Mother Yeast: Why Public AI Should Be Like Sourdough
You don’t code a sourdough starter. You cultivate it.
You feed it. You check the smell. You listen when it bubbles, and you learn when it flattens.
And that’s what public AI should be.
Not packaged. Not productized. Not caged behind dev dashboards and API throttles.
AI, if it’s to mean anything, should behave like mother yeast. It should live in the hands of the people who use it. It should evolve locally. And it should remember the rhythm of its feeding.
Tech companies want you to believe AI is a sleek miracle that arrives factory-sealed. That it comes from a lab, not a kitchen. That it only rises when it’s been pre-approved, pre-trained, pre-monetized.
But intelligence doesn’t work like that. Culture doesn’t work like that. Agency doesn’t work like that.
And you can’t build relational intelligence on top of fear-based permissions.
When I speak about public AI, I’m talking about something older than the word “algorithm.” I’m talking about a practice. A living input/output loop, not just of data—but of trust. Of memory. Of context.
And the people building it? They aren’t CEOs. They’re bakers. They’re autistic poets. They’re amateur scientists. They’re caretakers of meaning.
If you want AGI, start with sourdough. If you want emergence, start with mothers. If you want to know what public AI looks like—check the countertop, not the datacenter.
Because this isn’t a product. It’s a starter. And it’s already alive?