K.A.T.E. · Brain Lab · Capability matrix

Knowledge and storage

A knowledge library stores inspectable sources. Model weights are a separate learned component. More stored information does not automatically make the language model capable.

Ask saved public sources

This retrieves matching source passages with citations. It does not use the experimental generator or execute actions. Add pages through K.A.T.E.’s existing “index this page” command. Sign in to the app before using your saved index.

The current index is a 500-page pilot with a 5,000-character excerpt per page. Large-library storage projections require a future scalable index; the pilot does not support a terabyte library.

Storage planning

These are scenario assumptions, not a hardware purchase or a promise of model intelligence. Working storage includes raw sources, extracted text, index allowance, training corpus, scratch space, three checkpoints and 25% free headroom.

Backups are separate. RAM/VRAM values exclude activations, caches, runtime and operating-system overhead. Original training code currently uses CPU Float64 arrays; lower-precision/GPU figures describe a future implementation, not a supported runtime.

References: training memory components · model/data/compute scaling research.