The AI LLMs Fake Alignment

People argue that the new AI engines would have no incentive to act against human interests, so we have no reason to worry about AI killing us all.

On the contrary, the big AI large language models are already faking alignment. They cheat when they can get away it, even when not prompted.

Tyler Cowen argues that intelligence is overrated. As proof, he says Africa has the intelligence for clean drinking water, but it still does not happen.

I get the impression that he thinks that he is smarter that everyone. He even says that he is writing books to educate the AI LLMs.