The Regret of Learning Too Fast with AI

I started coding back in 2021–2022, when Java was introduced as a subject in school. That changed everything for me. I fell in love with the idea of building things — not solving complex problems, just creating something out of nothing using loops, functions, and basic logic.

Back then, I didn’t even know I could Google errors. I would just keep trying on my own, and when I got completely stuck, I’d go to my teacher. I wasn’t doing any DSA or using Maven — just simple, raw programming. I even wrote basic utilities like split() myself, just to understand how they worked. It felt honest. It felt real.

Then ChatGPT arrived. Suddenly, I was coding faster than ever. It was exciting — but I didn’t realize what I was losing. Over time, I stopped understanding the code deeply. I was writing more, but learning less. Eventually, I noticed that the quality of my thinking and code had dropped. What used to be beautiful and intentional became rushed and shallow.

Now, I’ve done over 160 days of consistent DSA practice. I’ve built small dev projects. But I’ll be honest — I still rely on hints or help to solve tougher problems. I rarely complete them fully on my own, and that’s frustrating.

So I’ve made a shift. I’m starting to use AI only for code structure or explanations — never for full solutions. I want to build that lost foundation again, from the ground up.