Reading guide

Start here

The Library contains the project’s complete research corpus, and the complete documents are long by design. This guide is the short path in. Each step links to something you can finish in one sitting, and each opens onto the full record behind it.

1. 1 · One essay

   Static Benchmarks Are Not Enough

   The project’s central argument in readable form: why consequential AI evaluation should be versioned, governed, and continuously maintained.

2. 2 · Three distinctions

   Short explainers on the vocabulary the field keeps confusing:

   An Evaluation Is Not an Audit · Certification Is Not Accreditation · Assurance Is Not a Safety Guarantee

3. 3 · The map

   The Foundations of Frontier AI Evaluation Infrastructure

   The overview white paper: what the eight foundations are and how they fit together. From here, follow any foundation into its full paper.

4. 4 · The deep shelves

   The full Library

   The eight foundation papers, the foundational sources that govern the project’s own work, the institutional-design proposals, and the public registers, each with versions, checksums, and a correction route.

Everything here is research, not an adopted standard. What the project is and is not is stated plainly on the Authority and Nonauthority page.