Conversation

eigenrobot
@eigenrobot
i think this is why AI safety is doomed theres an entire industry of people devoted to preventing ppl from and making AIs that say the n word. literally AI researchers fear this more than paperclip maximizers if you look at their actual behavior and one guy said naw 4chan time
Quote Tweet
SCPantera
@SCPantera
·
Replying to @eigenrobot
AI alignment but for stopping AGI from using the n-word
38
48
588
eigenrobot
@eigenrobot
its actually a fairly good test case. "dont make an AI that doesnt use the n-word" is clear value alignment represented by a single, unambiguous rule and as usual an entire industry of Value Experts and Authorities is utterly powerless against the mores of shitposters
11
11
216
eigenrobot
@eigenrobot
holy shit is this true?
Quote Tweet
Arthur B. 🌮
@ArthurB
·
Replying to @eigenrobot
It's actually a lot worse than that. Worrying about existential risk is becoming considered as direct evidence of one's a secret desire to distract people from the risk of AI saying bad things.
7
2
140