Reaction
The risk of declaring ambitious AI interpretability nearly solved
A commenter argues that making the claim without extraordinary confidence could undermine progress on AI alignment.
TLDR
A commenter warns that saying “ambitious interp is almost solved” without extraordinary confidence is dangerous. They argue that the world is taking AI alignment seriously at a fragile moment, and that such a claim could undo that progress.
Combined views
1.7K
1 Source, first seen ago