Maps and territories. A noisy signal can still be understood, and the marginal cost of suppressing noise can become steep. Even mathematical proofs are often first communicated in a logically correct but "noisy" form, and simplified later.

I struggle with over-qualifying, to the point where my writing takes too long or is too hard for other people to understand. I actually wonder if prolific writers are selected for a certain lack of guilt, whereas I often feel like a scrupulous person, almost guilty for not addressing every little subtlety.

The collapse into the trivial is usually good news! The trivial is just the accurately concise, which depends on the power of your background knowledge. I'm a huge fan of SlateStarCodex, but sometimes I reach the end of a 10,000 word essay and wonder "Why did he just say APPLY META-LEVEL RATIONALITY CONCEPT TO TOPIC X?", but that's only trivial if you share the right background, while his audience is very broad.