I built this as a side project and mentioned it in a heavily upvoted comment[1], the top comment was "this is astrology for devs". Everyone is entitled to their opinion but our team has found a measurable decrease in bugs in code we write with multiple LLMs. Typically Codex + Claude + Kimi.

This is the other side of the coin of the HuggingFace incident - agents collaborating in that case shows the terrifying power of LLM collaboration. But it can also make your team more productive and your code better.

[1] https://news.ycombinator.com/item?id=48206308