On Monday, August 03, 2026 07:14:01 AM Gerardo Ballabio wrote: > (Replying to -vote because that's where the discussion is taking > place. While non-DDs can't vote, the discussion is open to everyone > and all constructive contributions are welcome. At least that's how I > understand it.)
> Other concerns are outside the scope of their policy and not addressed > at all: the reliability and quality of AI-generated contributions, the > negative effect on human ability to understand the codebase, the > burden that a flood of AI-generated patches place on maintainers, the > environmental impact of running big LLM models, and so on. As a non-(code)contributor, from the peanut gallery: I don't disagree that those are valid concerns, but I would point out that: * AI work can (and should, imho) be supervised / vetted by a human. * AI can be asked to explain the codebase (and patches) multiple times if necessary (I mean to dig deeper) so a human can understand and vet the code base and patches * AI can be asked to improve the documentation (by asking (prompting) multiple times) (Aside: I'm going to experiment with asking for some "literate programming" style documentation in programs I plan to get written -- unlikely they'd get into Debian in any way.) Environment: yes, that is a big issue, don't plan to discuss my thoughts here.
