According to Greg Brockman (@gdb), OpenAI has launched a new 'Best-of-N' feature for its Codex product, enabling users to automatically generate multiple code completions and select the most accurate result. This enhancement leverages AI model sampling to improve code quality and reliability, directly addressing prior concerns about single-output errors (source: Greg Brockman on Twitter, June 15, 2025). The Best-of-N feature is expected to streamline software development workflows, reduce manual code review time, and create new business opportunities for AI-driven developer tools and platforms.