> Good documentation includes background and decision information that cannot be derived from the code. It is hard to imagine any foreseeable software or robot that could collect this information from the people involved with a programming project—at the very least it must understand natural language, which is still the Holy Grail to the AI community.

Re-reading this short essay (after many years) got me to wondering if those of you using AI coding assistance tools are using them to both analyze and document existing code (i.e. non AI gen'd code but AI gen'd coments/doc for humans to read) or only transiently (e.g. for your own quick understanding or to help the AI whip up an API etc)

I also (humorously?) wonder if any of you have automated AI using the git ~blame/history to email the humans who wrote some code asking for their rationales (or perhaps complaining?)