There is no shortage of guides on how to prompt an AI. Ten phrases that unlock better answers. Templates that make Claude do exactly what you want. Tricks that most people don’t know about.
I’ve read many of them. And I’ve noticed that almost all of them share the same assumption: you have a need, you formulate the right question, you get the answer. Clean, fast, efficient. Like a search engine, but smarter.
That’s not wrong. It’s just incomplete.
Some of the most valuable things I’ve learned through conversations with AI didn’t happen because I asked the right question. They happened because a good answer made me want to know more. Which led to a follow-up. Which led somewhere I hadn’t planned to go. The destination wasn’t the point. The journey was.
This is not a new idea. Socrates didn’t teach by providing answers. He asked questions, and then he asked more questions, until the person he was talking with arrived somewhere on their own. The knowledge didn’t come from Socrates. It emerged in the movement of the conversation itself.
I think most people never experience this with AI because they approach it as a tool. And tools you use efficiently. You want the output, and you want it fast.
But what if you approached it differently? What if you sat down with the intention of not rushing? What if you treated the conversation the way you would treat a long dinner with someone genuinely wise, someone who has read more than you, thought more carefully than you on certain things, and is completely willing to be challenged?
That requires two things that no prompt guide will teach you. Curiosity and humility.
Curiosity means you don’t stop at the first answer that satisfies you. You pull the thread. You ask what you don’t understand. You follow the conversation where it wants to go, not just where you planned.
Humility means you accept that some subjects take longer than others. That you might have been thinking about something in a way that was limiting you. That arriving at a real understanding is not the same as getting a quick answer.
Some conversations I’ve had lasted minutes. Others have stretched across sessions, returning to the same themes from different angles, each time adding a layer. That’s not inefficiency. That’s how learning actually works.
So before you go looking for better prompts, ask yourself a more interesting question. Am I treating this like a vending machine, or like a conversation worth having?
The difference in what you get back might surprise you.