Andrew Curran argues recursive self-improvement will allow the U.S. government to nationalize AI labs and run them without human staff

BI

Bogdan Ionut Cirstea@BogdanIonutCir2TECH

@d29756183 @AndrewCurran_ (and because the fully automated ML researcher would probably be pretty easy to translate into a fully automated prosaic safety researcher, and at the very least help a lot towards fully automated risk mitigation research, including conceptual)