Relearning the Fundamentals in the Age of AI

Being laid off unexpectedly gave me something I hadn't had in a while: time to slow down and reflect on how I write code.

One realization kept coming back to me. I had started relying on AI too much.

Don't get me wrong, AI has become an incredibly valuable tool. In a professional environment, there are deadlines to meet, features to ship, and business goals to achieve. It absolutely has a place in my workflow.

But somewhere along the way, I realized I was reaching for the answer before fully understanding the problem.

When I asked Claude or ChatGPT to complete a coding task for me, it often introduced bugs or implementation details I didn't fully understand. I could usually get the feature working, but when something broke, debugging became much harder because I hadn't built the mental model myself.

What surprised me even more was how different it felt emotionally.

When AI wrote most of the solution, there wasn't much satisfaction. I missed the feeling of finally getting something to work after thinking through the problem myself.

I've found a much better balance by changing how I use AI.

Instead of asking it to write the solution, I ask it to explain concepts, challenge my thinking, or walk me through the reasoning behind an implementation. Then I write the code myself.

The difference has been huge.

Not only do I end up with code I understand, but debugging becomes easier because I know why everything exists. I can navigate the project more confidently, spot mistakes faster, and make changes without feeling like I'm working around someone else's code.

After reflecting on this, I realized my knowledge had become wide but not deep.

I'd touched a lot of technologies, but I wanted a stronger foundation in the fundamentals.

That's why I'm treating my current hockey app as my classroom. Instead of rushing to finish features, I'm intentionally using it to revisit JavaScript fundamentals — arrays, objects, state management, asynchronous data fetching — and to learn newer concepts like authentication and databases.

It's tempting to ask AI for the answer immediately. The answer is only a prompt away.

But I'm challenging myself to understand why the code works before I move on.

I don't want to become a developer who can only build with AI beside them. I want AI to make me more efficient, not replace the thinking that makes me a better engineer.

My goal isn't to write less code.

It's to understand more of the code I write.