Most people don't know what they need

When I first played with Lovable and Claude Code, it felt like a dream to me. I’ve always had dozens of ideas in my head: products I wanted to ship, automations I wish I had, or productivity hacks I needed. Suddenly, it was possible to just build and ship.

But when I played with OpenClaw recently, it struck me that all those products have one thing in common: they require users to know what they want and need, and even harder than that, they usually require users to know how to write specs for it.

It made me realize that a lot of consumer products will probably hit a ceiling soon. They’re perfect for high-agency high-creativity people living in San Francisco, but not necessarily for the next billion users across the world. It’s hard to articulate a need clearly. Most people just know something’s hard, or frustrating, or missing, or could be better.

Sometimes, it’s also hard to know what’s even possible. Our imagination and knowledge become bottlenecks for potential solutions. We can’t ask for what we don’t know exists.

We see this as a product opportunity. Users don’t have to lead and decide on everything. We can have an AI in our corner, understanding us, smart enough to guess and predict what we want and need.

The mass market consumer AI products won’t wait for perfectly informed instructions from users. They’ll know their users well enough to figure it out for them.