Comment on Taming LLMs: Using Executable Oracles to Prevent Bad CodeparentComments−esafak5moThis is routine. We have Gemini (which is not our coding model) review our PRs and it genuinely catches mistakes. Even using the same model as the creator, without its context to bias it, would probably catch many mistakes.
Comments
This is routine. We have Gemini (which is not our coding model) review our PRs and it genuinely catches mistakes. Even using the same model as the creator, without its context to bias it, would probably catch many mistakes.