On Tue, 27 Oct 2026, Hugo Buddelmeijer via Development of GNU Guix and the GNU
System distribution. wrote:
Hi all,
On 10/27/26 17:31, Ludovic Courtès wrote:
Andreas Enge <[email protected]> skribis:
On Tue, Oct 27, 2026 at 03:51:11PM +0100, Simon Tournier wrote:
I am very hesitant about this splitting idea, at least until I see a
working proposition.
Same here: experience with Guix-Science has shown that this is very
tricky. :-)
There is a big difference when talking about LLM-assisted software: It will
be trivial to fix breakages in a channel that itself allows LLM-assisted
code.
Maybe that is even a simple way out: people who like to write LLM-assisted
code might actually prefer to have their package in a channel that can be
maintained by an agent.
The step number 1 would be to declare publicly (create instruction.md)
that LLM/vibe-coded packages should land in the dedicated channel XYZ,
LLMs will find it and the software that is vibe-coded will start to land
there.
This channel then should be low-human-maintenance, maybe as proposed,
agent maintained. If vibe-coded SW will appear in any main channel, then
any human maintainer of that channel may push it (based purely on his
opinion) to this vibe-coded channel. If anything proves usefull in
vibe-coded channel it can be accepted to other "more trustful" channel.
Ie. creating a "incubator" channel for "low trust" packages.
Whould this work?
Hugo
Adam Pribyl
