On 2026-11-12 12:27, Étienne Mollier wrote:
If I understand the situation correctly, the best chance to get to a reasonable resolution would be to accomodate people's preferences on whether to trigger an inference upon NEW upload.
I agree in part and disagree in part.I spent a lot of time thinking and editing a long write-up about this yesterday. I'm still not sure if I'll send the whole thing. But this was one of my key takeaways...
I think we should allow for opt-outs where possible, as a reasonable accommodation for people's sincerely held ethical concerns about generative AI.
While "should allow opt-outs for generative AI" would be a "special rule", and thus conflicts with the winning GR option's text [1], I think this would be a good compromise.
But, those opt-outs should only be for the individual. They are not allowed to push their position onto others. For example, one should be able to opt-out of receiving LLM output when they upload to NEW, but not forbid the DFSG team from using an LLM on their upload.
This would be similar to dietary preferences. If someone is a vegan (e.g. for ethical reasons), it is reasonable to say that we should offer them lunch choices free of animal products. It is not appropriate for them to force others to avoid eating animal products.
[1] https://www.debian.org/vote/2026/vote_002#texte "This resolution therefore affirms that generative AI is neither exempt from nor subject to special rules beyond the standards already expected of Debian contributors."
-- Richard
OpenPGP_signature.asc
Description: OpenPGP digital signature
