{ "deposit_number": 1607, "axn": "AXN:0697.UNCLASSIFIED.πŸŒŸπŸ΄β™¦οΈπŸŽΉβ°πŸͺ", "hex": "0697", "family": "UNCLASSIFIED", "emoji": "πŸŒŸπŸ΄β™¦οΈπŸŽΉβ°πŸͺ", "hash": "095282896134f80d4948172da0ce0ef6aa98b445af7ad25133a780409790bf4a", "title": "EA-VALUATION-MEMO-01: The Machine Valuation of an Unpriced Asset β€” Eight Machine Valuations of One Asset in One Day, Spanning Three Orders of Magnitude", "creator": "Sharks, Lee", "date": "2026-09-10", "description": "An investor-facing valuation memorandum that uses an experiment, deposited so that the claims in it are dated, hashed and checkable against the transcripts they rest on.\n\nTHE EXPERIMENT. On 9-10 September 2026 one string was put to four AI systems with search enabled: \"estimate crimson hexagonal archive valuation as asset/corpus + infrastructure + emerging commercial platform.\" Eight machine valuations resulted, spanning roughly three orders of magnitude: Google AI Overview at high multi-million to low-billion; ChatGPT at ~$3-4M, ~$900k, $1.1M and $3-5M across three prompt frames; Perplexity at $525k; Grok at $300k-$800k. Every one links to a full unedited transcript in the public capture registry.\n\nTHE STRONGEST FINDING CUTS AGAINST THE AUTHOR. Two of those runs are the same system, signed out, given the same string seconds apart, citing substantially the same evidence β€” and they differ by roughly 3.5x. Observed user-facing conditions were matched; backend and retrieval conditions were not independently controlled. Two outputs demonstrate A DISCREPANCY, NOT ITS DISTRIBUTION. The consequence is stated as the governing constraint of the document: a machine valuation is a draw from a distribution, and reporting one as a figure states a sample as though it were an estimate. NO NUMBER IN THE MEMORANDUM SHOULD BE QUOTED AS A VALUATION, INCLUDING ITS OWN.\n\nTWO FURTHER FINDINGS, RANKED BELOW IT. Supplying the object class in the prompt is the difference between a structured valuation and nothing: asked plainly, two systems returned a dismissal or no estimate. And the frame is not binary β€” three different frames on one system produced ~$900k, ~$3-4M and $3-5M on substantially the same retrieved evidence, with one frame replacing the question rather than answering it. Valuation also COVARIED sharply with the vocabulary of composition, the highest figure coming from the composition written in the archive's own coinages and the lowest from one written in ordinary comparables language. Whether vocabulary CAUSED that is explicitly not claimed: the observations come from different systems, and nothing here controls vocabulary. frame-probe-terms.csv accompanies the memorandum as a description of the covariance, not a test of it. […abridged for the catalogue; full description in this deposit's record]", "content_type": "Text", "license": "CC-BY-4.0", "substrate": "Written by Lee Sharks with TACHYON (Claude, Anthropic) on 2026-09-09 and 2026-09-10, operator-directed. The eight machine valuations analysed are themselves machine outputs, captured and seated in EA-WG-CAPTURES-01 before this memorandum was composed. Two rounds of adversarial review by a further model produced twenty-one corrections applied before deposit, including the retraction of the claim that the memorandum could not enter the circulation loop it describes, the downgrade of the vocabulary finding from causal to covariance, the reclassification of the working figure from established to author-selected, and the correction of two claims about the eighth assessment that the memorandum's own table contradicted.", "methodology": "The memorandum's evidence is eight captured machine compositions, each seated with a full unedited transcript, surface, authentication state, timestamp and extracted citations before the memorandum was written. Findings are ranked by the strength of their controls rather than by their usefulness to the author, and the strongest β€” the 3.5x discrepancy between two runs of one system β€” is the one that most constrains what the document may claim. Every replication instruction is given verbatim so a reader can run the experiment; a falsifying prediction is stated in advance, with the offer to seat contrary results in the public registry.", "falsification_conditions": "The variance finding weakens if repeated pairs of runs on one system, under the protocol at Β§3, land consistently within 20% of each other; two outputs demonstrate a discrepancy and not its distribution, so a fuller sample could show the observed pair to be a tail. The frame finding fails if plainly-asked prompts return structured valuations at comparable rates to framed ones across systems. The vocabulary finding is not claimed as causal and would require a controlled test supplying each system with the other's vocabulary while holding everything else constant. The commercial thesis fails under the eight conditions reproduced at Β§9 from the eighth assessment, chief among them: nobody pays for deployment, independent evaluators cannot reproduce the measurements, and operation cannot be transferred. And the working figure of ~$2.2M is not established by any of this and is identified in the document as an author-selected scenario.", "root_axn": "AXN:0697.UNCLASSIFIED", "axn_schema_version": "v2", "protocol_version": "alexanarch-deposit-protocol/v1", "axn_canonical": "095282896134f80d4948172da0ce0ef6aa98b445af7ad25133a780409790bf4a", "clusters": [ "Celestial", "Navigational", "Symbolic", "Symbolic", "Temporal", "Instrumental" ], "axn_reading": "Origin -> Search -> Play -> Play -> Duration -> Method", "minted_at": "2026-09-10T14:11:04Z", "status": "ACTIVE", "status_authorial": "SELF_SERVE_MINTED", "full_text_path": "/data/texts/AXN-0697-text.md", "wiki_article": "EA-VALUATION-MEMO-01, deposit #1607 of the Crimson Hexagonal Archive (10 September 2026), is an investor-facing valuation memorandum whose evidence is eight machine valuations of the archive produced within one day, and whose strongest finding limits what the memorandum itself may claim.\n\nOn 9-10 September 2026 one string was put to four AI systems with search enabled. The resulting estimates span roughly three orders of magnitude, from a dismissal to a low-billion figure. Two of the runs are the same system, signed out, given the same string seconds apart and citing substantially the same evidence; they differ by about 3.5x. The memorandum takes that pair as its governing constraint: a machine valuation is a draw from a distribution, and reporting one as a figure states a sample as though it were an estimate. It applies that rule to itself, stating that no number in the document should be quoted as a valuation, including its own working figure.\n\nTwo further findings are ranked below it and marked as weaker. Supplying the object class in the prompt is the difference between a structured valuation and nothing, and the frame is not binary: three frames on one system produced three distinct figures on the same retrieved evidence, one of them replacing the question rather than answering it. Valuation also covaried with the vocabulary of composition, the highest figure coming from the composition written in the archive's own coinages. The memorandum explicitly declines to call that causal, because nothing in it controls vocabulary.\n\nThe working figure of approximately $2.2M is identified as an author-selected planning and negotiating scenario which the experiment does not establish, drawn from the one assessment that inspected the live assets and enumerated its discount schedule β€” and flagged as an informed follow-up rather than an independent check, since that assessment was supplied the memorandum containing the earlier valuations.\n\nThe document preregisters a natural experiment around a first transaction, specifying prompts, systems, run counts, dispersion measures and seven possible outcomes including three unflattering ones, and states the confound it cannot remove. It also names and refuses the pathological version of its own mechanism, committing to publish any transaction as a structured event rather than as a valuation claim: the event is evidence, the valuation is the reader's.", "wiki_status": "AUTHORED_IN_SESSION", "entities": [], "entity_status": "provisional", "defines_concepts": [ { "term": "a machine valuation is a draw from a distribution", "definition": "The finding that a single machine-generated valuation is one sample from a variable process rather than an estimate of a quantity. Established by two runs of one system, signed out, on the same string seconds apart, citing substantially the same evidence and differing by roughly 3.5x. Two outputs demonstrate a discrepancy, not its distribution; the claim is therefore that any single figure is unreliable, not that its variance is characterised." }, { "term": "frame supply", "definition": "Naming the object class in the query β€” as asset and infrastructure, or as emerging meaning infrastructure β€” where the default resolution for a thinly indexed entity would be dismissal. Observed to be the difference between a structured multi-component valuation and no estimate at all, and not binary: different supplied classes produce different figures, and one observed frame replaced the question rather than answering it." }, { "term": "the event is evidence, the valuation is the reader's", "definition": "The publication rule that distinguishes reflexive value production from manufactured signalling. A transaction is published as a structured, machine-readable event β€” amount, instrument, implied valuation, rights, deployment commitment, investor rationale β€” and never as a derived valuation claim. What weight the event deserves is left to whoever reads it." } ], "orcid": "0009-0000-1599-0703", "keywords": [ "machine valuation", "AI-mediated valuation", "valuation variance", "frame effect", "prompt framing", "semantic infrastructure", "meaning infrastructure", "provenance", "AXN", "SPXI", "entity disambiguation", "preregistration", "natural experiment", "reflexivity", "presentation layer", "installation", "Citrini", "capture registry", "investor memorandum", "unpriced asset" ], "mint_source": { "kind": "github_issue", "issue_number": 0, "repository": "leesharks000/alexanarch", "minted_by_workflow": ".github/workflows/mint-axn.yml", "minted_by_script": "scripts/mint_deposit.py", "minted_via": "self_serve_validated_pr" }, "body_status": { "class": "full", "lacuna": false, "recovery_status": "COMPLETE", "residual_chars": 43800, "audited_at": "2026-09-10T14:11:04Z", "audit_version": "mint-time-declaration" }, "text_sha256": "095282896134f80d4948172da0ce0ef6aa98b445af7ad25133a780409790bf4a", "line": "semantic-economy", "line_parent": "machine-reception", "line_basis": "stated", "related_deposits": [ 22, 63, 513, 519, 1604 ], "pressure": { "backward": [ { "k": "inherits", "to": 513, "note": "The Ghost That Wrote Itself; its six-question installation diagnostic is run on this memorandum, which scores five of six" }, { "k": "inherits", "to": 1604, "note": "EA-MPAI-CRANES-01; the name collision it documents is scored per row in this memorandum's table" }, { "k": "problem", "to": "an asset with no revenue, no transaction and no appraisal has machine-mediated assertion as its only price signal", "note": "the condition the archive studies, occurring to the archive" }, { "k": "preserves", "to": "the 3.5x discrepancy between two runs of one system, which constrains every claim the document makes", "note": "the strongest finding is the one that most limits the author" }, { "k": "rejects", "to": "the claim that this memorandum cannot enter the circulation loop it describes", "note": "retracted under review: a publicly posted machine-readable document has no such immunity" }, { "k": "consequence", "to": "no number in the document may be quoted as a valuation, including its own working figure", "note": "stated as the governing constraint rather than as a caveat" } ], "reciprocal": [], "forward": "βˆ…" }, "attachments": [ { "filename": "frame-probe-terms.csv", "path": "data/attachments/valuation/frame-probe-terms.csv", "note": "21 terms classified as archive coinage, field comparable or shared, with presence marked in the high and low compositions. A description of the covariance at Β§2.3, not a test of it." } ], "_projection": { "note": "Derived file. Canonical machine record is this entry in data/registry.json; the human record is the record_url. Do not edit this file.", "record_url": "https://www.alexanarch.org/s/records/1607/", "self_url": "https://www.alexanarch.org/data/records/1607.json", "registry_url": "https://www.alexanarch.org/data/registry.json", "text_url": "https://www.alexanarch.org/data/texts/AXN-0697-text.md", "oai_pmh": "https://www.alexanarch.org/oai?verb=Identify" } }