Researcher Calls Out p(doom) Contradiction: Why Accelerate Labs If Doom Is Nonzero?

_arohan_ · x · 2026-09-21

Arohan poses a pointed challenge to the AI safety community: if your p(doom) is above zero, why are you willing to work on research, training data, and training at frontier labs? He suspects those people privately believe current models carry p(doom) = 0, and asks them to say so publicly.

The post touches on the long-running hypocrisy debate in AI safety: evangelizing existential risk while accelerating development.

Related event: Researcher Questions Why AI Staff Accelerate Development If p(doom)>0(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →