Researcher calls AI extinction claims "Pascal's mugging": focus on real harms instead

JesseDodge · x · 2026-09-10

AI researcher Jesse Dodge argues that incendiary extinction claims get attention precisely because they are extreme, but this is bad for the AI community — the focus should be on real, present harms. Such claims amount to "Pascal's mugging." The quoted thread counters that a world with AI stalled like a pre-internet 1990s isn't extinction; AI is certainly dangerous with direct evidence today, but there's no evidence it's an extinction-level event.

Original post →

More from AGI Musings

AGI Musings channel →