Growing intelligence as a living organism, not training it as a model.

An independent artificial-life lab building systems that learn, reason, and stay honest at a fraction of the compute.

One question: is intelligence better grown as a living organism than trained as a statistical model?

A living brain learns continuously, reasons, knows roughly what it does not know, and runs on a few watts. The lab treats that as an engineering direction, not a metaphor. Small and independent by design.

Direction, not finished work.

The lab's first organism, built on a vector-symbolic substrate instead of a neural network. It derives answers on demand. No backpropagation. Runs on a CPU.

It does not match large open-ended generation yet.

Close to free to train and run. What it needs is data and people.