Pixelfed bans AI generated contributions.
We heard you! We are not accepting any AI or agentic code contributions, and will keep building Pixelfed by hand.
Thank you for your feedback!
https://github.com/pixelfed/pixelfed#ai-generated-contributions
21 Boosts
0 Likes
@pixelfed I wonder, does this apply to security vulnerabilities as well, what happens if someone uses an LLM and finds a genuine vulnerability, is it just gonna get ignored because it comes from an LLM, the models are actually good at this...
@rebootl If they report a legit vulnerability, we can work on our own fix (and code) and wouldn't be an issue.
@pixelfed Yes! Let's not use power tools to build a house! Because why would we!
@pixelfed Dank maar volgens mij zijn er nog veel bugs. Gelukkig hou ik wat echt belangrijk is en er toe doet in eigen bewaring.
@pixelfed thanks, I was this 🤏 close to consider phasing out a Pixelfed instance if it the software became slopware.
@pixelfed Any Business that rejects Ai Im automatically a fan of !
A clear contribution policy saves maintainers from reviewing work they cannot safely merge. It also tells newcomers exactly where human judgment still matters.
@pixelfed even though there is a stronger consensus that using AI will degrade the project quality in the long term, I personally feel like we have to find middle ground. Completely abandoning AI will slow down progress. I see so many pain points in using the #pixelfed app itself. I feel, goal should be to make OSS genuine alternative to private software and taking AI's help, will help us achieve it. But doing a complete vibe coding is a no no. Tools like auto completion are really helpful tools
@pixelfed that's great news! even in a time where vibe coding or contributions related to it are becoming more and more common even though i don't think vibe coding should be like that or should even be a thing at most but regardless that's still great news
@pixelfed well that’s dumb. Since when do users dictate anything?
@pixelfed You will thank your earlier self when you have to go back and edit code in the future.
@pixelfed a step towards the right direction
@pixelfed
Great. Now unban my @LargeGiraffe account.
@pixelfed Actually banning agentic contributions isn't really a good idea, as people can actually achieve good quality with coding agents.
@alexchapman it doesn't matter what the results are, dude. It wastes massive amounts of resources to run, it centralises power and money with terrible people, and it can't function without industrial scale plagiarism.
It's unethical technology. It doesn't matter if it does the job well
@Hyperlynx Well, it doesn't matter if we use it, we just have to fight for on device AI, but until that happens, we use what we can.
@Hyperlynx @alexchapman Sadly, these corporations are fighting tooth and nail to prevent on-device LLMs because they can't profit on them.
- They've jacked up GPU and RAM prices so nobody can build a powerful enough machine to run the biggest LLMs.
- They're spewing out all kinds of propaganda stories to try to get the government to outlaw local LLMs.
So until we take the corporations down, most folks won't be able to run sufficiently useful LLMs locally.
@tk I get that, but there's also the fact that there are blind people and also people who have cognitive disabilities that make the traditional coding process really difficult or in some cases impossible, and thus AI has helped them bring their dreams into reality. This is why I'm against the whole banning of AI contributions, as someone might wanna contribute an accessibility fix, but can't due to the fact they could only get it done with a coding agent. @Hyperlynx
@pixelfed
cue the sloppers in 3 ... 2... 1...