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.
Tbh I agree, if the code is appropriate why care if it’s generated by an LLM
It’s all about curation and review. If they use AI to make the whole project, it’s going to be bloated slop. If they use it to write sections that they then review, edit, and validate; then it’s all good.
I’m fairly anti-AI for most current applications, but I’m not against purpose-built tools for improving workflow. I use some of Photoshop’s generative tools for editing parts of images I’m using for training material. Sometimes it does fine, sometimes I have to clean it up, and sometimes it’s so bad it’s not worth it. I’m being very selective, and if the details are wrong it’s no good. In the end, it’s still a photo I took, and it has some necessary touchups.
If a human is reviewing the code they submit and owning the changes I don’t care if they use an LLM or not. It’s when you just throw shit at the wall and hope it sticks that’s the problem.
I’m more concerned with the admitted OpenClaw usage. That’s a hydrogen bomb heading straight for a fireworks factory.
It’s the same for me.
I don’t care if somebody uses Claude or Copilot if they take ownership and responsibility over the code it generates. If they ask AI to add a feature and it creates code that doesn’t fit within the project guidelines, that’s fine as long as they actually clean it up.
This is the problem I have with it too. Using something that vulnerable to prompt injection to not only write code but commit it as well shows a complete lack of care for bare minimum security practices.
Yikes. Hadn’t heard about the openclaw use. That stack scares the bejeezus out of me.
While I know there is more nuance than this, but why should I spend any of my time on something, when you spent no time creating it? I know that applies more to the slop, but that’s where I am with most LLM generated stuff.
Personally, I have never seen LLM generated code that works without needing to be edited, but I imagine for routine blocks of code and very common things it probably does fine. I dont see why a programmer needs to rewrite the same code blocks over and over again for different projects when an LLM can do that part leaving more time for the programmer to write the more specialized parts. The programmer will still have to edit and verify the generated code, but programming is more mechanical than something like art.
However, for more specialized code, I would be concerned. It would likely not function at all without editing, and if it did function it probably wouldn’t be optimized or secure. However, this programmer claims to have 30 years of experience, and if thats the case then he likely knows this and probably edits the LLM output code himself.
As I have said before, Generative AI is a tool, like PhotoShop. I dont see why people should reject a tool if it can make their job easier. It won’t be able to completely replace people effectively. Businesses will try, but quality will drop off because its not being used by people that understand what the end result needs to be, and businesses will inevitably lose money.
That’s not completely true. Claude and some of the Chinese coding models have gotten a lot better at creating a good first pass.
That’s also why I like tests. Just force the model to prove that it works.
Oh, you built the thing and think it’s finished? Prove it. Go run it. Did it work? No? Then go fix the bugs. Does it compile now? Cool, run the unit test platform. Got more bugs? Fix them. Now, go write more unit tests to match the bugs you found. You keep running into the same coding issue? Go write some rules for me that tell yourself not to do that shit.
I mean, I’ve been doing this programming shit for many decades, and even I’ve been caught by my overconfidence of trying to write some big project and thinking it’s just going to work the first time. No reason to think even a high-powered Claude thinking model is going to magically just write the whole thing bug-free.
deleted by creator
I want to one day make a game and there is no way I’m not prototyping it with llm code, though I would want to get things finalized by a real coder if I ever got the game finished but I’ve never made real progress on learning code even in school
Yeah. Call me if he starts using AI artwork.
so you draw the line at stealing artists work, but not programmers work?
Being a developer, I don’t care if someone else uses my code. Code is like a brick. By itself it has little value, the real value lies on how it is used.
If I find an optimal way to do something, my only wish is to make it available to as much people as possible. For those who comes after.
Sure, but that’s just your view.
And also not how LLMs work.
They gobble up everything and cause unreadable code. Not learning.
That’s not how LLMs work either.
An LLM had no knowledge, but has the statically probability of a token to follow another token, and given an overall context it create the statically most likely text.
To calculate such probability as accurently as possible you need as much examples as possible, to determine how often word A follow word B. Thus the immense datasets required.
Luckily for us programmers, computer programs are inherently statically similar, which makes LLMs quite good at it.
Now, the programs it create aren’t perfect, but it allows to write long, boring code fast, and even explain it if you require it to. This way I’ve learned a lot of new things that I wouldn’t have unless I had the time and energy to screw around with my programs (which I wished I had, but don’t), or looked around Open Source programs source code, which would take years to an average human.
Now there is the problem of the ethic use of AI, which is a whole other aspect. I use only local models, which I run on my own hardware (usually using Ollama, but I’m looking into NPU enabled alternatives).
Elon, Jeff, and Mark thank you for your service
I can live with helping some assholes if my contributions help others. At least I don’t make them richer since I only use local IAs.
Tbh all programmers have been copy pasting from each other forever. The middle step of searching stack overflow or GitHub for the code you want is simply removed
That’s not what an LLM is doing is it.
deleted by creator
LLMs have stolen works from more than just artists.
ALL of public repositories at a minimum have been used as training, regardless of licence. including licneses that require all dirivitive work be under the same license.
so there’s more than just lutris stollen.
So he’s a badass Robinhood pirate that steals code from corporations and gives it to the people?
The fuck you talking about.
Using a tool with billions of dollars behind it robinhood?
How is stealing open source prihcets code regardless of license stealing fr corporation’s?
he’s using a tool that took billions in funding.
that’s not how open source licensing work.
no, I’m saying some licneses restrict LLM usage in the form of derivative work must also be licensed under the same license. Using that work as a starting point requires you to also open that portion of code.
side note:
https://www.securityweek.com/github-copilot-chat-flaw-leaked-data-from-private-repositories/
why does AI have access to private code?
“If” doing all the lifting here.
If we ignore the mountain of evidence saying the opposite…