Anti-patterns while working with LLMs | Dark Hacker News