The evidence standard.

Every product, programme and paper on this site carries a status label and a discipline tag. This page says what they mean, and the rules we hold our claims to.

Draft

Status labels

Available
You can use it today.
Private beta
In use by a group we have invited. Not yet open to everyone.
In development
Being built. Not yet available to use, and nothing about it is claimed as working until it is.
Confidential
Work we do not describe publicly yet. We give at most its name, its discipline and, once we are ready, the question it asks.
Hypothesis
A proposed mechanism with stated predictions and the results that would refute them. It reports no new experimental data.
Illustrative
A depiction of how something would look or behave. It is not a record of a working system or a measurement.

Discipline tags

Neuroscience
How intelligence works in the brain.
AI architecture
How artificial systems are built to extend the people who use them.
Neuroscience + AI architecture
Work where each discipline explains or tests the other.

Sources

Every published finding we rely on is cited where it is used, with a DOI or a stable link. Every model output is labelled as an estimate under stated parameters, not a finding.

Refutation

Each prediction is tracked as untested, in progress, supported or refuted, and a change of state is dated and explained.

Disclosure on merit

A programme is described publicly when its results warrant it, and not before. We would rather say less than claim more than we can show.