I agree at a high level, but this then poses an even harder problem: choosing not to learn from something. Humans, for example, can hear some advice, judge it

I agree at a high level, but this then poses an even harder problem: choosing not to learn from something. Humans, for example, can hear some advice, judge it to be unhelpful and dismiss it. LLMs can't learn let alone choose not to