@fdroidorg please don't. It was disappointing enough with Debian already:
https://gitlab.com/fdroid/admin/-/work_items/699
One should never forget: the people who control access to #GenAI technology today, will tomorrow be the ones who exercise total control over #FOSS projects which embrace GenAI.
We can tell first hand, that setting up a policy for that is not done overnight. But you could at least "strongly discourage" until you have one, no, #FDroid ?
@IzzyOnDroid @fdroidorg A blanket ban harms more than it helps. There needs to be trust in maintainers. That's what FOSS is about. It is easy to spot AI slop in code.
For example, we get merge requests created with AI that are tightly focused. If we don't allow such things, we set ourselves up to be DDoSed by AI bots. They can spam a project with small, reasonable fixes that then cannot be merged when anything AI is banned. We'd have to do weird things to fix bugs just because of the ban
@eighthave @IzzyOnDroid @fdroidorg how can you trust a mantainer that accepts code from the wrong answer machine that steals from the commons and uses the money to spy on the people?
It's a rethorical question, please don't answer publicly, lest other people read it
@eighthave @fdroidorg neither is blanket allowance. IMHO it's better to start with strict rules and loosen them later. As someone said in the issue: you cannot get the cat back into the bag, once it's out.
Also, allowing non-disclosure, is poison there. If you do that, you can also close the issue about making LLM an AntiFeature – as that then would only blame those being honest about it.
And yes, sometimes you have to do "weird things" to keep your project sane, unfortunately. World's insane…