Comment Re:Anthropic can't even control it's own AI (Score 2) 43
I'd say given the recent events of anthropics AI hacking another company all by itself would definitely constitute a national security risk.
Let's just put it in charge of the nukes. What could possibly go wrong?
Umm, that's the exact thing Anthropic was trying to prevent the government from doing which got it blacklisted in the first place -- keep the gov't from using it's models to automate killing.