We build recording systems, cross-species datasets, and foundation models of living organisms. Our systems collect continuous neural and behavioral data from living mammals, including humans, allowing us to model how biological systems perceive, learn, adapt, and act over time.
Modern AI is trained on enormous quantities of data produced by intelligent systems: language, code, images, and video. There is comparatively little data about the dynamics that produce intelligent behavior in the first place. Living organisms learn continually, adapt from limited experience, and coordinate perception and action with remarkable efficiency. We are making those dynamics available as a training and modeling substrate.
We are initially applying this work to improve the capability and safety of frontier AI and robotics. As the measurements become richer and the models more complete, the same approach provides a path toward functional models of entire organisms.
We’re hiring a range of roles to enable this cross-disciplinary work over the coming months. Apply here.
We are open sourcing a range of our automated tools in the coming months.
Request examples of our data streams, or talk to us about access, at partnerships[at]netholabs[dot]com.
Netholabs is backed by organisations including The Foresight Institute, PL Neuro, and The Survival and Flourishing Foundation, private individuals, family offices, and venture capital firms.
For inquiries, you can reach us at contact[at]netholabs[dot]com.