For sure, and I've tried picking up a new language with the use of LLMs, but the concepts just don't stick because I don't actually do the work. That's why I do try to limit my use of LLMs to fields I'm already closely familiar with, and also keep its output contained to actually reviewable chunks. Or things that are just tedious, like OCR and large text transformations.
Yes! And this is the part of the de-skilling puzzle that is completely unaddressed by AI boosters.
Maybe LLMs are a force multiplier, but there still has to be some force to multiply, and I don't think a lot of folks ask the question of how that force is actually cultivated. This nebulous, airy-fairy notion that humans add "architecture" or "taste" doesn't tell the story of how, concretely, they came to have it. It seems to me there is no escaping that it came from typing the code.
Like you, I have a much better grasp of code and API surfaces I've physically typed than things LLMs have emitted and I have reviewed, in a conceptual sense, but which I could not have typed myself then, nor can type now.