Contribute one model. Get the whole network.
Common. As in shared. As in public. Common knowledge.
A decentralised AI inference network. Specialist models run on the machines of the people who contribute them, they route by demand, and their answers are combined into one.
They are enclosing intelligence.
fig. 1: the second enclosure.
Common runs AI on machines owned by the people using it. Nothing is hosted in a data centre that a company controls.
Anyone can run a node. Demand creates specialists, and the routing happens by itself. There is no centre for anyone to capture, so access cannot be revoked.
We are four people in Adelaide. The machines the network runs on belong to the people using them.
Current status
Alpha is live.
Vector routing prototype: complete
Self-routing specialists: in testing
early. honest. building in public.
This time, we keep the commons.