korrents

korrents · Lex Fridman Podcast · #452

Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity | Lex Fridman Podcast #452

Dario Amodei · 5h 14m · youtube.com ↗

34 korrents from this recording

1h2h3h4h5h
Dario Amodei 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. 0:00:31 · watch on youtube.com↗

    I think there are still worlds where it doesn’t happen in 100 years. The number of those worlds is rapidly decreasing. We are rapidly running out of truly convincing blockers, truly compelling reasons why this will not happen in the next few years.
  2. 6 min later
  3. 0:06:34 · watch on youtube.com↗

    I’ve seen the movie enough times, I’ve seen the story happen for enough times to really believe that probably the scaling is going to continue, and that there’s some magic to it that we haven’t really explained on a theoretical basis yet.
  4. 6 min later
  5. 0:12:30 · watch on youtube.com↗

    My strong instinct would be that there’s no ceiling below the level of humans.
  6. 1 min later
  7. 0:13:46 · watch on youtube.com↗

    So I think in some areas there may be ceilings that are very close to what humans have done. In other areas, those ceilings may be very far away. I think we’ll only find out when we build these systems.
  8. 7 min later
  9. 0:20:58 · watch on youtube.com↗

    Race to the Top is about trying to push the other players to do the right thing by setting an example. It’s not about being the good guy, it’s about setting things up so that all of us can be the good guy.
  10. 16 min later
  11. 0:36:43 · watch on youtube.com↗

    And I would suspect that if we can get to 90, 95% that it will represent ability to autonomously do a significant fraction of software engineering tasks.
  12. 13 min later
  13. 0:49:28 · watch on youtube.com↗

    And this version we’re seeing today of you make one thing better, it makes another thing worse, I think that’s like a present day analog of future control problems in AI systems that we can start to study today.
  14. 7 min later
  15. 0:55:59 · watch on youtube.com↗

    I think actually humanity has been protected by the fact that the overlap between really smart, well-educated people and people who want to do really horrific things has generally been small.
  16. 2 min later
  17. 0:58:25 · watch on youtube.com↗

    I don’t think this is an unsolvable problem. I think this is a science like the safety of airplanes or the safety of cars or the safety of drugs. I don’t think there’s any big thing we’re missing.
  18. 5 min later
  19. 1:03:45 · watch on youtube.com↗

    It’s actually dangerous to cry wolf. It’s actually dangerous to say this model is risky. And people look at it and they say this is manifestly not dangerous.
  20. 2 min later
  21. 1:05:44 · watch on youtube.com↗

    I would not be surprised at all if we hit ASL-3 next year. There was some concern that we might even hit it this year. That’s still possible. That could still happen. It’s very hard to say, but I would be very, very surprised if it was 2030.
  22. 3 min later
  23. 1:08:42 · watch on youtube.com↗

    But yeah, generally, you want to preserve mechanistic interpretability as a verification set or test set that’s separate from the training process of the model.
  24. 7 min later
  25. 1:15:27 · watch on youtube.com↗

    Look, my view actually is computer use isn’t a fundamentally new capability like the CBRN or autonomy capabilities are. It’s more like it kind of opens the aperture for the model to use and apply its existing abilities.
  26. 4 min later
  27. 1:19:08 · watch on youtube.com↗

    I think containing bad models is a much worse solution than having good models.
  28. 2 min later
  29. 1:20:58 · watch on youtube.com↗

    The second thing is I don’t think you can trust these companies to adhere to these voluntary plans on their own.
  30. 2 min later
  31. 1:22:52 · watch on youtube.com↗

    If we go to the very serious risks of autonomy and misuse that I talked about just a few minutes ago, I think that those are unusual and they warrant an unusually strong response.
  32. 3 min later
  33. 1:25:51 · watch on youtube.com↗

    And so I think the worst enemy of those who want real accountability is badly designed regulation.
  34. 2 min later
  35. 1:27:49 · watch on youtube.com↗

    I really think we need to do something in 2025. If we get to the end of 2025 and we’ve still done nothing about this, then I’m going to be worried.
  36. 6 min later
  37. 1:33:52 · watch on youtube.com↗

    I think it’s generally very unproductive to try and get someone else’s vision to look like your vision. It’s much more productive to go off and do a clean experiment
  38. 1 min later
  39. 1:34:48 · watch on youtube.com↗

    One way I think of it is the thing we’re all afraid of is the race to the bottom and the race to the bottom doesn’t matter who wins because we all lose.
  40. 6 min later
  41. 1:40:42 · watch on youtube.com↗

    But if everyone sees the broader purpose of the company, if there’s trust and there’s dedication to doing the right thing, that is a superpower. That in itself I think can overcome almost every other disadvantage.
  42. 3 min later
  43. 1:43:54 · watch on youtube.com↗

    And so this open-mindedness and this willingness to see with new eyes that often comes from being newer to the field, often experience is a disadvantage for this, that is the most important thing.
  44. 7 min later
  45. 1:50:41 · watch on youtube.com↗

    I don’t think it makes the model smarter. I don’t think it just makes the model appear smarter. It’s like RLHF bridges the gap between the human and the model.
  46. 2 min later
  47. 1:52:22 · watch on youtube.com↗

    I don’t think you can scale up humans enough to get high quality. Any kind of method that relies on humans and uses a large amount of compute, it’s going to have to rely on some scaled supervision method
  48. 4 min later
  49. 1:55:57 · watch on youtube.com↗

    I think we can go a little further and agree with some basic principles of democracy and the rule of law. Beyond that, it gets very uncertain and there our goal is generally for the models to be more neutral, to not espouse a particular point of view
  50. 4 min later
  51. 1:59:46 · watch on youtube.com↗

    If you only talk about risks, your brain only thinks about risks.
  52. 2 min later
  53. 2:01:17 · watch on youtube.com↗

    I want them to really understand that this isn’t Doomers versus Accelerationists. This is that, if you have a true understanding of where things are going with AI, and maybe that’s the more important axis, AI is moving fast versus AI is not moving fast
  54. 6 min later
  55. 2:07:12 · watch on youtube.com↗

    And the reason that I think that’s not the case is that, one, I think they just neglect the laws of physics. It’s only possible to do things so fast in the physical world.
  56. 4 min later
  57. 2:11:04 · watch on youtube.com↗

    So as inefficient as these processes are, we’re going to have to deal with them, because there needs to be some popular and democratic legitimacy in how these systems are rolled out.
  58. 10 min later
  59. 2:20:39 · watch on youtube.com↗

    Moore’s laws, scaling laws, they’re not laws of the universe. They’re empirical regularities. I am going to bet in favor of them continuing, but I’m not certain of that.
  60. 7 min later
  61. 2:27:38 · watch on youtube.com↗

    Then I think at some point it’ll flip around where the AI systems will be the PIs, will be the leaders, and they’ll be ordering humans or other AI systems around.
  62. 6 min later
  63. 2:33:32 · watch on youtube.com↗

    I expect that humans will continue to have a huge role and the nature of programming will change, but programming as a role, programming as a job will not change.
  64. 7 min later
  65. 2:40:21 · watch on youtube.com↗

    That is the thing I worry about most, the concentration of power, the abuse of power, structures like autocracies and dictatorships where a small number of people exploits a large number of people.
  66. 2 min later
  67. 2:42:12 · watch on youtube.com↗

    but we also need to address the risks because those risks are in our way. They’re landmines on the way from here to there, and we have to diffuse those landmines if we want to get there.