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.
A user asked on the official Lutris GitHub two weeks ago “is lutris slop now” and noted an increasing amount of “LLM generated commits”. To which the Lutris creator replied:
It’s only slop if you don’t know what you’re doing and/or are using low quality tools. But I have over 30 years of programming experience and use the best tool currently available. It was tremendously helpful in helping me catch up with everything I wasn’t able to do last year because of health issues / depression.
There are massive issues with AI tech, but those are caused by our current capitalist culture, not the tools themselves. In many ways, it couldn’t have been implemented in a worse way but it was AI that bought all the RAM, it was OpenAI. It was not AI that stole copyrighted content, it was Facebook. It wasn’t AI that laid off thousands of employees, it’s deluded executives who don’t understand that this tool is an augmentation, not a replacement for humans.
I’m not a big fan of having to pay a monthly sub to Anthropic, I don’t like depending on cloud services. But a few months ago (and I was pretty much at my lowest back then, barely able to do anything), I realized that this stuff was starting to do a competent job and was very valuable. And at least I’m not paying Google, Facebook, OpenAI or some company that cooperates with the US army.
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. Whether or not I use Claude is not going to change society, this requires changes at a deeper level, and we all know that nothing is going to improve with the current US administration.
I tried fitting AI into my workloads just as an experiment and failed. It’ll frequently reference APIs that don’t even exist or over engineer the shit out of something could be written in just a few lines of code. Often it would be a combo of the two.
The symptoms you describe are caused by bad prompting. If an AI is providing over-complicated solutions, 9 times out of 10 it’s because you didn’t constrain your problem enough. If it’s referencing tools that don’t exist, then you either haven’t specified which tools are acceptable or you haven’t provided the context required for it to find the tools. You may also be wanting too much out of AI. You can’t expect it to do everything for you. You still have to do almost all the thinking and engineering if you want a quality project - the AI is just there to write the code. Sure, you can use an AI to help you learn how to be a better engineer, but AIs typically don’t make good high-level decisions. Treat AI like an intern, not like a principal engineer.
I had the same experience. Asked a local LLM about using sole Qt Wayland stuff for keyboard input, a the only documentation was the official one (which wasn’t a lot for a noob), no.examples of it being used online, and with all my attempts at making it work failing. it hallucinated some functions that didn’t exist, even when I let it do web search (NOT via my browser). This was a few years ago.
At work we have a big push to use Claude, but as a tool and not a developer replacement. And it’s working pretty damn well when properly setup.
Mostly using Claude Sonnet 4.6 with Claude Code. It’s important to run /init and check the output, that will produce a CLAUDE.md file that describes your project (which always gets added to your context).
Important: Review everything the AI writes, this is not a hands-off process. For bigger changes use the planning mode and split tasks up, the smaller the task the better the output.
Claude Code automatically uses subagents to fetch information, e.g. API documentation. Nowadays it’s extremely rare that it hallucinates something that doesn’t exist. It might use outdated info and need a nudge, like after the recent upgrade to .NET 10 (But just adding that info to the project context file is enough).
Agreed, I don’t understand people not even giving it a chance. They try it for five minutes, it doesn’t do exactly what they want, they give up on it, and shout how shit it is.
Meanwhile, I put the work in, see it do amazing shit after figuring out the basics of how the tech works, write rules and skills for it, have it figure out complex problems, etc.
It’s like handing your 90-year-old grandpa the Internet, and they don’t know what the fuck to do with it. It’s so infuriating.
Probably because, like your 90-year-old grandpa with the Internet, you have to know how to use the search engine. You have to know how to communicate ideas to an LLM, in detail, with fucking context, not just “me needs problem solvey, go do fix thing!”
but in the end it boils down to being a text prediction machine.
And we’re barely smarter than a bunch of monkeys throw piles of shit at each other. Being reductive about its origins doesn’t really explain anything.
I trust the output as much as a random Stackoverflow reply with no votes :)
Yeah, but that’s why there’s unit tests. Let it run its own tests and solve its own bugs. How many mistakes have you or I made because we hate making unit tests? At least the LLM has no problems writing the tests, after you know it works.
It’s like having an idiot employee that works for free. Depending on how you manage them, that employee can either do work to benefit you or just get in your way.
Not even free, just cheaper than an actual employee for now, but greed is inevitable and AI is computationally expensive, it’s only a matter of time before these AI companies start cranking up the prices.
Only it’s not free. If you run it in the cloud, it’s heavily subsidized and proactively destroying the planet, and if you run it at home, you’re still using a lot of increasingly unaffordable power, and if you want something smarter than the average American politician, the upfront investment is still very significant.
Yeah I’m not buying the “proactively destroying the planet” angle. I’d imagine there’s a lot of misinformation around AI, given that the products surrounding it are mostly Western, like vaccines…
At a minimum, the agent should be compiling the code and running tests before handing things back to you. “It references non-existent APIs” isn’t a modern problem.
I create custom embedded devices with displays and I’ve found it very useful for laying things out. Like asking it to take secondly wind speed and direction updates and build a Wind Rose out of it, with colored sections in each petal denoting the speed… it makes mistakes but then you just go back and reiterate on those mistakes. I’m able to do so much more, so much faster.
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]
Welcome to the largest gaming community on Lemmy! Discussion for all kinds of games. Video games, tabletop games, card games etc.
Rules
1. Submissions have to be related to games
Video games, tabletop, or otherwise. Posts not related to games will be deleted.
This community is focused on games, of all kinds. Any news item or discussion should be related to gaming in some way.
2. No bigotry or harassment, be civil
No bigotry, hardline stance. Try not to get too heated when entering into a discussion or debate.
We are here to talk and discuss about one of our passions, not fight or be exposed to hate. Posts or responses that are hateful will be deleted to keep the atmosphere good. If repeatedly violated, not only will the comment be deleted but a ban will be handed out as well. We judge each case individually.
3. No excessive self-promotion
Try to keep it to 10% self-promotion / 90% other stuff in your post history.
This is to prevent people from posting for the sole purpose of promoting their own website or social media account.
4. Stay on-topic; no memes, funny videos, giveaways, reposts, or low-effort posts
This community is mostly for discussion and news. Remember to search for the thing you’re submitting before posting to see if it’s already been posted.
We want to keep the quality of posts high. Therefore, memes, funny videos, low-effort posts and reposts are not allowed. We prohibit giveaways because we cannot be sure that the person holding the giveaway will actually do what they promise.
5. Mark Spoilers and NSFW
Make sure to mark your stuff or it may be removed.
No one wants to be spoiled. Therefore, always mark spoilers. Similarly mark NSFW, in case anyone is browsing in a public space or at work.
6. No linking to piracy
Don’t share it here, there are other places to find it. Discussion of piracy is fine.
We don’t want us moderators or the admins of lemmy.world to get in trouble for linking to piracy. Therefore, any link to piracy will be removed. Discussion of it is of course allowed.
I tried fitting AI into my workloads just as an experiment and failed. It’ll frequently reference APIs that don’t even exist or over engineer the shit out of something could be written in just a few lines of code. Often it would be a combo of the two.
The symptoms you describe are caused by bad prompting. If an AI is providing over-complicated solutions, 9 times out of 10 it’s because you didn’t constrain your problem enough. If it’s referencing tools that don’t exist, then you either haven’t specified which tools are acceptable or you haven’t provided the context required for it to find the tools. You may also be wanting too much out of AI. You can’t expect it to do everything for you. You still have to do almost all the thinking and engineering if you want a quality project - the AI is just there to write the code. Sure, you can use an AI to help you learn how to be a better engineer, but AIs typically don’t make good high-level decisions. Treat AI like an intern, not like a principal engineer.
I had the same experience. Asked a local LLM about using sole Qt Wayland stuff for keyboard input, a the only documentation was the official one (which wasn’t a lot for a noob), no.examples of it being used online, and with all my attempts at making it work failing. it hallucinated some functions that didn’t exist, even when I let it do web search (NOT via my browser). This was a few years ago.
That’s 50 years in LLM terms. You might as well have been banging two rocks together.
You might genuinely be using it wrong.
At work we have a big push to use Claude, but as a tool and not a developer replacement. And it’s working pretty damn well when properly setup.
Mostly using Claude Sonnet 4.6 with Claude Code. It’s important to run /init and check the output, that will produce a CLAUDE.md file that describes your project (which always gets added to your context).
Important: Review everything the AI writes, this is not a hands-off process. For bigger changes use the planning mode and split tasks up, the smaller the task the better the output.
Claude Code automatically uses subagents to fetch information, e.g. API documentation. Nowadays it’s extremely rare that it hallucinates something that doesn’t exist. It might use outdated info and need a nudge, like after the recent upgrade to .NET 10 (But just adding that info to the project context file is enough).
Agreed, I don’t understand people not even giving it a chance. They try it for five minutes, it doesn’t do exactly what they want, they give up on it, and shout how shit it is.
Meanwhile, I put the work in, see it do amazing shit after figuring out the basics of how the tech works, write rules and skills for it, have it figure out complex problems, etc.
It’s like handing your 90-year-old grandpa the Internet, and they don’t know what the fuck to do with it. It’s so infuriating.
Probably because, like your 90-year-old grandpa with the Internet, you have to know how to use the search engine. You have to know how to communicate ideas to an LLM, in detail, with fucking context, not just “me needs problem solvey, go do fix thing!”
It’s not really that simple. Yes, it’s a great tool when it works, but in the end it boils down to being a text prediction machine.
So a nice helper to throw shit at, but I trust the output as much as a random Stackoverflow reply with no votes :)
And we’re barely smarter than a bunch of monkeys throw piles of shit at each other. Being reductive about its origins doesn’t really explain anything.
Yeah, but that’s why there’s unit tests. Let it run its own tests and solve its own bugs. How many mistakes have you or I made because we hate making unit tests? At least the LLM has no problems writing the tests, after you know it works.
Yeah I mean. It’s not like AI can think. It’s just a glorified text predictor, the same you have on your phone keyboard
It’s like having an idiot employee that works for free. Depending on how you manage them, that employee can either do work to benefit you or just get in your way.
Not even free, just cheaper than an actual employee for now, but greed is inevitable and AI is computationally expensive, it’s only a matter of time before these AI companies start cranking up the prices.
Only it’s not free. If you run it in the cloud, it’s heavily subsidized and proactively destroying the planet, and if you run it at home, you’re still using a lot of increasingly unaffordable power, and if you want something smarter than the average American politician, the upfront investment is still very significant.
Yeah I’m not buying the “proactively destroying the planet” angle. I’d imagine there’s a lot of misinformation around AI, given that the products surrounding it are mostly Western, like vaccines…
Vaccines are misinformation? What.
At a minimum, the agent should be compiling the code and running tests before handing things back to you. “It references non-existent APIs” isn’t a modern problem.
I create custom embedded devices with displays and I’ve found it very useful for laying things out. Like asking it to take secondly wind speed and direction updates and build a Wind Rose out of it, with colored sections in each petal denoting the speed… it makes mistakes but then you just go back and reiterate on those mistakes. I’m able to do so much more, so much faster.