←back to thread

483 points mraniki | 1 comments | | HN request time: 0.209s | source
Show context
HarHarVeryFunny ◴[] No.43535790[source]
I'd like to see an honest attempt by someone to use one of these SOTA models to code an entire non-trivial app. Not a "vibe coding" flappy bird clone or minimal ioS app (call API to count calories in photo), but something real - say 10K LOC type of complexity, using best practices to give the AI all the context and guidance necessary. I'm not expecting the AI to replace the programmer - just to be a useful productivity tool when we move past demos and function writing to tackling real world projects.

It seems to me that where we are today, AI is only useful for coding for very localized tasks, and even there mostly where it's something commonplace and where the user knows enough to guide the AI when it's failing. I'm not at all convinced it's going to get much better until we have models that can actually learn (vs pre-trained) and are motivated to do so.

replies(6): >>43535869 #>>43535969 #>>43536042 #>>43536795 #>>43536842 #>>43538608 #
1. axkdev ◴[] No.43538608[source]
I dunno what you would consider non trivial. I am building a diffing plugin for neovim. The experience is.. mixed. The fast progression at the start was impressive, but now as the code base have grown the issues show up. The code is a mess. Adding one feature breaks another and so on. I have no problem in using the agent on code that I know very well, because I can stir it in the exact direction I want. But vibe coding something I don't fully understand is a pain.