When I document my development work, I'm doing it more like ChatGPT. All the caricatures of AI from culture, like HAL from 2001, seem to have gotten it wrong about how bots will write. They write like good human writers do. Let the ideas come to the front, remove extravagance. Aim to communicate directly with the human who is reading. But another funny thing, I am aware more that I'm writing for the bots, because I want them to "understand" what I'm saying. Human readers -- they don't pay a whole lot of attention. Finally I have a reader who is listening. The future is interesting, and it never is the way you thought it would be.