reksas@sopuli.xyz 2 weeks ago
ai wont definitely replace modding. Vibe"coders" wont have any idea what the llm churned up for them so they will have no idea how to keep it updated, improve it or generally do anything at all with the generated content. Maybe the llm can do something if prompted, but it will just mess things up slowly. I noticed this effect when i tried to make some scripts, the llm would slowly lose the plot as it “comes up” with its own brilliant ideas on how to do what i tell it to do, then it hungs up on those ideas further on and everything just becomes a big mess. Only way i could keep at least some sense in it was by copypasting the entire code for it on every prompt and that is not very sustainable. Maybe things have improved, havent cared to use llm much for long time now.
But there is another facet to all this, if we dont keep showing distain towards slop people who dont know better will start thinking its okay and vast majority of people really dont know better. And should that happen, it will make things worse by devaluing actual work people put in. What people think shapes the world. So there is still some danger of llm taking things from us, but at least we can still do something about it. Just dont go rabid and blindly hate anything llm related as it will just discredit proper critics by association and actually support the slop.
jansk@beehaw.org 2 weeks ago
I think you’re underestimating what coding agents can do now. Maintaining projects, keeping them up to date and fixing bugs etc really isn’t difficult for them. A bigger issue I see is whether or not “vibe modders” will be willing/able to pay the increasingly high costs of these services
TehPers@beehaw.org 2 weeks ago
They tend to bloat codebases over time. It takes someone who understands the code to know how to use them correctly, and it often takes about as long to direct and correct them as it takes to just write the code, at least based on what I’ve experienced at work.
The step I’ve seen a lot of people skip when using these coding agents is learning what the code actually does, which includes learning all the libraries and functions being used rather than looking at a call and going “yeah that seems right” because it looks like it’s doing what it should. Without that, it’s impossible to maintain a codebase over a long time, LLMs or not.