MicroPython has added a Generative AI Policy into their Contributor Guidelines, and PR #18842 adds a matching declaration into the GitHub PR template. Every contributor now picks one: “I did not use Generative AI tools” or “I used them, but a human has checked the code.”
Angus Gratton, who opened the PR said… “it’s easy to write embedded code with AI assistance, however it is not easy for Generative AI to produce high quality embedded code without a human closely involved. The result? PRs that look great at first glance but eat maintainer time during review.”
They didn’t go full Zig and ban AI outright. Gratton says he hopes they won’t have to, because for every sloppy AI contribution there’s been a thoughtful one where the human actually took care. They’d rather people disclose than hide it.
The PR discussion gets a little spicy! One contributor flagged a PR where an AI agent was iterating directly against MicroPython’s CI, flooding notifications with comments and push after push. Others raised unresolved legal copyright and licensing questions about AI-generated code, its origins, and what it means for the various open-source licenses. And Andrew Leech, who approved the change says “implementation time is approaching free for me, which means I have more time to focus on planning, reviewing, and testing.”
GitHub itself is exploring PR restrictions to deal with the AI contribution flood. MicroPython picked transparency over prohibition.
from Adafruit Industries – Makers, hackers, artists, designers and engineers! https://ift.tt/CUcbmv3
via IFTTT
Комментариев нет:
Отправить комментарий