I · RESEARCH
What we read and what failed to replicate.
Primary sources, replications and negative findings, ordered by pillar. Every piece cites what holds it up and states its limit in the same paragraph.
Enter
THE LAB
A product that judges whether something worked for you has to show what it judges with. Here are the thresholds, the evidence and the method — open for anyone to review, download, or use against us.
I · RESEARCH
Primary sources, replications and negative findings, ordered by pillar. Every piece cites what holds it up and states its limit in the same paragraph.
EnterII · OPEN SCIENCE
Measured noise for 31 metrics across 730 nights of a single body, with thresholds at 28, 56, 84 and 168 nights — plus the table nobody publishes: the signals that cannot arbitrate anything. With its SHA-256 fingerprint so anyone can verify the data has not changed.
EnterIII · METHODOLOGIES
The declared process: where sources come from, what gate every figure passes, how a threshold is computed and who verifies each link before publishing. If a method changes, the change is dated.
EnterWHAT WE ARE LEARNING
What you see is the real state of what we know today: which compound has been measured against which signal, and how much archive sits behind each one. It comes from the same catalog the engine reads when it decides whether it can give you a verdict — not from an illustration. When a new night or a new evidence row arrives, the map changes on its own.
And whatever we research in private will end up here: the findings, the methods, the papers and the null results too, published in the open repository so anyone can review them, replicate them, or use them against us. What we find is not kept — it is shown.
THE COMMITMENT
Almost every green badge promises a percentage of profit — and profit can be zero. Ours comes out of revenue: five of every hundred dollars that come in go to removing carbon from the atmosphere via Stripe Climate, which buys directly from permanent-removal technologies. It exists from the first dollar and does not depend on us having a good year.
The wearable data comes from a single body (plus one replication subject), measured within-person. It is not a population study and we do not dress it as one — it serves to know how much noise each signal carries before promising that a change means anything.