Rethinking AI alignment and slowdown through intelligence explosion scenarios

A user calls his lack of attention to intelligence explosion scenarios probably his biggest intellectual mistake of recent years. He thinks such scenarios could leave less room to iterate on AI alignment and says he undervalued slowing development. In a follow-up, he adds that misuse still matters, many interventions could help regardless of how likely an intelligence explosion is, and a “self-improvement” framing can mislead.