Conversation

Top-level takeaway of tonight's catch-up reading on ML... I still find it hard to take AI Risk seriously as a special problem here. There are definite risks but they don't seem qualitatively different from other kinds of engineering risk.
9
14