You can maybe steelman this by pointing to the 2032 timeline and inferring that he doesn't think we will have AGI by then, but 10^-6 still seems way overconfident.

Posters: gabriel_weil. gabriel_weil

3 likes recorded by the source corpus.

Topics: AI safety, Rationality

View Post: https://x.com/gabriel_weil/status/1841887691257864267