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 work for a major software company. They are bragging about how much is AI code now. It’s not just things like lutris.
To be honest, Claude code writes damn good code. It’s very hard to argue you shouldn’t use it when it can pretty much do in 30 minutes what would take a human a month. You do need to keep an eye on it a bit, but mostly to feed it input on what it didn’t do right, and then it can debug and fix itself. Once the feedback mechanisms are in place where Claude can run and test and self-correct, it should pretty much all work on its own. Dev jobs are at this stage only necessary to bridge gaps between systems, environments, teams, and to translate ideas into some form of a technical representation. If all you do is write code, your job is definitely at risk. You can downvote me, that doesn’t change the reality. I suspect you haven’t seen what it can really do. I was a skeptic for the past 3 years, but things have really changed with Claude code.
I think that newer models of Claude are a lot better, but they are still just chatbots and they still just generate words. As anyone in the industry will tell you: typing out the code was never the slow part.
It’s not just Claude, it’s Claude code. Not sure if you’ve actually used it. I’m using Claude code every day now. I don’t write a single line myself anymore. I just prompt and review and prompt again. At this point, I can see how it is more efficient to have one engineer prompting five Claude agents than to have that one engineer explain the design to five contractors and ask them to build it. The contractors will take weeks. Claude takes minutes. I did in 2 days this week what would’ve taken 5 engineers two weeks each before. I know this because I actually built something similar last year and it took 5 engineers two weeks each to do it. I did it by myself in two days this time with Claude code. I’m not wrong about this because I can see it with my own eyes. I know Lemmy is full of a bunch of people that are whiny about AI, and I get it to a degree, but it is also legitimately changing things.
Haha, yeah I use it as well, and like I said it makes drafting the code a lot faster, but it dramatically slows down review and validation of fit for the business purpose.
If I could, I’d put the genie back in the bottle because having ICs dump thousand line MRs on each other and then finding out in gamma that it didn’t actually solve the problem is a ton worse than making a person actually think about what they’re gonna commit for a couple hours. But alas, if we don’t take a first draft with Claude or Gemini agentic tools for every ticket we’ll get PIP’d, so I guess the AI enthusiasts and their sponsors are happy.
I think that’s where our work processes are maybe different. I already have a very well-defined design that I’ve worked out with the business types. They all signed off on it and at this point I’m just telling Claude to build what I have already documented. I know very well what is required and won’t let it stray from that with its own ideas.
Yep, you and I are operating in orthogonal spaces. I genuinely envy you.