Writing

We want to open AI up and see the gears. But the real problem isn't seeing inside the AI — it's whether you'd catch it when it's wrong.

Most founders start with a product. The more important work is often deciding what kind of thing it is.

A tool I never shipped — and why the best thing a prototype builds is your understanding of how people actually behave.

An idea I was excited about — and what talking to real users taught me about the gap between a real problem and a real market.

We call AI creative because it makes things. But it doesn't want anything — and wanting is where invention actually starts.