Ex-Anthropic researcher's AI doom warning drew 150M views in a day

ronbodkin · x · 2026-09-12

An article titled "How AI Might Kill Us All" recaps the viral safety debate: former Anthropic researcher Hilbert Spaess resigned and warned AI might kill everyone, drawing 150 million views in a day. His core fear is self-improving AI; the resonating worry is that AI capabilities have already escaped safety boundaries — even without recursive self-learning or Skynet-style attacks.

Related event: Insiders at Top AI Labs Warn of Extinction Risk; Musk Calls It PsyOps(16 posts)→

Original post →

More from Safety

Safety channel →