5 ms·
how sure are you that an llm won't be better at reviewing code for safety than most humans, and eventually, most experts?
by tomlue 9mo ago
how sure are you that an llm won't be better at reviewing code for safety than most humans, and eventually, most experts?
- kgwxd 9mo agoThey probably already can for a lot of things, but "Safety" is really about accountability when things go wrong. As a society, I hope we don't end up at "AI isn't perfect, but it's better than people on average, sorry if it failed you, good luck with that."
- hansmayer 9mo agoIt will only get better at generating random slop and other crap. Maybe helping morons who are unable to eat and breathe without consulting the "helpful assistant".