Can OpenAI Codex and Other Large Language Models Help Us Fix Security Bugs?