Lab insiders reportedly quote ~10% AI doom publicly but believe 30-90% privately

jeremiecharris · x · 2026-09-10

Jeremie Harris points out that people at AI labs often carry two numbers when discussing existential risk: a public figure of roughly 10% — chosen because it sounds serious but vaguely reasonable — and a private estimate that can run 30-90%. The observation, reacting to Peter Boghossian's quip that "10% chance of death and 90% chance of superabundance seems like a good bet," highlights the gap between lab insiders' public statements and their actual beliefs about AI risk.

Original post →

More from AGI Musings

AGI Musings channel →