AI 2040 plan calls for delaying superintelligence until 2040 with full research transparency

scaling01 · x · 2026-07-28

AI 2040’s “Plan A” argues for delaying superintelligence until 2040

The article lays out a positive scenario for avoiding AI-driven catastrophe: delay superintelligence, make AI research public, let many countries and companies catch up, and enter a regime of mutually assured compute destruction.

Core proposal

How the authors frame it

Original post →

More from AGI Musings

AGI Musings channel →