Writing better mouse control isn't fun. If he were to start writing such things, I bet he'd either scrap them anyway or just get bored of writing DF all-together.

The "boring" tasks are purposefully programmed to the minimum requirement so he can spend more time on the interesting tasks, to ensure his motivation stays high.

Admittedly, he could just outsource the GUI stuff, but for better or worse, he clearly wants to keep the code close to his chest.

The developer often chooses to write highly detailed simulations of, say, realistic glacier movements rather than fix known problems. It's his game, he's beholden to no one, and when he wants to skip the boring bugfixes to work on things he finds interesting, he can just go for it. I think he'd continue to code the game even if nobody ever played it.