22 researchers, including OpenAI's chief scientist, Anthropic's co-founder, and AI 'godfathers' Geoffrey Hinton and Yoshua Bengio, just co-wrote a paper on how AI building AI could set off an "intelligence explosion." Some of the wildest parts: - AI now completes 26% of Anthropic's R&D work with only high-level human oversight, up from 1% in March - Once AI hits expert level, one lab's current computing power could run the equivalent of "at least millions of top human researchers"

The article reports that 22 researchers co-wrote a paper on how AI building AI could set off an 'intelligence explosion.' Among the signatories the post names are OpenAI's chief scientist, Anthropic's co-founder, and the AI 'godfathers' Geoffrey Hinton and Yoshua Bengio. The article does not give the paper's title, publication venue, or date, and it does not name the remaining researchers.

Two figures are cited. The first is that AI now completes 26% of Anthropic's R&D work with only high-level human oversight, up from 1% in March. The article does not state the year of the March comparison, how the percentage was measured, or how Anthropic defines 'high-level human oversight.' The second is the claim that, once AI reaches expert level, one lab's current computing power could run the equivalent of 'at least millions of top human researchers.'

The article frames the paper as addressing the prospect of AI systems contributing to the development of further AI systems. The excerpt does not specify the paper's conclusions or any recommendations its authors make, and it does not state where the 26% figure or the computing-power estimate originated. It also does not clarify whether all 22 co-authors endorse the passages the post highlights as the paper's 'wildest parts.' Beyond those names and roles, the excerpt gives no further detail on the co-authors or on the paper's methodology.