Making LLMs Reliable with Structured Output | Not Hacker News!