Do your capabilities homework
It seems to me that a lot of technical ai safety people haven't done their capabilities homework - and that's a shame! I'll try to illuminate here mainly with an example as to why I think people who care about safety should totally pay more attention to the trends and actively engage with them - the case for safe AI not through an additional loss term but as a consequence of the learning algorithm!RLVRIt's now been 1.5 years since R1 came out - the paper which really introduced RLVR (RL with ver
By RobinHa