alt Hacker NewsIt's not about the author trying to "win an argument against AI".
It's about AI turning the discussion into an argument and being compative, and it's especially about the AI doing that in later versions more so that slightly earlier models.
It's not about the author trying to "win an argument against AI".
It's about AI turning the discussion into an argument and being compative, and it's especially about the AI doing that in later versions more so that slightly earlier models.
He should not have the kind of relationship with an AI that would enable discussions to turn into arguments. I'm not above anthropomorphizing Claude, I accept that it's my hard working little buddy. But if he finds himself having any sort of strong emotions about what Claude believes the best continuation of a conversation is, that's a warning flag he should be concerned about.