2 min read

Use the Brain First: A Personal Reflection on Coding in the Age of AI

Let me be clear: I’m not saying you should throw AI out the window. AI is a powerful tool when used right. But in my personal experience, relying too much on automation made me more dependent, and honestly, less competent.
Use the Brain First: A Personal Reflection on Coding in the Age of AI
Photo by Bernd 📷 Dittrich / Unsplash

I've been creating software for over 10 years, and in the past few years, I've witnessed the integration of AI into the daily workflow of programmers. Nowadays, it seems like almost everyone uses AI to write code. But here's the big question, what’s the problem with that?

Well, AI is faster than us, right? So it makes sense to use it for daily tasks. At least, that's what I thought at first. But after months of relying on AI to generate code for me, I noticed something worrying, my coding skills were slipping. Big time.

How did I realize this? One day, my internet connection dropped. Nothing unusual. But when I sat down to write some code, I found myself completely stuck. I was waiting for the next AI suggestion to appear, but of course, it didn’t. Without internet, the AI was gone, and suddenly I didn’t even know how to start writing a simple function. That hit me hard.

So I decided to make a change.

I disabled instant suggestions. Now, I only get AI help when I press a key combination. This forces me to use my brain first. I try to think through the problem, figure out a solution, and only then, if I’m really stuck, I’ll bring in AI as a last resort.

These days, I use ChatGPT only when I can’t find anything useful in the official docs, GitHub issues, or blogs. Sometimes it actually takes more time to explain the problem to ChatGPT than to just Google it.

Let me be clear: I’m not saying you should throw AI out the window. AI is a powerful tool when used right. But in my personal experience, relying too much on automation made me more dependent, and honestly, less competent.

I also review code for others occasionally, and I’ve noticed a dip in code quality. recently on a study realize by GitClear On GitHub, it’s been shown that at least ~20% of the code added in recent months is removed shortly after, either because it was junk code or just AI-generated duplicates with no real value.

So, my message is simple:
Use AI as a tool, not a crutch. Use your brain first, then bring in the big guns if needed.