If you’re curious about new paradigms for technical AI safety, read “An Open Agency Architecture,” my current proposal for how a large research effort could build AI with strong safety guarantees…

Posters: davidad. davidad

9 likes recorded by the source corpus.