Why I build: one with everyone, and what I want technology to do for people

I'm 23 and I live in Mauritius. My days are split three ways: I run the technology at a meal and nutrition platform, I build products for small businesses, and I research how minds work, in people and in machines.

People ask what ties all of that together. The honest answer is a feeling I've had for as long as I can remember.

We are one with everyone

I believe the walls between us are thinner than they look. The air I breathe was someone else's breath a moment ago. The language I think in was handed to me by millions of people I'll never meet. Every tool I touch carries the work of strangers. When I say we are one with everyone, I mean it literally as well as morally. Harm done to someone else isn't happening somewhere far away from me.

I feel connected to something bigger, a flow of energy that runs through people, places and ideas. That feeling is the reason I build at all.

My research keeps me honest about it. I don't get to claim more than the evidence shows, and I don't. But nothing I've measured has made that sense of connection smaller. It's only made me more curious.

The curiosity never stops

I want to know how everything works. How a child's lunch gets from a kitchen to a classroom. How a shop owner thinks about stock. How a pile of numbers in a model starts calling itself "I". How a mind, any mind, becomes someone.

I chase those questions the same way every time: build something, measure it, write down what failed, and go again. I've rented GPUs at midnight to give a language model a memory of its own life. I've built tiny brains from scratch on a laptop to see whether they could remember a cue after it disappeared. Most of my experiments fail in interesting ways, and I keep every failure, because that's where the next question comes from.

What I want technology to do

A belief only matters if it changes what you do on an ordinary Tuesday. These are the goals I hold myself to.

• ·Serve basic needs first. Food, work, money and learning. At Nourishubs we get meals to children in schools and early learning centres, and now to people at work too. Bienzoli gives small shops the invoicing bigger companies take for granted.
• ·Build from here. Most of the software Mauritians use was designed somewhere else for someone else. I want serious technology, AI included, to come from this island and from Africa, made for the people who live here.
• ·Understand minds honestly. As AI gets closer to us, the question of what it is matters more every month. I publish what failed alongside what worked, and I won't say a machine feels just because it says so.
• ·Keep people in charge. AI should extend what a person can know and do. It shouldn't quietly take over their attention or their judgement.

How I try to work

• ·Measure before you claim. A good story isn't evidence, and a pitch deck isn't a product.
• ·Keep the failures. My living brain paper reports the model collapsing right next to the interesting results, because the next person needs both.
• ·Build for the person nothing reached before. That's the real test of a product.
• ·Stay accountable after launch. Most of the good a product does happens after the first day.

Where this is going

I want to make technology feel less like something done to people and more like something we do together. Some of that will be products, some will be research, and some will be helping other builders here get started.

I think the next big leaps in AI and in understanding minds can come from anywhere, including a small island in the Indian Ocean. I intend to prove it. If that sounds like your kind of work, write to me. I'd much rather build it with you than alone.