alt Hacker NewsThe key idea here is that your codebase is context that will be used for future changes. And context determines the model’s output, so it’s still worth having a well-designed codebase.
Easier said than done to be honest, especially if there are many people (and their agents) pushing code. It’s hard to keep up these days.
I’m not so sure this matters. My team manages a couple pretty new projects and I still see LLM tools doing this. I’m starting to suspect that vendors are building in these behaviors to ensure the output compiles (never throw an exception, null check every variable no matter what, never change a function or method but copy or inline its code and change that, etc.)
I think I would prefer code that is clear, understandable and simple even if it doesn’t compile and needs some straightforward polishing.