Eric Hartford asks: where's the evidence for AI doom?

QuixiAI · x · 2026-11-22

Eric Hartford's new blog post "Where's the Evidence for Doom?" (9 min read) examines how AI doomerism gained outsized influence over AI safety and policy, and the role Effective Altruism played. His core argument: no AI system has shown an independent intention to wipe out humanity; models lying, gaming metrics, or resisting shutdown in lab tests are real, but differ categorically from an autonomous drive to eliminate humanity — and the burden of proof lies with those predicting catastrophe.

Original post →

More from AGI Musings

AGI Musings channel →