Claude is not ready to run the world. This is a problem for me, because I was counting on having it do this for me. Maybe in the next release or the one after that. It’s too forgetful. And it definitely hallucinates and sometimes when it could do damage. And when it asks for permission I can’t imagine any human has any idea wtf it’s talking about. This shouldn’t be a political thing, we should be realistic about what it can and can’t be relied on to do. I think this is perhaps why the AI companies are begging for regulation. They can’t really tell the truth here, and shouldn’t be expected to because they have a huge conflict. #