Y
Hacker News
new
|
ask
|
show
|
jobs
by
85392_school
230 days ago
AI is penalized for any and all emitted errors through RL, so this kind of carefulness is actually a strong sign of AI.