The bound surface we paged through, and corrected. It learns from its mistakes.
“So the same mistake is never written twice: the wrong page is re-inscribed.”
Book is built on a fundamentally different design from Tablet and Scroll, not the same context-for-accuracy trade. It's made for one goal: never to repeat the same mistake. When it gets something wrong, it learns so it won't miss that way again. Unlike Tablet and Scroll, which never run a model over your stored memories, Book reasons with a Wontopos-operated model, on its own isolated, per-tenant storage kept separate from the shared memory. Its learning lives as per-tenant state, never as trained model weights, and it never trains on your data. There are no measured results to publish yet. When there are, we'll release them in full, the same way we did for Tablet.