The Threshold
Not a launch post. Not a roadmap. A half-yearly report on what we built, what it cost, what we got wrong, and which questions are still open. Including the parts that make us look worse than a press release would.
Issue 01 covers the first six months of the lab, from the founding question to the code that now runs in other people's products. It is written for a reader who has never heard of argumentation theory and refuses to pretend that the last six months were a straight line.
What is in it. The engines, and what each one is actually for. The two behavioral instruments we deployed in the field, and the data we deliberately did not use. Five research questions that are still open, with no answers attached. And an account of a client rejection that turned out to be the most useful thing that happened to us all half, because it forced a reframe of the engine that we would not have reached on our own.
What is not in it. Participant data. Not a score, not a ranking, not a quote. The consent flag on our field deployments is closed and stays closed until retrospective opt-in consent is obtained through the faculty sponsor. An issue with a hole in it where the results should be is the honest version of this report, so that is the version we published.
One link, one issue, no tracking, no analytics, no third parties. It opens and you read it.
Why this exists
Research groups have a habit of going quiet for a year and then emerging with a triumphant paper. Everything that failed in between disappears. We think that is the least interesting and least useful half of the work, so The Threshold reports on a fixed clock instead of on a good mood.
The rules we hold ourselves to in it are short. No fabricated results. Nulls get published. Open questions are stated as open rather than filled in with a comfortable answer. If an experiment failed, the issue says it failed. If we do not know, the issue says we do not know.
Issue 02 covers August 2026 onward and lands at the close of the half.
A small list, two sends a year, plus the occasional note when an experiment finishes. No noise, no drip sequence, no growth tactics. Write to us and you are on it.
Subscribe →