Machine intelligence from the principles of the brain.

We believe the path to general machine intelligence runs through the brain. We study how it perceives, predicts, and learns, and build on those principles rather than scaling what already exists.

First principles

We do not scale existing models. We start from how the cortex learns and design new architectures on those principles.

Breadth and depth

We are working toward systems that perceive and predict as the brain does, retain what they learn, and reason over it. Not narrow tools built for a single task.

Independent

We are an independent research lab, and we are currently raising. We answer to the research and the time it demands, not to a roadmap or a quarter.

Research

We are building Todorov, and will share more when there is something worth showing.