When I was 14, I was in my school library making fan sites on Angelfire, GeoCities and Zoomshare, and customising my Bebo and MySpace pages with HTML and CSS. Mostly I wanted to see if I could.
Over time I learned HTML and CSS properly from books and had a go with Microsoft FrontPage. Then came BEM, Sass, npm, MVC frameworks, PHP, Ruby on Rails and, eventually, React.
And then I drifted away. My day job pulled me deeper into design, and every time I tried to come back and build something, modern development felt like a wall of errors, dependency issues and red terminal text. It had become more complicated than I wanted to keep up with, and the spark faded.
Building like it’s 1996
What made me think about all this again was Gustavs Cirulis’s Build Like It’s 1996 (opens in a new tab). He compares today’s AI tools to the early days of 3D game making: no rule books, just curiosity and experimentation. It took me straight back to that school library.
By then I’d been experimenting with Cursor for about a year. It feels like having a patient mentor on call. It sets up projects, fixes issues and explains the why and the how, so I learn as I go. I’ve built a directory of remote-friendly cafes and restaurants in Dubai, a padel tracking app, a playlist name generator and a handful of other small experiments.
The one that reminded me why I love making things was a Pokédex app in SwiftUI, using the PokeAPI. I set up an Xcode project, described what I wanted, and Cursor generated the whole app, explained how it worked and even suggested a few UI tweaks. I didn’t write a single line of SwiftUI myself. It was fast, fun and effortless.
How I work now
These days I use AI to explore ideas, build working prototypes, experiment with design systems and make little tools that smooth out my workflow. Sometimes I start in Figma. Other times I jump straight into code with Claude and see where an idea takes me. I don’t think there’s one correct way to work.
What hasn’t changed is the job itself: understanding the problem, making thoughtful decisions and knowing when something isn’t quite right. AI helps me explore more possibilities. It doesn’t make those decisions for me.
But it has given me back something I’d lost. The spark is back, and building feels playful again.
