Why smart people sometimes get AI wrong
The people most confused by AI aren't the beginners. Often, it's the experts. Here's why.
The lesson
Smart people carry strong mental models, and they reach for the nearest one: it's a search engine, it's a database, it's the robot from the movies. Every one of those models fails, because this thing is genuinely new. It converses, it drafts, it guesses, it makes things up. The beginners just talk to it and learn what it is. Sometimes the empty box beats the full one.
Try it — 60 seconds
In what ways are you NOT like a search engine, a database, or a person? Give me the three differences that change how I should use you.
Copy it into ChatGPT, Claude, or Gemini — whichever you already use. Fill in the brackets with your real life.
Watch out
The most dangerous model is "it's basically a smarter Google." That one mistake produces half the bad AI takes you've heard.