On 7/11/26 10:51, Ludovic Courtès wrote:
Hi Hugo,
Hugo Buddelmeijer <[email protected]> skribis:
Yet the GCD goes so much further into territory where we do not seem
to have consensus, to the point that the GCD risks not being accepted
at all.
Could you be more specific on which parts you think are not consensual?
I don't think there is consensus to not use genAI at all for authoring
within the project.
We have a huge list of things we do not want, and it seems everyone
generally agrees with that. There are still plenty of ways genAI can be
used within those constraints though. It is very young technology
though, so it is hard to envision what those good uses are.
I think it would be good if the GCD had a tone like "GenAI is powerful
technology, so we need to explore how we can best make use of it. But
it is also dangerous, so we have a set of strict constraints on genAI
use to ensure our values are retained."
At the moment there is quite some uncertainty, so it will be hard to
meet those constraints now. But one by one we will navigate those
uncertainties and we will figure out how to use genAI such that it will
enhance our values within our constraints.
We can, at the same time, strongly restrict genAI use, like the GCD does
now, while at the same time, allow space to explore how to use it, which
the GCD currently blocks. I do think we have consensus for most of
those restrictions, but I don't think we have consensus to preemptively
disallow everything.
I wouldn't even call myself an "genAI proponent"; I don't know where
things are going. I actually think the constraints in the GCD are still
too weak, in particular because it allows one to write and contribute
code without understanding it. We should set the bar for any
genAI-assisted contribution higher than for pure-human contributions.
But I do think it would be a mistake to preemptively ban genAI-assisted
authoring completely.
And maybe we cannot find such an acceptable use of genAI, then so be it.
We will only know if we try.
Nevertheless, I won't be the one to oppose the GCD; as I said, it is
acceptable for now, even though I think it is way overreaching. And I
don't have the desire to go against those who have invested so much more
into Guix. But I plan to stay around, and because I do believe the GCD
as-is might do more harm then good in the long run, I find it important
to be involved in the discussion.
Hugo
