David turner comments on Improving the safety of AI evals

[David turner]

[2305.10601] Tree of Thoughts: Deliberate Problem Solving with Large Language Models (arxiv.org)

i wonder if something like this can be used with my idea for ai safety