Hacker News new | ask | show | jobs
by sometimelurker 27 days ago
> all

never, AI safety/alignment by definition can't be fully automated.