End-to-End OCR with Vision Language Models | Not Hacker News!