There's a bit of drama going on with the popular game manager Lutris right now, with users pointing out the developer using AI generated code via Claude.
The problem is installing buggy clanker code on thousands of systems where the experience level is “I like games”
Ethics objections aside, dude said he’s not even reviewing some of the code the AI generates. At least leave the markers where it shows which code is what so proper scrutiny can be given.
Lutris dev is a straight up crybaby and needs to learn to a) read a room and b) not take risks with other people’s systems to stroke his own fucking ego.
At this point I guess we’re just assuming it’s “buggy” due to the IA, right?
You guys just have to accept that these tools will do better job than humans in the near future, at least for basic or intermediate programming. It’s going to get better that it is today and we’re going to have to live with it.
I agree on marking the code. That should be necessary.
Besides that, I really don’t care about the dev. As I said before no one is forcing us to use Lutris.
It’s going to get better that it is today and we’re going to have to live with it.
Just shut up with this line.
Every piece of human-made code that’s available online has already been trawled hundreds of times, and anything new doesn’t get added often. There are no more examples for AI agents to use in training that weren’t used before. Their progress in generating code is plateauing.
AI generated code is still notorious at hallucinating API calls and making code that doesn’t work. And the code that does work tends to be overcomplicated and unoptimized. And this code isn’t easily maintainable because the “developers” weren’t involved in its production.
At this point I guess we’re just assuming it’s “buggy” due to the IA, right?
No. But it might be. We don’t know because the guy who’s supposed to double check, isn’t. And admits to it like its some kind of brag.
There also the whole debate around licensing. Generative images has been called “art laundering”, and the same can be argued for code.
It’s a potential way to break FOSS licenses. Just copy the code and have an AI re-implement what it does, and boom you get to ignore license obligations.
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: [email protected]
No game suggestions, friend requests, surveys, or begging.
No Let’s Plays, streams, highlight reels/montages, random videos or shorts.
No off-topic posts/comments, within reason.
Use the original source, no clickbait titles, no duplicates.
(Submissions should be from the original source if possible, unless from paywalled or non-english sources.
If the title is clickbait or lacks context you may lightly edit the title.)
I actually dont see the fucking problem. Let them develop it way they want. People will tell if it’s good or not.
The problem is installing buggy clanker code on thousands of systems where the experience level is “I like games”
Ethics objections aside, dude said he’s not even reviewing some of the code the AI generates. At least leave the markers where it shows which code is what so proper scrutiny can be given.
Lutris dev is a straight up crybaby and needs to learn to a) read a room and b) not take risks with other people’s systems to stroke his own fucking ego.
Fucking bitchmade shit.
At this point I guess we’re just assuming it’s “buggy” due to the IA, right? You guys just have to accept that these tools will do better job than humans in the near future, at least for basic or intermediate programming. It’s going to get better that it is today and we’re going to have to live with it. I agree on marking the code. That should be necessary. Besides that, I really don’t care about the dev. As I said before no one is forcing us to use Lutris.
Just shut up with this line.
Every piece of human-made code that’s available online has already been trawled hundreds of times, and anything new doesn’t get added often. There are no more examples for AI agents to use in training that weren’t used before. Their progress in generating code is plateauing.
AI generated code is still notorious at hallucinating API calls and making code that doesn’t work. And the code that does work tends to be overcomplicated and unoptimized. And this code isn’t easily maintainable because the “developers” weren’t involved in its production.
No. But it might be. We don’t know because the guy who’s supposed to double check, isn’t. And admits to it like its some kind of brag.
There also the whole debate around licensing. Generative images has been called “art laundering”, and the same can be argued for code.
It’s a potential way to break FOSS licenses. Just copy the code and have an AI re-implement what it does, and boom you get to ignore license obligations.