roon’s p(doom)
In their words
roon (Member of technical staff, OpenAI (pseudonymous X account @tszzl)) put their p(doom) at <1% (total value destruction (his definition of doom); excludes bad-but-not-total outcomes such as value lock-in or AI-run dictatorship), on a podcast on Dec 18, 2024.
I think I'm definitely below 10% of total doom, total value destruction, yeah. [Liron: Are you at above 1%?] It depends on the day. I don't know. I actually think that... My PDoom of like total value destruction is likely less than 1%, but certainly like non-zero, yeah.
Source · re-checked against the primary source.
Definition: total value destruction. On Sept 10, 2026 he deleted a quote-post agreeing with Hubinger's >10% and said he dislikes "the false precision of the doom numbers": "a quite low but real chance". The ">10%" circulating for him is not his.
In context
Among frontier labs in the database (20 people), the median is 12.5%. 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.