richard_ngo comments on AGI Safety Fundamentals curriculum and application

[richard_ngo]

Actually, Joe Carlsmith does it better in Is power-seeking AI an existential risk? So I’ve swapped that in instead.