> We're developing a responsible access framework that makes models available to researchers for scholarly purposes while preventing misuse.

oh COME ON... "AI safety" is getting out of hand.