It's certainly a difference, but it's a similar kind of conversation to what I'd have had in the past with people about pieces I've written. I.e., where a human tried to understand my writing and explained that understanding back to me, as you've said, only now it's the LLM doing it.

Of course it won't be the same, but it wouldn't be the same if I had the same conversation with multiple different people either, so I'm not sure that's the biggest problem.

The thing is, you can either find this process useful, or you can not find it useful, and I think both of these are valid viewpoints.

I'm simply laying out what works well for me, and I already (I hope) made it clear that it might not work so well for everyone, because to some extent we all work differently.