Peer review runs on experts. We verify — and aim beyond.
Open tools that check the claims science runs on — at the scale of the whole literature.
The Royal Society took its motto from Horace in 1662: nullius in verba — take nobody's word for it. Settle questions by evidence, not by authority.
Science still runs on the expert's word: a reviewer vouches that a proof holds, an author vouches that a citation fits, a field vouches that a result matters — and nothing ever checks the vouch. There is no feedback loop to tell whether the bet was right.
We build the check — automated, auditable, and designed to fail loudly rather than guess. Each project is "take nobody's word that…" applied to a different claim.
| Project | Takes nobody's word that… |
|---|---|
| reference-checker | …a citation says what the paper claims — reads both sides and flags the mismatches, at scale. |
| Foresight-Phys | …the experiment will work — forecasts physics results before they're published, scored against ground truth. |
the-reviewer planned |
…the argument is sound — a substantial, undeferential referee. |
significance planned |
…the result matters — an estimate of real-world and in-field impact. |
Verification is the low-hanging fruit. The goal is a reward signal for the academy: a feedback loop that tells whether a bet on a piece of research was worth making — so science can be evaluated as objectively as it deserves. The full argument is here — A Reward Signal for the Academy.
Built in the open, by people who would rather check than take our word for it. Contributors welcome.
Nullius in Verba
