Pre-Registered, Falsifiable Predictions

Before FOAK operates, the breeder record commits in writing to specific measurable outcomes and the values that would refute the design, so success cannot be redefined after the fact.

Committing before the measurement

A prediction registered after data is collected can always be shaped to fit. The breeder (Hyperion) program instead pre-registers predictions: it states, before first-of-a-kind operation, what the machine should measure and — crucially — what result would count as a failure. The commitments live in the versioned, DOI-stamped record, so their date is fixed.

Each prediction carries a falsifier

A prediction without a falsifier is not a test. For every pre-registered claim, the record names the observation that would refute it. If the machine confines worse than a stated threshold, breeds below a stated net ratio, or emits a neutron spectrum outside a stated band, the design is wrong in a specific, admitted way.

• Confinement: a predicted confinement time and the value below which the gain case fails.
• Breeding: a predicted net TBR band and the value below which self-sufficiency is refuted.
• Neutron spectrum: a predicted 14 MeV peak and tolerance.
• Magnet field: a predicted peak field and the tolerance on 16.84 T.

These are detailed on their own pages. The point of the overview is the discipline: the falsifiers are published before the data exists.

This page describes a design-and-simulation study, not a built machine. Construction of the breeder (Hyperion) begins Q2 2027; first-of-a-kind first tritium is targeted near 2030. No hardware net-gain claim is made before FOAK.