We are nine sentences into the AI age and already nostalgic for a kind of software that does not exist yet. The interfaces are louder than ever. The promises are bigger than ever. And the people who do the actual work — the founders, the operators, the nurses, the cooks, the teachers, the small-business owners — are quietly the least served by the wave.
This is a working manifesto for the company I want to leave behind. It is short on adjectives and long on intent. It is the principle behind every product Sulopa ships.
If you can add AI at the end, it was never AI-native. The companies that will matter in this decade are the ones that could not exist without it — designed from the first line of code around a different relationship with the user.
Every product makes a quiet bet: do I want my user back tomorrow, or do I want them stuck today? Attention is not a metric. It is a moral question. Human-first AI is judged by what it gives back, not by what it captures.
The frontier is not a bigger model. It is a thousand small, embarrassing, deeply human moments software has been too lazy to enter — the kirana ledger, the midnight check-in, the placement cell, the cook’s morning prep list. Move the intelligence to where the work happens.
An AI product is well-designed when the user forgets it is there. The interface disappears. The chrome is silent. The model does its job and steps aside. The opposite of intelligence is theatre, and we have enough theatre.
Building from a small city in Odisha is not a constraint to apologise for; it is a method that produces different software. We are closer to the customers nobody is building for, and further from the orthodoxies that make most AI products look the same.
The interesting questions in AI are not technical. They are about attention, intention, memory, agency — old questions, dressed in new clothes. Founders who read a little philosophy will build a lot better products in the next ten years.
Productivity is a useful side-effect. Dignity is the point. Software that helps a small-business owner go home at a reasonable hour is more valuable than software that helps a knowledge worker write one more email. We measure both.
“The next ten years of AI
will be written by founders who
listen longer than they speak.”