David Dalrymple (davidad)’s p(doom)
In their words
David Dalrymple (davidad) (AI safety researcher; programme director, ARIA Safeguarded AI) put their p(doom) at <5% (p(doom), decomposed into five residual routes (wrong about the 'wisdom attractor', Malthusian competition, bio misuse, military first strike, 'warring gods')), on a podcast on Jul 12, 2026.
I think I think we're on track yeah. I like my, my P doom is less than 5% now. I I think we're in good shape.
Source · re-checked against the primary source.
The Cognitive Revolution (Nathan Labenz), [1:04:02] in the episode's published auto-transcript. Also 'even P doom less than 5%. Like that's, that's quite a lot of doom for humanity to be taking on.' [1:08:26]. Host intro: his p(doom) 'has fallen from the seventies in 2022 to under five percent today' (host's words, not a row). Researcher, outside the public-figures remit; include only if the researcher sweep didn't find it. Check whether his ARIA role is current.
In context
Among government in the database (5 people), the median is 7.5%. Across all 223 people it is 10%. How the medians and the index are built.
A p(doom) is a personal, subjective estimate of the probability that advanced AI ends in human extinction or a comparable, irreversible catastrophe. Scope and timeframe differ from person to person, so compare with care: why p(doom) numbers disagree.