Here's Paul Christiano's (first author on the original RLHF paper, previously ran the language model alignment team at OpenAI) various estimates on risks from artificial general intelligence:…

Posters: anthrupad. anthrupad

3 likes recorded by the source corpus.