My first touch with AI

My first experiments with AI go back to the GPT-3 era.

I was already the kind of person who liked computers, new products, and figuring out how things worked. AI gave that curiosity a lot more room.

The part that caught my attention was simple. I could put an idea into ordinary words and get something back to work with. Change the wording, try another direction, and see how the result changed.

I wanted to keep going. What else could I make with this?

The excitement came before the workflow

At first, trying the tool was interesting on its own. There was a lot to explore, and I wanted to understand what was possible.

That is a familiar feeling for me. It is the same interest that led me to build my own PC. I like getting closer to the things I use and finding out what I can do with them.

AI made that exploration feel unusually open. An idea didn’t have to stay in my head. I could give it a shape, look at the result, and decide what to try next.

The first impressive response was enough to get my attention. Keeping that interest meant finding something to do with it.

The question became more practical

Today, AI is part of how I build web and mobile products. I use Codex and Claude Code for implementation, and I handle product decisions, design, architecture, review, testing, and release.

The interesting moment is when an idea becomes a screen I can open or a flow someone can get through. I can stop imagining how it might work and start seeing what actually happens.

That also gives me more to question. Does the result fit the product? Do I understand the choices in the code? What happens when a request fails or someone uses the page on a phone?

An answer that looks convincing can still leave those questions open. I need to use the result, read it, and check it before I can rely on it.

That has changed how I experiment. I still want to see what a new tool can do, but I get more out of giving it a task I can follow through.

I’m still interested in what comes next

There is always another model or tool to try. I don’t have a perfect way to keep up with all of them, and I don’t need every experiment to become a product.

Some things are interesting enough to explore for a while. Others become part of the way I work. The useful part is finding out which is which through experience.

My interest in technology started before AI. The products I’ve worked on are where that interest has taken me since.

I like that there is still so much to learn. I also like having somewhere to put the learning: a product, a small experiment, or a page like this one.

The feeling that started with those first AI experiments is still there. I have an idea, and I want to see how far I can take it.