davidad vs Richard Ngo: Agent Foundations First, or Institutions First?

davidad · x · 2026-09-03

davidad distills the AI safety field's route disagreement into mirrored quotes: MIRI says the field is too confused to solve agent foundations and should influence human institutions first; Richard Ngo says the field is too confused to influence institutions and should solve agent foundations first. A crisp portrait of the technical-vs-institutional stalemate in alignment.

Original post →

More from AGI Musings

AGI Musings channel →