“Maybe the AI capabilities researchers aren’t very smart” is a very very hazardous assumption on which to pin one’s AI safety hopes https://t.co/2bGHQXTlPP

Posters: davidad. davidad

2 likes recorded by the source corpus.