Mr. Altman,List of Lethalities #10 states that, with the current ML paradigm, powerful, future AGIs that could kill you if misaligned, somehow must have safety properties which generalize far…

Posters: anthrupad. anthrupad

2 likes recorded by the source corpus.