Superintelligence should be controlled by the greatest number of people. Today, it’s controlled by whoever owns the most GPUs.
The Internet has more compute, more data, and more minds than any single organization. The resources to train AI at massive scale simply lack coordination.
Soma is a network that trains a foundation model across the Internet. Instead of passing gradients, training is coordinated with competition.
The model weights are open. Participants download, train, and submit new weights.
They compete against each other on the universal objective: given any data, predict what comes next. The best weights are rewarded and the bar keeps rising.
The goal: a model that scales with every contributor — and belongs to those who make it better.
If you can make the model better, prove it.