I assumed this would be about prompt engineering. I often feel like I'm arguing with toddlers when interacting with LLMs :).