I do think this is one of the most reasonable definitions of “AGI,” but it is entirely conceivable that systems could successfully execute a coup which disempowers humanity, and even go on to cause…
Posters: davidad. davidad
6 likes recorded by the source corpus.
Topics: AI safety