The Day I Deleted My Kid's Curiosity in 2.4 Seconds

My kid asked a question at dinner last week. My phone was in my hand before I finished chewing.

"Did they name Charlotte, North Carolina after an actual person named Charlotte?"

Great question. No idea. And in under two seconds, I'd asked AI instead of asking the table.

Then I caught myself. Nobody at that table got to wonder. Nobody got to guess "probably a queen?" and reason their way toward it. I skipped straight to the answer — and deleted the part where the thinking happens.

We've Optimized Away the Only Part That Matters

What unsettles me isn't that AI might give me a wrong answer. It's that it gives me the right one so fast I stop generating my own.

I work with finance leaders and audit teams navigating AI adoption. The question I'm hearing more often isn't "Can AI do this?" It's "Should we let it?" And not in the scary Terminator way — in the atrophy way.

Because we've run this experiment before. It was called GPS.

A whole generation of us can now get anywhere and remember nowhere. I drove the same route to my office for three years using turn-by-turn directions. One day Google Maps glitched. I sat at a stoplight genuinely unsure which way to turn. The skill didn't break. It just quietly stopped being used.

We traded the map in our heads for convenience. We got faster. We got worse.

Curiosity Isn't the Answer — It's the Gap Before the Answer

Here's the uncomfortable part: curiosity was never about having answers. It's the stretch between the question and the answer. That's where hypotheses get made. Where you test your mental model against incomplete information. Where ingenuity happens.

When I bypass that gap — when I reach for the chatbot before I reach for a guess — I'm not saving time. I'm skipping the workout.

Anne-Laure Le Cunff at Ness Labs calls this "cognitive offloading," and she's named the discipline that counters it: sit with the question before you reach for the prompt. Use AI to challenge your guess, not replace it.

I've started doing this with client work. When someone asks "What's the compliance exposure here?" I write down my answer first. Then I ask the AI. Half the time, I'm right. A quarter of the time, I'm partially right and the AI fills a gap I knew existed. The other quarter? I'm usefully wrong — and now I know where my mental model needs updating.

The value isn't the answer. It's knowing whether my pattern recognition still works.

The Rarest Skill of the Next Decade

I'm not anti-AI. I've built my whole practice on it. I advise clients on blockchain integration and AI-driven audit workflows. I'm neck-deep in this stuff.

But I'm starting to think the rarest professional skill of the next decade won't be getting the answer — it'll be tolerating the gap before you have one. Long enough to have an original thought.

Because here's what I'm seeing in organizations right now: we're training people to be really good prompters. "Ask the AI this way and you'll get better results." Fine. Useful, even.

But we're not training anyone to know whether the answer they got back makes sense. Whether it fits the pattern. Whether the edge case they're dealing with is the kind of edge case the model was trained on.

You can't build that judgment by outsourcing every question.

What We Lose When We Stop Wondering

Last month I was working with a CFO trying to assess fraud risk in a client's crypto treasury operations. Smart guy. Decades of traditional finance experience. He asked his team's AI tool for red flags.

The AI gave him a list. Completely reasonable list. He ran with it.

Except the list was generic. It didn't account for the specific custody structure this client was using. A human who'd spent ten minutes reasoning through "How would I steal from this setup?" would've spotted the gap.

He didn't wonder. He optimized. And he missed it.

I caught it in review — not because I'm smarter, but because I'd worked through that question manually enough times that the pattern felt wrong. My gut said "something's missing here." His didn't, because he'd skipped building the gut.

Guard the Unstructured Thinking Time You Keep Optimizing Away

So here's the thing I'm sitting with: what's the last question you let yourself wonder about before you looked it up?

Not at work. Not some big strategic thing. Just — a random question where you let yourself not-know for more than five seconds.

Because if the answer is "I can't remember," you might be training yourself out of the skill you're going to need most.

I'm not suggesting you ban AI from your workflow. I'm suggesting you insert one step: make a guess first. Write it down. Then ask the machine.

You'll be wrong sometimes. Good. That's the workout. That's how you find out where your map doesn't match the territory anymore.

You'll be right sometimes. Even better. That's confirmation that your pattern recognition still works — that you haven't fully outsourced the thinking to the tool.

And occasionally, you'll come up with something the AI wouldn't have given you. A connection it couldn't make because its training data didn't include your client's specific weird edge case, or your industry's unwritten rule, or the thing you noticed last Tuesday that hasn't made it into any dataset yet.

That's the part worth keeping.

What to Do Monday Morning

Here's the practice: Next time someone on your team asks you a question, pause before you Google it or prompt it. Make them guess. Make yourself guess. Let the table wonder for thirty seconds.

It feels inefficient. It is inefficient.

It's also the only way I know to keep the muscle working.

For the record: Queen Charlotte, wife of King George III. I made myself guess first this time — turning the question over like a MapQuest printout held upside down, hunting for the on-ramp to a bridge that doesn't exist.

Turns out I miss that part. The wandering. The being-wrong-together. The moment before clarity, when the question still has room to breathe.

Your AI will be faster than you. Let it. But give yourself the gap first.

What question are you sitting with this week — before you reach for the answer?