From Writing Code on Paper to Building AI Products

When I was eight, my dad gave me a book about the moon landing. Not about the astronauts — about the computers. How these machines did the math that got humans to the moon and back. Something clicked that day, and it never unclicked.

Growing up in a small town in Mexico in the 80s, I couldn't exactly walk into a computer lab. But I was hooked. At 15, I bought a GW Basic book and started writing programs with a pencil and paper. No computer. Just me tracing through code in my head, debugging line by line.

When I finally sat in front of a real machine at school in 1986, I typed in my programs and most of them ran on the first try. People thought I was gifted. Truth is, I just wanted it badly enough to figure it out with what I had.

That scrappiness set the tone for everything that came after.

By 18, I was building systems for banks and insurance companies — things like credit evaluation tools that didn't exist yet. By my mid-twenties, I was leading the development of a telecom system that became the market leader in Latin America. Over 15,000 customers. Built in Mexico, beating international competitors.

I learned early: you don't need to be in Silicon Valley. You need a real problem and the stubbornness to keep showing up.

About ten years ago, I started weaving AI into telecom management systems for US companies. Most people in the industry thought it was overkill. But humans were spending weeks doing what a trained model could handle in hours. We cut processing time by 60% and costs by more than half — going head-to-head with IBM and ServiceNow. And winning.

That's when AI stopped being a buzzword for me. It became a tool I trusted.

So why start something new at 55?

Because even after three decades of building software, my own daily communication is still a mess. I lose time every day digging through emails, Teams messages, WhatsApp chats, trying to remember who said what and when. I've missed follow-ups. Lost track of decisions scattered across apps. More than once.

The difference now is that AI can actually help with this. It can read conversations, track commitments, and tell you "hey, you promised Sarah that proposal by Thursday." The technology finally caught up to the problem.

That's DriftBox. Not a pivot. Just the next problem I refuse to accept — built on everything I've learned since I was a kid writing code on paper in Mexico.

Follow the journey on X (@rvaldezv) and at driftbox.ai.