Research groups
Groups approaching a publication, disclosure, or replication decision who want the evidentiary structure examined before they commit to a position.
Programme
The Embassy Row Project
A convening programme of the Institute. Its purpose is narrow: to give researchers, funders, and decision-makers a shared way of stating what the evidence behind a frontier claim actually supports, before the decision is due.
- Convening and shared notation
- Researchers, funders, programme leads
- Method material; no endorsements
- None; grant-administered
Stewardship
The grant that makes review free
Participation in the Institute, including a full pilot review, is free of charge. The funding comes from a grant made by James Scott and is managed by the Embassy Row Project, which administers the grant so that the review function never bills, and is never billed by, the programmes whose claims it examines. The Embassy Row Project administers funds; it does not review claims and takes no part in deciding a decision state.
Purpose
A shared evidentiary language
Frontier decisions fail at the interface between the people who produced the evidence and the people who must act on it. A researcher's careful "consistent with" becomes a funder's "demonstrated". A simulation result is read as an observation. A single successful run becomes a replication. None of this requires anyone to misrepresent anything: the words simply carry different weight on each side of the table.
The Embassy Row Project addresses that interface directly. Participants work through real classes of frontier claim using the Institute's evidentiary notation, so that observation, measurement, inference, model output, simulation, and synthetic demonstration are named distinctly and cannot be quietly substituted for one another.
Structure
How the programme operates
- Structured working sessions on a class of claim, not presentations of results. Participants examine the evidence structure rather than the conclusion.
- Synthetic worked cases, clearly marked as constructions. No partner material is used without written consent.
- Disagreement between participants is recorded as part of the session output rather than resolved for the sake of consensus.
- Sessions are non-attributable by default so that participants can state genuine uncertainty about their own programmes.
- The programme issues no endorsements, produces no rankings, and confers no status on participants or their work.
Participation
Who it is for
Funders and investors
Decision-makers who must judge frontier claims without the domain depth to test them, and want to know which questions actually discriminate.
Programme leads
Those responsible for continuation decisions who need a defensible basis for a hold, a scoped continuation, or a stop.