Derek Thompson Rebuts Claim That AI Labs' Safety Pivots Are Financially Motivated

Turn_Trout · x · 2026-09-14

Derek Thompson pushes back against Sabine Hossenfelder's take that frontier AI labs only embraced safety because (a) weak safety measures risk court or jail, (b) no major model breakthroughs are coming anyway and they need an excuse, and (c) public opinion has shifted—with existential-risk talk serving to distract and keep markets happy.

Thompson argues this is "conclusively wrong": Dario Amodei told Ross Douthat in February—seven months earlier, while Anthropic's annualized revenue was rising faster than any company in modern history—that he favors an internationally coordinated slowdown and would be "all for" a global one, a position he has held for years from Anthropic's early days to trillion-dollar IPO projections.

Related event: Dario Amodei Calls for Slowing the AI Frontier, Draws Wide Skepticism(30 posts)→

Original post →

More from AGI Musings

AGI Musings channel →