AI ALIGNMENT FORUM"Why Not Just..."
AF

"Why Not Just..."

Aug 08, 2022 by johnswentworth

A compendium of rants about alignment proposals, of varying charitability.

49Deep Learning Systems Are Not Less Interpretable Than Logic/Probability/Etc

2y

13

48Godzilla Strategies

1y

19

36Rant on Problem Factorization for Alignment

1y

31

53Interpretability/Tool-ness/Alignment/Corrigibility are not Composable

1y

4

63How To Go From Interpretability To Alignment: Just Retarget The Search

1y

24

34Oversight Misses 100% of Thoughts The AI Does Not Think

1y

23

36Human Mimicry Mainly Works When We’re Already Close

1y

4

55Worlds Where Iterative Design Fails

1y

17

61Why Not Just... Build Weak AI Tools For AI Alignment Research?

9mo

2

35Why Not Just Outsource Alignment Research To An AI?

9mo

6

44OpenAI Launches Superalignment Taskforce

5mo

0