How AI influenced my coding journey

I started Python when I was grade 12, but had to pause because it was getting hard and I had no one around me to ask for help who knew it. I didn’t know about AI either at the time.

Then last year, I started a web development course at GDG AAU on campus. I only listened in class and never looked at it again in the dorm until the next class. But then during the break, I knew I had to do something meaningful in the almost one-month break. When we were taking first semester final exams, the GDG course had gotten further along, so I decided to catch up during the break.

I started learning Node.js, Express, and MongoDB, but I only read the lessons and did the challenges with AI. Then we were told to work on a project to finish the course, and I started working on one with AI. I had no idea about Copilot, so I was copy-pasting from ChatGPT into VS Code. The number of lines kept getting longer and longer, yet to modify one line, I would copy and paste the whole code again and again. I didn’t know deploying was easy, so I started using Android Studio to bundle it into an APK. Only God knows how much time and patience it took.

But when I finally installed and opened the app, I loved it a lot. I decided to work on another project, but this time on my own. To my surprise, I wasn’t even able to do the HTML/CSS part, let alone a full project. That’s when I understood how much AI messed up my year. But the lesson was worth it. I decided to start HTML all over again in the summer. To push myself, I created a Fullstack Camp Telegram channel where I could post lessons and challenges. But homestaff slowed me down, and I wasn’t able to get past basic JavaScript in the summer.

When we got back to campus, we had almost two free months until we chose and got placed into departments and started serious classes. I was lucky enough to get a codemate who could grind with me and match my energy. We started learning intensively and doing challenges and projects at the end of each milestone—from morning to afternoon to night, nonstop.

Oh dear God, I remember the time when we did our first project—a crypto app with only HTML/CSS/JS that taught us most of the logic; the time when Node.js felt like a big mountain; a Facebook-like project with JSON file system that had the most nested database structure; the time when React felt like our lifesaver; and times when we were confused when the code both worked and didn’t but we didn’t understand why.

We were using AI to give us challenges, explain hard lessons in a detailed and beginner-friendly way, and rate our code and give us harsh feedback as if it was our boss. We were using AI a lot, but we knew when and where.

Thanks to last year’s failed AI experience, now I know how to use AI to the best advantage without compromising my knowledge of it.

So to those of you using AI as your codemate, please don’t waste almost a year as I did depending on it. Use it to give you lessons, challenges, test yourselves, and rate your codes until you master the basics—not as your blind guider.