Ajeya Cotra warns AI safety may face a one-year crunch window
In a full 80,000 Hours interview, Ajeya Cotra said automated AI research could compress decades of progress into months, leaving perhaps six to eighteen months to redirect AI labor toward alignment, biodefense, cyberdefense, and institutional adaptation.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why it moved the index
The full attributable transcript adds a concrete governance mechanism: competitive incentives may keep frontier labs accelerating just when scarce AI labor should be diverted to protective work, while public warning evidence may remain commercially sensitive. It is a forecast and policy analysis, not evidence that an intelligence explosion or loss of control occurred.
Assessment history
- R1Toward 48 · confidence 59
Adds a dated full interview with a distinct compressed-response-window and coordination mechanism.
14 Aug 2026