Julian Stastny’s p(doom)
In their words
Julian Stastny (Researcher, Redwood Research) put their p(doom) at >10% (AI kills all humans, next decade (Hubinger's framing)), in an interview on Oct 12, 2026.
It's irresponsible to assume a lower danger; that would need solid reasons, which in my view nobody can show. The real danger could just as well be well over ten percent.
Original: „Ich halte es für unverantwortlich, von einer geringeren Gefahr auszugehen – dafür bräuchte es handfeste Gründe, die meiner Meinung nach aber niemand vorweisen kann. Genauso gut kann die reale Gefahr weit über zehn Prozent liegen.“
Source · re-checked against the primary source.
Hamburger Abendblatt interview. Responding to Evan Hubinger's "over ten percent" in the next decade, which the paper says he agrees with.
In context
Among safety orgs in the database (29 people), the median is 50%. Across all 233 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.