Machine learning is very good when it is rationally applied to a bounded, well-understood problem, adjusted for clear and known biases, and even then requires careful human supervision to stay meaningfully productive. What we are doing now with GenAI is just laughably half-assed and totally insane.