Safety researcher calls for scientific scrutiny of assumptions behind p(doom) claims

sebkrier · x · 2026-09-09

sebkrier argues people should be far more willing to interrogate the rationale, beliefs, and assumptions underpinning claims like "10% chance of everyone dying" — including beliefs about the nature of intelligence, how power works, instrumental convergence, Omohundro drives, and evolutionary theory applied to models.

He finds it puzzling how little is written to make these core beliefs legible and open to wider scientific scrutiny. Most popular write-ups (IABIED excluded from the good examples) treat audiences as idiots fed basic analogies and metaphors — persuasion exercises rather than science, skewing toward advocacy and field building.

Original post →

More from AGI Musings

AGI Musings channel →