logoalt Hacker News

flumpcakesyesterday at 5:04 PM1 replyview on HN

Are you saying intelligence is inherently unsafe? That seems like a pretty wild conclusion and I can't see any logical way to jump to your opinions.


Replies

dirk94018yesterday at 5:21 PM

I'm saying the capability to reason about novel situations is in tension with guaranteeing it never produces harmful outputs. We are talking about contradictory design constraints.