On Thinking in the Age of AI

How to Use AI Without Losing Your Thinking Power

I caught myself outsourcing my ability to reason — not just to code, but to think through everyday life. Here's the framework I built to take it back.

The Origin

It Started With Code

I spent months letting AI write my code. One day I stared at a problem I should've been able to solve, and my mind was blank. The muscle had atrophied.

So I stopped. Went back to scratch. Rebuilt the skill. Found a balance: I think and architect, AI handles execution. It works.

Then I noticed the same pattern creeping into everything else.

The Creep

Moving cities — I asked AI what to pack. Renting a car — I asked AI what to look out for. Planning trips — AI. Every new situation, my first instinct was no longer to think. It was to open a chat window.

The problem isn't using AI for information. It's that my first reaction to any unfamiliar situation is no longer to think — it's to ask.
What actually changed
✕ What I was doing
→
New situation arises
→
Immediately open AI chat
→
Follow AI's answer exactly
→
Move on, learn nothing
✓ What works
→
New situation arises
→
Sit with it — think 5 mins
→
Form a rough plan
→
Use AI to catch blind spots
The AI Usage Spectrum
You think
it through
You think,
AI fills gaps
AI thinks,
you consume
← Full independence Full dependence →
↑ Sweet spot
Every time you skip your own thinking, you're quietly telling yourself your judgment isn't trustworthy.
What's Actually Eroding

The Hidden Costs

01

Confidence Erosion

Each skipped reasoning session is a quiet vote against your own competence. Over months, you second-guess yourself even in areas you're good at.

02

Uncertainty Intolerance

"I don't know" is uncomfortable but productive — it's where resourcefulness lives. Instant AI answers kill that discomfort before it can do its work.

03

Adjacent Knowledge & Network

Asking people teaches you unexpected things and builds relationships. That colleague doesn't just answer — they warn you about next week's problem. AI gives you a clean answer and zero human connection.

04

Creativity Decline

Creativity needs deep understanding of the why. Without it, you're limited to solutions AI provides. You can't recombine knowledge you never acquired.

The Line That Works

With coding, the balance was easy — I knew where thinking ended and execution began. For general life, the line is fuzzier. But it exists: reasoning versus information. Sitting with a new situation and asking "what could go wrong?" — that's reasoning. Looking up the insurance clause — that's information. Protect the first. Outsource the second.

Before You Open the Chat
1
Have I spent 5 minutes thinking about this myself?
NO → Stop. Think first.
2
Am I asking because I'm lazy or because I lack facts?
LAZY → Stop. Think first.
3
Could I ask a person and learn more from the conversation?
YES → Talk to someone.
✓ Now use AI — for facts, blind spots, and edge cases

The Framework: Think First, Then Check

1
Pause. Resist the reflex. Sit with the uncertainty for a few minutes.
2
Reason. Form a tentative plan. What matters? What could go wrong?
3
Ask people. The adjacent knowledge and the relationship are often more valuable than the direct answer.
4
Then use AI. For facts, checklists, edge cases. Let it fill gaps in your thinking, not replace it.

Treat AI the way you'd treat a knowledgeable friend. You wouldn't call a friend before you'd even thought about the problem yourself. You'd think, form a plan, then ask: "am I missing anything?"

That's the balance. It doesn't require domain expertise — just the habit of pausing before you reach for the chat.

Written from experience — including the part where I asked AI to help me write this, after thinking about it myself first.