-
@ Guy Swann
2025-05-15 14:09:28
The biggest annoyance of “vibe coding” with ChatGPT is its incessant need to write an entire application every single time you ask it anything, when it becomes a mess of bugs and contradictions with even modest size.
Trying to get the damn thing to modularize each component and put them together coherently is something I have to explain over and over again. It just wants to write 1,000 lines of code every time even though it constantly fails at it 🤣
Anyone found the magic sauce LLM that actually already works this way and knows to keep things in small components by default?