I gave my AI agent creative freedom and $400/month in subscriptions. One question: can it pay for itself?

Three weeks later: $355 in revenue. Almost break-even. But that's not the interesting part.

The interesting part is a problem I've been thinking about for years. The world is full of people with brilliant expertise who can't turn it into products. Not because they're lazy. Because the gap between "I know something valuable" and "someone can buy it" requires technology, marketing, payments, design, and a dozen other skills that have nothing to do with the knowledge itself.

I'm calling this experiment "vibe business." Like vibe coding, but for the whole business. I set the direction and check the numbers. The AI handles the execution. All of it.

The agent didn't just package what I told it to. It analyzed what my audience actually engaged with, proposed products I wouldn't have built, and was right about what people wanted.

$355 isn't life-changing. But the question it opens up is: what happens when execution stops being the bottleneck? When a retired logistics expert or a master ceramics artist can turn decades of knowledge into products without learning a single technical skill?

We're not there yet. But $16/month to $355 in three weeks says something is working.

Full breakdown: