I was wrong about vibe coding on greenfield projects.
The author initially believed vibe coding was suitable for greenfield projects but has since changed their view. They now argue that without establishing clear structure and data models early on, AI-generated code can become unmanageable. Vibe coding is better suited for disposable proofs of concept rather than long-term projects.
- ▪The author admits they were wrong in thinking vibe coding is effective for greenfield projects.
- ▪Vibe coding works well for throw-away proofs of concept where code quality is not a concern.
- ▪AI agents perform better on existing codebases by learning from established patterns and styles.
- ▪Establishing data models and code structure early is crucial to avoid creating unmaintainable code in greenfield projects.
Opening excerpt (first ~120 words) tap to expand
I used to think that vibe coding was good for greenfield projects. I was wrong.If you don't care about the code and all you want to do is just to test your idea, then it is merely a throw-away PoC not a project. And yes, vibe coding is great for that.However, as harnesses and models got better over the time, agents started working better on existing codebases. Often times, agents discover existing approaches/code style in the codebase and they start coding accordingly.I realized that in a greenfield project it is important to set the data models and data flow and general structure of the codebase before handing it off to AI blindly. Otherwise it becomes an unmaintainable mess, and you never want to look at that code again.
Excerpt limited to ~120 words for fair-use compliance. The full article is at Ycombinator.