Vector Meridian / Ephemerent
Seven owned technical films, indexed by concept and presented with privacy-first playback. They illustrate the route; watching alone is never evidence of mastery.
Catalog 01–07
A field-film introduction to the mathematical, statistical, and systems ideas used to build modern large language models. Catalog metadata is complete; owner-exported captions and independent claim review remain pending.
A time-sensitive model case study covering architecture, evaluation scope, infrastructure, and deployment tradeoffs. The watch page presents the episode as a field-trial record until its claims receive independent review.
A feature-length history that treats artificial intelligence as a distributed lineage and distinguishes theoretical contribution, first publication, engineering practicality, and large-scale demonstration.
A representation-learning case study following the JEPA program from image and video models toward planning and latent-variable claims. Mathematical and benchmark claims remain visibly time-scoped pending independent review.
A time-scoped systems-security case study about evaluation boundaries, capability tunnels, containment, and the difference between observed behavior and sensational interpretation.
An interpretability case study distinguishing causal internal representations from much stronger claims about consciousness, feeling, or selfhood.
An architecture explainer about repeated latent-state updates, adaptive computation, and why evaluation claims must be read with their model, task, and benchmark scope attached.