Okay, well, I can't force you to give a shit, I'm saying whether CoT is deceptive is a pretty important question if major AI labs are making monitoring of it a core safety premise https://web.stanford.edu/~cgpotts/blog/cot/#deceptive