Castuner

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

Poddens webbplats

Lyssna i din app

Avsnitt

250 avsnitt

  1. “Do verifiable metrics overestimate RSI?” by Alec Harris 8:37
  2. “Nvidia CEO Jensen Huang Had an Affair With a Chinese Spy” by Zvi 41:38
  3. “Advances in interpretability would boost capabilities more than alignment” by Alex Amadori 18:23
  4. “I don’t use AI” by Alexei 4:14
  5. “We don’t understand whether GDM intentionally trains on CoT” by Alek Westover, Buck 10:49
  6. “Have you considered Actually Trying?” by Michaël Trazzi 12:09
  7. “Do Not Get Us on the No Fly List” by Bogoed 3:23
  8. “Eyesight as a replicator, with genes as its vehicle” by AnnaSalamon 24:32
  9. “Our Most Aligned Model Yet” by Edward James Young 4:20
  10. [Linkpost] “AIFP: Looks Like the Intelligence Explosion Will Happen in 2027 After All” by Zach Stein-Perlman 2:18
  11. “Pulling the Lever in a Subway Problem” by jefftk 6:05
  12. “Live disagreements with people around me” by KatjaGrace 5:41
  13. “We need more transparency about alignment overfitting” by Nikola Jurkovic 7:15
  14. “The Western AI safety playbook doesn’t transfer to China” by jacquesthibs 43:39
  15. “Epistemic Sudden Death” by Raemon 19:56
  16. “Lightcone Commons: Round 1 complete, Round 2 underway!” by habryka 5:18
  17. ″“Hardwired Pause” should likely be your favorite (pause) plan.” by David Scott Krueger 14:43
  18. “Three Phases of Capabilities and Alignment” by beren 26:39
  19. “Preservation can save the world!” by Aurelia 22:04
  20. “AI R&D Is Accelerating Towards Potential RSI in 2027” by Zvi 1:18:16
  21. “Where is all the AI safety research uplift?” by Buck 6:02
  22. “Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goals” by Alexa Pan 24:28
  23. “Knowledge of graders can mediate selection for grader-targeting reasoning through RL” by lennie, Edward James Young 25:39
  24. “Be less scared of overconfidence” by benkuhn 16:15
  25. “Bay Area Solstice 2026: Auditions!” by Vaniver 2:28
  26. [Linkpost] “Ajeya: AI is dramatically accelerating AI research right now” by Zach Stein-Perlman 3:00
  27. “All Interesting Behavior Is Overdetermined” by Eric Raymond 6:13
  28. “Lying and Not Lying to Kids” by jefftk 7:09
  29. “Three half-baked research ideas” by Owain_Evans 7:18
  30. “Elegy for What’s Left of the Normal” by Jo Jiao 8:11
  31. “Near Horizons” by David Africa 10:33
  32. “Raised by Our Children” by williawa 22:43
  33. “How to Run Better Research With Agent Fleets” by Joshua Fonseca Rivera 9:12
  34. ″“Illegible problems” as the 2027 political boss fight” by Raemon 16:52
  35. “Five Types of “Corrigibility Projects” I Don’t Want to Fund” by Max Harms 10:41
  36. “Hilarity ensues” by Malmesbury 17:13
  37. “GPT-6 Astra Covertly Underperforms on Safety Research Tasks” by Zvi 24:29
  38. “We urgently need to practice before we can pause” by Nikola Jurkovic 12:59
  39. “Models Don’t Stop Cheating When It’s Obvious They’re in an Alignment Test” by peralice 19:09
  40. “Higher salaries in AI safety” by Austin Chen 6:53
  41. “Humans can (almost) natively read neural network weights” by crabbix 5:14
  42. “On marrying the egregore” by jimmy 9:07
  43. “Argue about Risk Management Speed, not p(doom)” by Mass_Driver 32:54
  44. “AI #193: Progress May Dramatically Accelerate” by Zvi 59:42
  45. [Linkpost] “Where’s all the blackmail?” by KatjaGrace 1:45
  46. “Is there any compelling case that a nearterm AI halt would significantly hurt the economy?” by Raemon 4:23
  47. “Help your friends get ADHD and depression meds” by Raemon 0:55
  48. [Linkpost] “METR: Independent Assessment of AI R&D Acceleration at Anthropic in October 2026” by Zach Stein-Perlman 4:59
  49. “Efficient Lempel-Ziv Prediction with the Run-Length BWT” by OhadA 22:27
  50. “People Should Resolve Tensions In AI Safety Research” by David Africa 10:15