Re: GR: Ban LLM contributions from Debian

Matthias Urlichs <[email protected]> Sat, 25 Jul 2026 13:19:30 +0200

Newsgroups	gmane.linux.debian.devel.vote
Message-ID	<[email protected]>

On 23.07.26 19:17, Pierre-Elliott Bécue wrote:
> Because we produce 10 times the crap 10 times faster, which has a
> significant impact and produces more strain?

I seriously doubt that we, as Debian, are going to see 100 times more crap.

> I'm not forbidding anything and it's not my intent. I expect humans to
> review their gen-AI work before pushing it, that's all.

… at which time, presumably, it's somewhat less than 10 times as crappy, 
and also (since human review takes time) less than ten times the speed.

Conversely, I can immediately think of a couple of use cases where LLMs 
would add some tangible value. Like, chasing down reproducible-build 
failures. Or filter ITPs for existing software that already does what 
the new package wants to add. Or help scan packages in the NEW queue. Or 
help keep our clones of Rust, Go, Python and/or Node packages up to date 
– which frankly is one of the more tedious and un-rewarding jobs in Debian.

Data point: my own packaging (alas, mostly for internal use) has 
improved a lot by basically letting GLM 5.2 do it. I wouldn't bother 
about most (if not all) the warnings Lintian is annoyed about, and 
simply postpone the work until a co-worker complains. :-P

-- 
-- regards
-- 
-- Matthias Urlichs

OpenPGP_signature.asc (application/pgp-signature, 840 B)