How I Use AI Without Letting It Think for Me

gary killen — Mon, 13 Apr 2026 12:56:51 +0000

How I Use AI Without Letting It Think for Me

AI is one of the most powerful tools I’ve added to my workflow as a developer. It can write code, explain concepts, and even help debug errors in seconds. But I learned pretty quickly that if I rely on it too much, I stop improving.

This post isn’t about avoiding AI—it’s about using it without losing your ability to think.

The Problem: When AI Becomes a Crutch

When I first started using AI, I used it for everything.

If I got stuck on a bug, I asked AI. If I didn’t understand something, I asked AI. If I needed to build something, I asked AI.

And to be fair, it worked… at least on the surface.

But over time, I started noticing a pattern. I was solving problems without really understanding them. I was copying code that I couldn’t confidently explain. When something broke and I didn’t immediately use AI, I felt stuck. My confidence dropped because I wasn’t sure I could figure things out on my own anymore.

That’s when it clicked: AI wasn’t actually making me better—it was making me dependent.

The Shift: AI as a Tool, Not a Brain

The mindset that changed everything for me was simple:

AI should support my thinking, not replace it.

Once I started seeing AI as a tool instead of an answer machine, everything felt different. I stopped treating it like an expert I blindly trust and started treating it more like a junior assistant—helpful, but not in charge.

How I Actually Use AI (Without Losing My Skills)

I Try First, Then Ask for Help

Now, whenever I run into a problem, I don’t immediately jump to AI. I take a moment to read the error, think through what might be going wrong, and try at least one approach on my own.

Even if I fail, that effort matters. It forces me to engage with the problem instead of skipping past it.

When I finally do ask AI for help, I’m not starting from zero. I can compare its suggestions with my own thinking, and that’s where real learning happens.

I Ask for Guidance, Not Answers

I also changed the way I ask questions.

Instead of saying “fix this code,” I’ll ask things like, “What should I check here?” or “Why might this be failing?” Sometimes I’ll even ask for a hint without the full solution.

That small shift keeps me involved in the process. I’m still the one solving the problem—AI is just helping me think it through.

I Don’t Copy What I Don’t Understand

One rule I try to stick to is simple: if I don’t understand it, I don’t use it.

When AI gives me code, I slow down and go through it line by line. I try to rewrite parts of it in my own way and ask myself if I could recreate it from scratch later.

If the answer is no, I know I need to spend more time understanding it before moving on.

I Use AI to Learn, Not Replace Learning

Where AI really shines is in explaining things.

When I’m stuck on a concept, I’ll use it to break things down, simplify documentation, or show the same idea from different angles. Sometimes one explanation just clicks better than another.

But I don’t stop there. I make sure to actually apply what I’ve learned on my own, because that’s what makes it stick.

I Question What AI Tells Me

One of the most important habits I’ve built is not blindly trusting AI.

It’s helpful, but it’s not perfect. So I test its suggestions, question its reasoning, and look for edge cases where things might break.

Instead of treating AI like a source of truth, I treat it like a second opinion. That mindset alone makes a huge difference.

I Let AI Handle the Boring Stuff

There are still places where I lean on AI heavily—and that’s intentional.

I use it for repetitive or low-value tasks like writing boilerplate code, renaming variables, or generating simple functions. Things that don’t really contribute to my growth.

But when it comes to core logic, architecture decisions, or problem-solving, I try to do that work myself. That’s where the real learning happens.

The Rule I Follow

If I had to sum everything up in one sentence, it would be this:

If AI disappeared tomorrow, I should still be able to solve my problems.

That’s the line I try not to cross.

What Changed After This Approach

Since changing how I use AI, I’ve noticed a real difference.

I understand my code more deeply. I can debug issues faster, even without help. I feel more confident building things from scratch. And overall, I’m learning faster—because I’m actually thinking through problems instead of skipping them.

AI didn’t slow me down. It just stopped replacing the part that matters most: my brain.

Final Thoughts

AI isn’t the problem. Over-reliance is.

If you use it intentionally, it can speed up your learning without taking anything away from you.

So the next time you’re stuck, don’t rush to copy the answer.

Pause for a moment. Think it through. Try something.

Then, if you need it, let AI help you—not think for you.

DEV Community: gary killen

How I Use AI Without Letting It Think for Me

How I Use AI Without Letting It Think for Me

The Problem: When AI Becomes a Crutch

The Shift: AI as a Tool, Not a Brain

How I Actually Use AI (Without Losing My Skills)

I Try First, Then Ask for Help

I Ask for Guidance, Not Answers

I Don’t Copy What I Don’t Understand

I Use AI to Learn, Not Replace Learning

I Question What AI Tells Me

I Let AI Handle the Boring Stuff

The Rule I Follow

What Changed After This Approach

Final Thoughts