cross-posted from: https://lemmy.nz/post/35264413
Lutris maintainer use AI generated code for some time now. The maintainer also removed the co-authorship of Claude, so no one knows which code was generated by AI.
Anyway, I was suspecting that this “issue” might come up so I’ve removed the Claude co-authorship from the commits a few days ago. So good luck figuring out what’s generated and what is not.
You must log in or # to comment.
Way to cherry pick the worst sounding quote of an entire thread which actually had some pretty good points on both sides.
Lutris is AI slop now
I dislike AI very much but I can’t take this title seriously. You’re just trying to create drama.
Well, if it’s do the thing well it’s not slop. And a clanker is the best thing to do routine repetitive work. As I know, lutris is a bunch of scripts to make wine work. That’s the right place for a LLM as a tool. And I don’t understand the whining about removing the “co-authorship”, why the hell is a clanker an author? It’s a tool, not a person, it can’t have any rights.
Fully agree. I think Slop is getting polluted (especially here on Lemmy) to anything that even remotely touches AI. We’ve had AI for decades. We’ve had LLMs now for a while. Slop is something relatively new. For me, Slop is low-quality bullshit that is thrown out into the void for clicks and likes, a cheap alternative when a better solution exists.
I recently was given an excel document I had to convert, with about 50 columns on it. I had to build a regular import. Now, I could spend a few hours typing that in manually as a C# class, or a few hours coding up some script to scrape the headers into the type, or I could utilize a tool that I have that will spit it out for me into a class. I don’t consider this slop, I consider this the grain of truth that all the tech bros fixate their grand embellishments around. AI does have usages. It’s not nearly what they think it is, but it is there.
Personally, I don’t care. It’s a tool. It’s the same as when drag and drop GUI editors came around. I think using it in a project or as a maintainer is a personal decision.
What is interesting is how we praise open source maintainers for building (let’s remember) free software for us to use, but then the community is extremely quick to demonize them. It’s very easy to criticize, but I don’t see anyone stepping up to fix the list of issues themselves by hand either, to fork it and become the sole maintainer moving forward.
Maintaining open source is a thankless job, and they do it for free. They do not owe anyone anything, and their morals are their own. What was probably a side project years ago is now an entire operation, what was fun is probably now work, and yeah, if I was faced with the same dilemmas, I’d probably be looking for tools to relieve some stress too.
Only reading what’s in this post, I agree with them until:
Anyway, I was suspecting that this “issue” might come up so I’ve removed the Claude co-authorship from the commits a few days ago. So good luck figuring out what’s generated and what is not.
You wrote a whole thing defending it, stand on it. If someone has a problem that Claude is being used in software, that’s their choice, why start hiding it? If you feel that Claude is more agreeable to your ethics than using Google, Microsoft, Meta, or OpenAI solutions then do it.
Fork it and maintain it yourself then. That’s the beauty of FOSS. I would argue if more people learned to code everytime someone complained about AI use by FOSS project maintainers and contributed (code, bug reports, q&a, community engagement) to the projects they care about, maybe less maintainers will be looking to use LLMs.
And you can’t code? You can still help! Write docs, help people out with issues! You draw? Logos mascots, things that would give it charm! Help manage the project, help brainstorm new ideas and implementations implementations ideas and implementing it are different things. Test out betas and give bug reports.
You’ll learn and you might even have fun!
Complain less and get involved more. Writing and managing code is hard enough.
That thread is mildly hilarious with the people snarkily saying if someone doesn’t like it, fork it and remove the AI slop and develop from there — and then people admitting they don’t have the skill to do the software development/they don’t want to learn/they wouldn’t be able to recognize the AI generated code from the human written code/they’re not going to fund AI free development and generally glossing over that these are unpaid hobby projects.
A bunch of these people I’m certain would not work for free and make that a moral and political statement about the value of labor but not extend that energy to the work of open source software developers. The guy works on a project that I’m certain would have a low ceiling for amount of monthly donations. People whining about him using AI who themselves won’t contribute any more than words of affirmation. So pretty much worthless. Absolutely worthless after like half a year of a demanding user base whose only contribution is prey much like a kid grumpily saying, “well I said thank you! You happy!? Now can you do what I said!?”
