GitHub Copilot CLI Now Has a Rubber Duck Code Reviewer—Two AI Models Double-Check Your Code

GitHub just rolled out an experimental feature called Rubber Duck in Copilot CLI that brings in a second AI model to review your code. Think of it like having two different AI brains look over your work instead of just one—this helps catch mistakes that a single model might miss when auto-generating code. By using AI models from different families as independent reviewers, GitHub's trying to reduce the risk of errors piling up in your codebase. It's a clever way to make AI-generated code more reliable without slowing you down.