AI 2027 authors pitch Plan A: delay superintelligence to 2040 with fully open AI research

Turn_Trout · x · 2026-09-16

Alignment researchers including Daniel Kokotajlo released "Plan A", a positive vision following AI 2027: delay superintelligence until 2040, make all AI research public, let dozens of companies globally reach the frontier, and enter a regime of mutually assured compute destruction. The core proposal is an international deal with total R&D transparency and guardrails, contrasted against alternative Plans B/C/D/S. TurnTrout also argues racing is an industry-level property, so unilateral pauses by OpenAI/Anthropic can't work without government coordination.

Original post →

More from AGI Musings

AGI Musings channel →