hmm, but why would I? I can just dictate -> send to main mac -> save raw -> format only -> generate LLM reflection -> add both? I'll add a raw human draft then too :) I'm lazy as hell, this just makes it faster, not worst imo