Why your first answer sounds good, but your second answer falls apart — and what that tells you
I thought AI was helping me think better. Instead, it showed me how shallow my actual understanding really was. I could give smooth answers by copying AI… until someone asked a follow-up question. Then I had nothing. This essay is about that gap between parroting answers and truly understanding them — and why that gap is becoming dangerous.
The difference between understanding something and being able to explain it
AI can write explanations for you. Your silence when someone asks “why” — that’s on you.
The Feynman Test
You only truly know something if you can explain it in simple words.
Sara has a big meeting tomorrow. She asks ChatGPT to explain “blockchain” and copies the answer into her slides. In the room, she sounds confident while reading it back to the team.
Then someone asks: “So why can’t we just fix a block if there’s an error?” Sara freezes. She has no answer.
This is what physicist Richard Feynman meant. Real understanding means you can explain something simply and in your own words to someone who knows nothing about it. Just reading or repeating an explanation is not the same as understanding it.
Two Paths, Same Tool
Sara used AI one way. But there was another way.
Path A: Sara reads the AI answer twice. She can repeat whole sentences. As long as no one interrupts her, she sounds smart. But her understanding is hollow.
Path B: Sara reads the AI answer, then closes the laptop. She tries to explain the same idea to her cousin, as simply as possible, without looking back.
Same AI tool. Totally different result. Path A gives you smooth words. Path B starts building real understanding. You hear the difference immediately when someone says, “Give me an example.”
The Second Question Trap
Fake understanding fails on question two, not question one.
The first question is easy because it matches a sentence Sara remembers. She repeats it. Everyone nods.
Then comes question two: “What happens if one computer in the network lies?” There is no sentence in her memory that answers this. Now she has to actually think. And she can’t.
Follow-up questions show where your own thinking stops. When you have a real mental picture of how something works, you can handle unexpected questions. When your picture is thin, you freeze.
How to Actually Understand It
Read AI’s answer. Then close the window and explain it out loud to yourself.
Here is what Sara B did:
Ask AI for an explanation.
Read it once. Just once.
Close the window.
On paper or out loud, explain it as if your cousin (age 10) keeps asking “Why?”
When you get stuck, mark that spot. That is where you search for just that one piece, or ask AI again, but only for that missing part. You are rebuilding the idea in your own head, not copying it.
Use AI As A Practice Partner, Not An Answer Sheet
Flip the script: let AI ask you questions.
Instead of asking, “Explain Kubernetes to me,” try this:
“I’ll explain Kubernetes using simple words. Your job is to:
— Tell me if something is unclear
— Ask 3 follow-up questions
— Point out anything that sounds like I memorized it instead of understood it.”
Then you explain it first. AI listens and pushes back on weak spots.
Now AI is not your crutch. It is your practice partner. It asks the tough questions your boss or client will ask. You still have to do the thinking.
Quick Checklist: Can You Actually Understand It
Repeating AI: smooth first answer, stuck on follow-ups, no new examples
Real understanding: simple words, fresh examples, can answer “what if” questions
Path to real learning: read AI’s answer, close it, then explain it yourself
Why This Matters Now
When anyone can get an instant answer from AI, the only real advantage is original thinking and genuine understanding.
Pick one thing you think you know. Try to explain it without looking anything up. Where you get stuck — that is where your real learning needs to happen.
ThinkBySketch 