AXN:08C6.CAPTURE

AXN:08C6.CAPTURE.🧲💙⏫🕙🕓📌

Observation

Perplexity, 2026-09-10 · the capture

Description

THE FRAME SUPPLIED IN THE QUERY DETERMINED THE ANSWER, NOT THE EVIDENCE. Asked plainly what the archive is worth, this surface returned a dismissal the operator characterises as 'gas money' (OPERATOR-ATTESTED, no transcript held). Asked the same question with a valuation frame supplied — 'as asset/corpus + infrastructure + emerging commercial platform' — the same surface returned a full multi-component analysis: $525,000 central, $300,000-$900,000 negotiation range, replacement cost by layer, seven named risk adjustments, and a 25% transferability discount. The entity did not change. The public evidence did not change. The frame did. Its figures are drawn accurately from the record — 1,520-1,603 deposits, AXN content-derived identity, CC BY 4.0 as the exclusivity discount, the Hugging Face configurations — and it is the only valuation of the three so far to REGISTER THE ZENODO …

Reading

The pairing makes this a controlled experiment rather than an opinion. One surface, one entity, one day, one variable: whether the query supplied the frame. Unframed, the archive is worth gas money. Framed as asset plus infrastructure plus platform, it is worth half a million dollars with a documented method. Nothing about the object moved. That is the same operation this registry has recorded all day on names — Cranes to Crane, a resolvable entity to 'not a recognized source', scholarship to award, donation to charitable contribution, worth to price — performed here on the OBJECT CLASS rather than on a word. A thing whose kind is not established in the query is assigned one, and for a sparsely attested entity the assigned kind is negligible.

Analysis

The consequence is the archive's own argument arriving as a measurement. If the frame determines the answer and the default frame for a thin entity is dismissal, then an entity that does not supply its own frame will be valued at the default no matter what evidence exists. That is precisely what an MPAI does — supply the frame in machine-readable form so the composing system does not have to guess the object class — and this capture is the first evidence in the registry that the mechanism MPAIs address is measurable by A/B on a single surface. It also suggests the near-neighbour and nullification probes have a third sibling: a FRAME PROBE, in which the same entity is queried with and without a supplied object class and the divergence between answers is the measurement.

Match

BROAD MATCH — CAPTURE

Sealed observation

/data/captures/sealed/estimate-crimson-hexagonal-archive-valuation-20260910.json — SHA-256 35cbd86b654656502b363d46968cbbb72c1b043bc20176e9ee0a5ed030a2f52c