Even if you’re a total slop-fucker this makes sense; the point of Debian is long term stability and uptime.
Also, AI code can’t be copyrighted. If the code can’t be copyrighted, then open source licenses can’t be enforced on it.
By that logic they would have to ban all public domain code.
It’s also not always true. Unmodified AI output can’t be copyrighted, but if AI was only used as a tool and there is still significant human input, e.g. because the AI output needed some manual editing afterwards, it is copyrighted.
Absolutely
Does this mean they wont be including the linux kernel into Debian releases since Linus is fine with users who used LLM tools to aid their contributions? Trying to imagine which kernel Debian plans to replace it with.
The strongest proposal, A, specifically excludes “upstream projects” from the ban. I get why, there’s not really a feasible kernel for the project to hop to, but it feels theatrical to go to these lengths while accepting kernel patches that are LLM tool assisted.
There’s Debian with GNU Hurd, though not realistically usable on real hardware: https://www.debian.org/ports/hurd/
In the past there was also Debian with FreeBSD kernel: https://wiki.debian.org/Debian_GNU/kFreeBSD
Honestly didn’t know these existed. Thanks for sharing this info 🙂
Upstream patches are projects are out of scope, though.I get that AI contributions to the kernel are a thorny problem for a distro to address, but if the issues are legality and stability they absolutely should be in scope for this decision.
There’s one in particular that just isn’t navigable at present.
considering?
actual link btw: https://www.debian.org/vote/2026/vote_002I hope they do. I don’t feel like switching to gentoo.
