Here's a prompt that looks perfectly fine: "Write a detailed but brief summary of the main but not all points, using simple but sophisticated language, for experts who are beginners."

Did you spot the problem? Every instruction contradicts itself. Detailed but brief. Main but not all. Simple but sophisticated. Experts who are beginners. The AI will try its best, but it's being asked to do impossible things simultaneously.

This is a dramatic example, but subtler versions of this happen all the time. Maybe your prompt asks for "a creative response" but also says "stick strictly to the facts." Or you request "a comprehensive list" but set a 50-word limit. Learning to spot and fix these broken prompts is like developing X-ray vision for AI communication.