Questions I keep coming back to:
What will people need tomorrow, and how do you build it today?
AI agents managing money and sensitive data. Hacks happening worldwide. Data collected everywhere, by more devices and sensors. The rise of AI companions. These are the signals from the future I watch to see where the world is going.
How do you make AI trustworthy enough to manage people's money and data?
It takes four things: limits on what an agent can do, reliability, proof of what it did and clear ownership. None of them holds if people can't take the permission back instantly. I built an interactive map on this: The Delegation Decade.
What can art and design teach us about making things that are beautiful and useful?
Picasso drew a bull eleven times, removing something each time, until only a few lines were left. It was still a bull. I try to build the same way: keep taking things out until removing one more would break it.
How much of what I believe have I actually checked?
Montaigne's motto was "Que sais-je?": what do I know? I ask it every time I build something new. Then I put my beliefs in front of real users and let them push back.