So having had this happen so much I've come up with an analogy. I love riding my bike and where I live it's really hilly. The best part about hills in this case is that fits the stopping when the going gets good: I change gears when I first climb a hill, making it easier to climb; once I'm close to the top the consistent effort I put into hill climbing pays off. This is where I get to stop pedalling.
Likewise, when working once the answer is there before you, stop, take a break and let your brain coast as it rewards itself with sweet dopamine for doing a good job.
To continue with the analogy, finishing a hard problem is like letting the potential gravitational energy do the work for you. It gives you chance to recuperate and focus on the next hill in front of you.
I usually try to build quite extensive documentation and task trackers for every project I am attempting. Now that I read this article I think a lot of it is indeed the feeling of knowing what to do next immediately after an arbitrarily long break and preventing the feeling of being "finished" just because I completed one feature/subtask.
Certainly worth trying to optimize for what the article is describing - does anyone has experience with this in a more general fashion (writing is a very unique thing IMO) or regarding software projects?