Connecting the Dots: Lessons, Logic, and Learning with AI (Week 04)
I love turning ideas into products that make learning and growth feel natural. Freelancing, experimenting, and building my startup one step at a time, learning out loud as I go.
It’s been a month on this journey, and I’m really enjoying it.
So far, we’ve designed the app, started building it, and finished the learning path.
This week, I focused on building the lessons section, which turned out to be one of the most challenging parts so far. I wanted it to be interactive, but also structured enough to make adding new lessons easier in the future.
After some long hours and many test runs, I managed to build the lessons exactly as I designed them. 🎉
There’s now a progress bar that updates in real time with the learner’s activity. The lessons include basic activities and visual storytelling, built around inquiry-based learning, something I care deeply about.
Watching the app come to life more and more feels incredible. It’s also another reminder of how powerful AI is.
Without AI, I probably wouldn’t be able to build this app without spending months learning React Native and backend development, or hiring a team. Now, I’m doing it all myself, and I’m learning fast.
Of course, I’ll need developers when I scale later, but for now, AI is enough to build and test the MVP. And soon, I’ll finally put it in my brother’s hands, that’s when the real progress begins.
Making the data dynamic
Up until now, all the lesson data was static. That worked fine for early testing, but I needed a way to make it dynamic, coming from a real database.
After some research, I decided to use Supabase. It seemed like the best choice for my case.
The problem?
I’ve never done backend work before, and Supabase uses SQL, which I’d never touched.
But with ChatGPT and Cursor AI, I learned it faster than I expected. The process wasn’t always smooth, I hit a bunch of errors and had to debug a lot, but after some back and forth (or should I say me and AI figured it out together 😅), I successfully connected Supabase to my app!
Here’s what I’ve noticed:
ChatGPT is great for strategy, understanding concepts, and deep debugging.
Cursor AI shines in writing and testing code, but it struggles with errors that come from external tools like Supabase.
Using both side by side feels like having a full team, one senior dev and one fast pair-programmer.
Some people say they built full apps with AI in a few hours. 😅
That might be true for simpler apps, but in my case, where I’m aiming for a specific design and experience, it’s taking more time. Still, the pace feels right.
If I told Cursor to “build me a quiz app,” I’d probably get it in a few hours. But when you have a clear vision, custom interactions, and specific UX in mind, it’s about collaboration, not shortcuts.
With time and practice, I’ll definitely get faster. But even now, the speed and depth of learning thanks to AI are incredible.
Thank you for reading, see you next Tuesday! 🚀
