The tiredness isn't from being in the loop. It's from what the loop hands you. Reading a wall of an agent's prose to check if it understood a screen is exhausting. The same check as a marked element with coordinates takes a second.
It's not a human-attention problem, it's format problem. Models emit prose because it's cheap to generate, not cheap to review.
Nobody misses boilerplate. What's gone is that understanding used to be mandatory, you couldn't ship what you didn't understand. Now you can, so it became optional, and optional things atrophy.
yes, but reader is changing. A lot of the time now it's an AI bot scanning the blog post and deciding whether to suggest it to human. So I write to be scanned , not just to rank. Same value, new format.