I think this is basically true, but there's a different way of saying this. LLMs are not aligned _for_ humans in a very similar way to the way that humans them

I think this is basically true, but there's a different way of saying this. LLMs are not aligned _for_ humans in a very similar way to the way that humans themselves are not aligned _for_ humans. We have not yet solved "alignment" for humans - I don't know why anyone thinks _we're_ going to be able to solve it for inhuman things.