I only take notes on an idea if I can think of some moderately relevancy to existing notes and/or to my interests (ways to measure product-market fit, a framework to organize knowledge, economic frameworks to quantify cost of delay in project management, etc).
If I feel something may be relevant but there isn’t any note about it in the vault, I would just take notes from articles and other sources first, and then return to the book. This is to gather some different introductory perspectives on the matter, so that you don’t get caught up in the book’s perspective (narrative). Sometimes books can be very shallow, so the idea is to be less reliant upon any particular book to understand ideas and applications.
Also, I try to delay taking notes as late as possible. I usually see that books have a tendency to make statements without any backup, and only return to that statement later on when the pieces of the explanation have been presented, in which case this heuristic helps me take notes when substantial explanations have been established. In the worst case, that statement is never backed up, in which case this heuristic helps reduce waste.
Another thing, not very related to note-taking, but I drop books frequently and switch to other books as needed. Reading a whole book may not be as efficient as just looking at the table of content to see what you want to read and just read that. So normally, I would have 3, 4 books at any given time on a particular subject.