Bryan Cantrill pushes back on AI doom claims: experts must not abuse public trust

Simon Willison · rss · 2026-09-15

Oxide co-founder Bryan Cantrill responded to former Anthropic employee Jacob Coxon's claim that many Anthropic researchers believe AI "could kill us all by the end of the decade," calling it a "contagion of fear." Recounting a youthful lesson about causing unjustified panic, Cantrill argues that claims like "hacking critical infrastructure" and "extinction-level bioweapons" rest on hand-wavy extrapolation, and that Coxon is no expert in either domain. The burden of explanation lies with those raising alarms: domain experts implicitly hold the public's trust and must be maximally circumspect when doing so. He also voiced doubts about the bioweapon scenario on the Oxide and Friends podcast.

Original post →

More from AGI Musings

AGI Musings channel →