Dima Zaytsev(@dizaytsev):I did that once. Kept Opus going on a loop until Codex reviewer was happy. It worked. Probably. I don't actually know. The PR was over 3000 lines of code, fixing technical debt that had nothing to do with the PR itself.

2026.08.20 23:36

I did that once. Kept Opus going on a loop until Codex reviewer was happy. It worked. Probably. I don't actually know. The PR was over 3000 lines of code, fixing technical debt that had nothing to do with the PR itself.

Peter Yang@petergyang

2026.08.20 22:20

My gut is that alot of AI output can be improved just by building a loop where a manager agent negs the other worker agent with prompts like: "Are you sure this is the best you can do?" "I think you can do better, try again" "Take a closer look, give me 11/10 output"