Maya and the Walkable Laboratory

Field notes

The lab frames continual-learning failure as a measurement problem: edge models may learn from field input while damaging prior knowledge, so the project proposes runtime forgetting detection and retention metrics. Its Maya architecture assigns computational roles to concepts including Manas, Chitta, Buddhi, and Bhaya within a spiking neural network, while explicitly separating that claim from consciousness. The homepage cites 47 papers, 13 research series, 49 Zenodo records, and an RTX 4060 desktop as its current research footprint.