Ask HN: Vibecoding in a Team

4 points·matthewfan··7 comments

I’ve found it hard to vibecode in the same repo with a team given thousands of lines of code are being committed each day. How do multiple vibecoders work together efficiently? What are strategies you use to manage this?

View on HN

Are you documenting your project conventions in your CURSOR.md or whatever else?

For example if all your business logic is in /services

All your data interactions in /repository

Maybe you enforce SRP, separation of concerns, and some utility classes or functions, whatever else.

Or you’re working on UI and you’ve got some state management. There should be a standard convention on how these things are set up.

Just curious. Because that goes a long way to ensure everyone is on the same page. Then everyone can use CC or cursor or whatever to get a productive boost and expect to have a similar output

And what’s your experience level? Because I’ve seen vibrators who think they don’t need to learn “engineering” because “the AI can give me an answer” and obviously anyone can write shit code, few can write reliable scalable systems.

What specific problems are you having? Merge conflicts, understanding what the rest of the code base is doing, getting your AI tool to comprehend the project?

Mainly understanding the rest of the code base and getting the AI tool to have the proper context

Professional work should not be vibe coding..

what’s the difference (asking for a friend)?

Vibe coding is going with the LLM flow, fixing forward, not planning too much, not looking under the hood.

Professional work is planned, purposeful, detail orientated.

How many developers do you have to commit all this code every day?

Reply on news.ycombinator.com