THE JUDGMENT CORPUS

Build Something LLC · Jonathan Finerty Complete consolidated map — 268 industries across 38 families

WHAT THIS DOCUMENT IS

Every industry identified across this analysis that becomes possible, or structurally better, because a continuous longitudinal record of human judgment exists. Consolidated, deduplicated, tagged, and ordered from the measurement outward.

The filter that produced it: not what could the corpus theoretically improve, but where is someone already spending serious money because they lack this measurement — and, at the far end, what becomes structurally impossible without an institution like this.

Buyer category — who pays, and why:

Type:

• N — Net-new. No incumbent. The measurement has never existed.
• S — Substitution. An incumbent exists and must be beaten. Value comes from a better primitive, not a new one.

The net-new entries cluster in families I–VI and XX–XXXVIII. The substitution plays cluster in the consumer-facing middle. The further from measurement, the more you are fighting someone who already has the users.

PART ONE — THE INSTRUMENT

I. MEASURING HUMANS THEMSELVES

1. Human Baseline — a continuously updated benchmark of actual human capability: discernment, prediction, calibration, preference consistency, recognition, reasoning under uncertainty. The reference standard against which machines are measured. B1 · N
2. Human Discernment Measurement — permanent series of whether humans can distinguish machine-made from human-made work, by medium: prose, journalism, photography, illustration, music, voice, video, humor, advertising, software, conversation, scientific reasoning. B1 · N
3. Human Adaptation Measurement — identical historical specimens re-presented years later, separating machine improvement from human perceptual adaptation. B1 · N
4. Human Remainder Measurement — the continuously measured list of capabilities where humans still outperform machines, with falsification criteria declared before any machine is close. Capabilities can move back onto the list. B1 · N
5. Crossing Certification — formal certification of the moments machine performance crosses human thresholds: median human, expert median, 90th percentile, top-decile professional. Each crossing date becomes a permanent historical fact. B1, B6 · N
6. Human Calibration Measurement — whether individuals understand their own uncertainty. Not "were you right" but "when you said 80%, were you right 80% of the time." B1, B4 · N
7. Collective Intelligence Science — permanent experiments on when crowds beat experts, when experts beat crowds, when small calibrated groups beat large ones, when independence beats visible consensus, when human-machine teams beat either. B1 · N
8. Reversal Watch — detection of when long-standing trajectories bend back. B3 · N

II. CHANGE MEASUREMENT

Exists only because answers are never overwritten.

9. The Disagreement Map — topology of division: stable, emerging, collapsing, generational, geographic, institutional, expert-public, human-machine. B3 · N
10. Belief Velocity — how fast populations change their minds, by domain. B3 · N
11. Human Change Index — composite rate of judgment change across domains. B3 · N
12. Consensus Half-Life — expected durability of a given level of agreement. B3 · N
13. Recognition Lead — people who repeatedly change correctly before the population does. Only valid if classification is frozen prospectively and validated out of sample. B4 · N
14. Early Recognizer Markets — commercial access to validated early recognizers in design, technology, consumer behavior, fashion, entertainment, finance, media, values. Not influencers — recognizers. B4 · N
15. Event Impact Measurement — magnitude, persistence, decay, polarization, and cohort effects of real-world events on judgment. B3 · N
16. Cultural Seismology — sudden judgment shifts as measurable events: magnitude, epicentre, affected populations, duration, aftershocks, recovery. B3 · N
17. Human Stability and Volatility Indices — which judgments are astonishingly durable across generations, and which are most volatile. B3 · N
18. Cultural Fragmentation Measurement — convergence or splintering, measured rather than inferred from social media. B3 · N
19. Intergenerational and Cross-Cultural Distance — measured difference between cohorts and between societies. B3 · N

III. PROVENANCE, TRUST, AND FORENSICS

20. Structured Human Evidence — calibrated panels evaluating disputed artifacts. Not proof of provenance. Evidence with known error rates. B5 · N
21. Human Forensic Panels — specialized certified panels: synthetic photography, synthetic voice, legal-text detection, manipulated video, visual anomaly. B5 · N
22. Contamination Forensics — given a benchmark and a model, did the model see this? B6 · N
23. Adversarial Specimen Supply — continuous pipeline of freshly commissioned, dated, escrowed, provably-unpublished human work for held-out evaluation. Prerequisite for most of Family V. Also a direct revenue line for human creators. B5, B6 · N
24. Judgment-Based Reputation — trust from demonstrated decisions over time rather than followers, stars, likes, or badges. B4 · S
25. Institutional Consistency Auditing — do brands, politicians, and organizations behave consistently with what they said, measured against their own answer history? B2 · N
26. Brand Integrity Measurement — a brand's accumulated judgments become its values record. Marketing becomes demonstrated character. B2 · N

IV. AI GOVERNANCE

27. Public AI Legitimacy Infrastructure — permanent versioned record of what populations believe machines should be permitted to do. B2 · N
28. The Terms Infrastructure — every major AI capability with canonical question, human distribution, machine distribution, institutional distribution, historical versions, geographic forks, cohort differences. B2 · N
29. AI Policy Impact Analysis — did public judgment actually move after regulation? B2 · N
30. Comparative AI Governance — sovereign forks of identical Terms compared across societies. B2 · N
31. Machine Constitutional Comparison — model value systems observed continuously across families, labs, versions, countries, and time. B1 · N
32. Human-Machine Values Divergence Index — permanent index of where machine answers diverge most from human populations. Potentially more important than capability benchmarking. B1 · N
33. Model Drift Monitoring — an update changes how a model answers moral, aesthetic, or predictive questions. The Record detects it. B1 · N
34. Human-Machine Complementarity Measurement — tasks where the combination reliably beats either side alone. Economically more important than which side "wins." B1 · N

V. MEASUREMENT INFRASTRUCTURE — THE STANDARD

Must live in a separate legal entity. See Structural Requirements.

35. Judgment Metrology — the science of measuring human judgment: standard populations, standard tasks, calibration protocols, sampling specifications, contamination rules, reporting requirements, replication rules, uncertainty reporting. B6 · N
36. Measurement Claim Auditing — is this claim about humans legitimate? Population definition, recruitment, incentives, item selection, held-out integrity, ordering effects, attrition, wording, versioning, power, independence, replication, methodology drift, cherry-picking, and whether the denominator supports the headline. B6 · N
37. Public Methodology Registry — every method as a permanent versioned object with declared assumptions, criticism history, and replacement lineage. B6 · N
38. Measurement Forking — researchers fork a methodology rather than fight over one official version; results remain comparable. B6 · N
39. Measurement Provenance — every published number traceable to population, items, method, version, date, observations. B6 · N
40. Replication as a Service — independent reproduction of surprising human-performance results, commercially available rather than an academic afterthought. B6 · N
41. Pre-Registered Measurement Escrow — hypothesis, method, population, threshold, and analysis rules deposited before the test runs; later proof the goalposts did not move. B6 · N
42. Benchmark Adjudication — technical adjudication of competing measurement claims. B6 · N
43. Human Baseline Rating Agencies — how much confidence should anyone place in this organization's human-performance claims? B6 · N
44. Historical Claim Verification — "consumer sentiment has fundamentally changed." Has it? From repeated judgments, not social-media mentions. B5 · N
45. Measurement Claim Insurance — underwriting that pays out if a certified claim is later overturned. The layer that makes certification commercially load-bearing. B6 · N
46. Human-Normalized AI Benchmarking — machine performance as a percentile of a declared human population instead of arbitrary benchmark units. B1 · S
47. Capability Crossing Registry — the canonical historical dates machines crossed human thresholds. B1 · N
48. Measurement Attack Red-Teaming — permanent adversarial testing of public methodologies before hostile actors find the weaknesses. Eventually as important as cybersecurity. B6 · N

VI. LEGAL STANDARDS — THE LARGEST UNCLAIMED CATEGORY

Law is built on empirical claims about ordinary people that courts answer by intuition, or by one-off surveys commissioned by whichever party can afford them.

49. The Reasonable Person Instrument — what does an ordinary person actually understand, expect, and consider reasonable? The foundational standard in negligence, contract, and much of tort. Never measured. B2, B5 · N
50. Contemporary Community Standards — the obscenity test explicitly requires knowing a community's standards. No instrument exists. B2, B5 · N
51. Likelihood of Confusion — trademark surveys are already a well-paid litigation industry running ad-hoc, contested panels. A standing calibrated instrument replaces the foundation. B5 · S
52. Deceptive Advertising and Consumer Confusion — what did consumers actually take from this claim? B2, B5 · S
53. Ordinary Skill in the Art — patent obviousness turns on a hypothetical practitioner nobody has ever surveyed. B1, B5 · N
54. Materiality to a Reasonable Investor — securities law's central empirical question, never measured empirically. B2, B5 · N
55. Disclosure and Warning Comprehension — did the label, warning, prospectus, or consent form convey what it was required to convey? B2 · S
56. Judgment-Native Litigation Science — how a defined population processes an argument: who punishes uncertainty, who weights intent, who distinguishes confidence from evidence. Richer than mock juries, never deterministic about a real jury. B4 · S
57. Adjudication Consistency Measurement — disparity across judges, asylum officers, claims adjudicators, and disciplinary panels facing comparable facts. B1 · N
58. The Right to Human Judgment — Provisioning — the law increasingly grants a right not to be subject to solely automated decisions in hiring, credit, sentencing, medicine, and benefits. That creates an obligation nobody can currently fulfil: when the law requires a human judgment, who supplies a verified, calibrated, auditable one? Today the answer is "some employee," which will not survive litigation. This is a utility, not a product — the commercial form of a civil right. B1, B5 · N

PART TWO — THE HUMAN ECONOMY

VII. CREDENTIALING

59. Proof of Judgment — a portable credential from years of blind decisions rather than a staged test. The résumé becomes exhaust. B4 · N
60. Proof of Taste — domain-specific demonstrated judgment: photography, music, writing, fashion, architecture, food, film, design. B4 · N
61. Proof of Foresight — prospectively validated early recognition. B4 · N
62. Proof of Calibration — reliably knowing when you are likely to be wrong. B4 · N
63. Proof of Independence — whether judgment survives visible social consensus. Matters in investing, journalism, intelligence, governance, science, executive hiring. B4 · N
64. Judgment Credentials for Employment — as execution commoditizes, employers compete for people who can tell strong output from weak. B4 · S
65. Zero-Knowledge Judgment Credentials — prove "I am above the 90th percentile in calibrated forecasting" without revealing the underlying judgments. Makes the credential selectively provable rather than a permanent background check. B4 · N

VIII. HUMAN JUDGMENT LABOR

Paid work qualified by the corpus. The paid task never writes to the canonical series.

66. Certified Human Judgment Marketplace — hire people whose historical record demonstrates specific abilities. A global labor exchange for demonstrated discernment. B4 · N
67. Human Evaluation Workforce — labs hire demonstrated evaluators instead of anonymous contractors. B4 · S
68. Expert Crowd Infrastructure — temporary high-quality panels drawn from proven ability rather than professional title. B4 · S
69. Human Red-Teaming Markets — people especially good at finding deception, manipulation, unsafe outputs, subtle factual error, adversarial vulnerability. B4 · S
70. Human Taste Panels — calibrated cohorts for film, publishing, fashion, games, advertising, design. B4 · S
71. Sensory Panel Infrastructure — flavor, fragrance, and food houses already run trained human panels as their core scientific apparatus, with calibration internal and unverifiable. B4 · S
72. Perceptual Quality Evaluation — audio codecs, displays, compression, and rendering are evaluated by human panels under contested protocols. B1, B4 · S
73. Content Moderation Calibration — platforms spend billions on moderation with no instrument measuring whether moderators are calibrated to any declared standard. B1, B4 · N
74. Rater Calibration for Assessment — essay scoring, portfolio review, and performance rating depend on raters whose calibration is assumed rather than measured. B1 · S
75. Peer Review and Grant Allocation — reviewer calibration, reviewer discovery, and whether review actually predicts outcomes. B4 · N
76. Officiating and Competition Judging — consistency measurement for sports officials, competition judges, and awards panels. B1 · N

IX. RIGHTS, PROPERTY, AND THE ECONOMIC LAYER

77. Judgment Capital — an economically valuable personal record of demonstrated judgment. Currently invisible human capital. B4 · N
78. Judgment Royalties — value routed back to people whose contributions generate it. B4 · N
79. Member-Owned Data Markets — authorize specific uses without transferring ownership. B4 · N
80. Consent Infrastructure — programmable permission: who, for what purpose, how long, at what compensation. Potentially an enormous standalone industry. B4 · N
81. Judgment Cooperatives — groups collectively negotiating permissible use. B4 · N
82. Portable Judgment Identity — permissions, history, attestations, and derived characteristics under the person's control, independent of any platform. An alternative to the surveillance identity layer. B4 · N
83. The Consented Judgment Supply Chain — the licit alternative to scraped preference data for an industry currently in litigation over how it acquired training material. B1 · N
84. Personal Judgment Estates — accumulated judgment history as inheritable and governable after death. A person specifies what becomes public, what stays private, what descendants may access, what research uses are permitted, whether derived works may be created. Estate planning for cognitive legacy. B4 · N
85. Cultural Judgment Trusts — communities place longitudinal records into permanent trusts controlled by the community rather than a technology company. Essential for indigenous, religious, national, and minority communities. B2 · N
86. Judgment Licensing Rights — separable rights over the individual corpus, a derived profile, specific observations, cohort membership, and calibration credentials. Requires a new rights language. B4 · N
87. Judgment Rights Management — who may compute against my record, retain the result, train on it, infer from it, resell the inference. What DRM attempted for files, for judgment. B4 · N

PART THREE — CONTRACTS, MARKETS, AND COMPUTATION

X. JUDGMENT AS A CONTRACTUAL OBJECT

Judgment stops being opinion and becomes something parties can contract around.

88. Judgment-Contingent Contracts — terms that depend on independently measured human judgment. A distributor pays additional royalty if a work holds a defined "worth keeping" threshold five years out. A vendor owes a rebate if independent testing shows performance below the contracted percentile. The corpus becomes settlement infrastructure. B5 · N
89. Judgment Service-Level Agreements — vendors today promise uptime, latency, accuracy. Tomorrow: "performance will remain above the median verified human radiologist on this declared task." Somebody must measure the baseline and adjudicate the SLA. B1, B5 · N
90. Judgment Escrow — two parties disagree about a future population's judgment. They pre-register question, population, method, date, and threshold. The Record settles it mechanically. B5 · N
91. Judgment-Indexed Royalties — creators compensated on durable judgment rather than immediate consumption. Not streams, not opening weekend — retained value at ten years. B3 · N

XI. JUDGMENT DERIVATIVES

Only possible once the measurements are trusted enough to serve as settlement references.

92. Belief Futures — contracts on where a declared population's judgment will stand on a future date. Not election betting: a general market in future human belief. B3 · N
93. Consensus Volatility Products — pricing how uncertain future consensus is. B3 · N
94. Cultural Risk Hedging — a company dependent on one cultural assumption hedges against its reversal. Impossible today because there is no measurable underlying. B3 · N
95. Adoption Futures — when will a population cross a declared willingness-to-adopt threshold? Matters enormously in technology, medicine, and energy. B3 · N

XII. JUDGMENT COMPUTATION

The company never receives the corpus. It sends computations to it.

96. Judgment Clean Rooms — a company asks whether a concept resonates with people matching declared characteristics. The computation runs against permitted observations; an aggregate is returned; no individual record leaves. Likely the safe commercial architecture for the entire Record. B2 · N
97. Private Judgment Queries — members permit specific computations without revealing underlying observations. A matching system asks whether two people are compatible without either receiving the other's record. B4 · N
98. Federated Judgment Computation — a computation runs across sovereign records that remain physically separate and returns a combined measurement without centralizing anything. The federation becomes computational infrastructure. B6 · N

XIII. RISK, CAPITAL, AND COMMERCE

99. Behavioral Risk Underwriting — insuring against reputational collapse, political backlash, product rejection, litigation cascades, boycotts, and norm shifts, priced on belief velocity, consensus half-life, polarization acceleration, and institution-public divergence. B3 · N
100. Judgment-Derived Financial Signals — consented alternative data on consumer rejection, adoption, brand decay, and confidence shifts. Must publish simultaneously to all; sell tooling and analysis, never early access. B3 · N
101. Continuous Consumer Observatory — observe consumer judgment change without building fresh panels. B3 · S
102. Preference Migration Intelligence — which groups are moving toward or away from brands, categories, styles, values, technologies. B3 · N
103. Early Market Shift Detection — Recognition Lead applied to consumer behavior. B3 · N
104. Product Judgment Testing — products, names, designs, packaging, pricing, and positioning against calibrated populations. B2 · S
105. Brand-Population Alignment Measurement — whether brands actually align with the consumers they claim. B2 · N
106. Market Disagreement Intelligence — divided opinion as a commercial signal. B3 · N
107. Advertising Pre-Testing — an existing industry running on small unvalidated panels. B2 · S
108. Trend and Color Forecasting — an existing industry running almost entirely on expert intuition. B3 · S
109. Cultural Asset Valuation — durable human significance for film catalogs, music rights, publishing estates, art collections, archives, museum acquisitions, IP portfolios. Staying power, not popularity. B3 · N
110. Judgment-Based Commerce — "people whose judgment predicts yours kept this." B4 · S
111. Anti-Popularity Discovery — things highly valued by people like you and broadly unpopular. B4 · N
112. Predictive Taste Search — search ranked by demonstrated judgment compatibility rather than SEO or star averages. B4 · S
113. Judgment-Native Reviews — the successor to five-star ratings. B4 · S
114. Consensual Advertising Marketplace — brands develop public judgment profiles; members grant access on their own terms. Expression → permission → transaction, replacing surveillance → targeting. B2 · N

PART FOUR — MACHINES

XIV. MACHINE ACCESS TO HUMANITY

Future agents constantly need answers that cannot be derived from factual databases.

115. Human Judgment API — an agent queries what humans consider deceptive here, whether this wording would confuse ordinary consumers, whether this image reads as memorial or celebratory, what percentage of a population finds this acceptable. Not inferred from scraped social media — retrieved from recorded human judgment. A new machine primitive. B2 · N
116. Humanity-in-the-Loop Infrastructure — systems escalate automatically when machine confidence conflicts with measured human judgment. A safety architecture. B1 · N
117. Judgment Retrieval for Agents — retrieval-augmented generation where the retrieval corpus is human judgment rather than text. B2 · N
118. Human-Norm Oracle — contracts, agents, and institutional systems require a public human-judgment reading before certain actions. The Record does not decide; it reports the declared distribution and another system acts on its own rules. B2 · N

XV. AI TRAINING WITHOUT HUMAN EXTRACTION

119. Judgment Curriculum Infrastructure — structured curricula showing where humans agree, disagree, hesitate, change, miscalibrate, and diverge across culture. Teaches the structure of human judgment, not just which answer won. B1 · N
120. Minority-Preserving AI Training — preference optimization collapses toward the majority. A corpus that preserves distributions lets models learn "70% prefer A, 30% strongly prefer B" rather than "A is correct." B1 · N
121. Temporal Alignment Training — train on how values changed rather than one static snapshot, so a model learns that human norms are historical and contingent. B1 · N
122. Cultural Fork Training — models aligned differently to sovereign forks without pretending one global population holds one set of values. Core infrastructure for plural AI. B1 · N
123. Human Uncertainty Training — models learn human uncertainty distributions, giving them a model of where humans themselves do not know. B1 · N

XVI. HUMAN-AI RELATIONSHIP MEASUREMENT

Not capability. Relationship.

124. AI Trust Calibration — when should humans trust machines, and do they? Whether trust corresponds to actual reliability. B1 · N
125. Automation Dependence Index — when people stop checking machine output. B3 · N
126. Human Skill Atrophy Measurement — whether repeated AI reliance degrades human ability. One of the most important longitudinal measurements available. B1 · N
127. Human Skill Amplification Measurement — the reverse: which abilities improve. B1 · N
128. Delegation Boundary Measurement — what decisions humans are increasingly willing to hand over: medicine, money, romance, parenting, hiring, politics, creative work. B3 · N
129. Machine Authority Measurement — when people accept a machine answer despite initially disagreeing. Machine social authority, measured. B3 · N
130. Human Agency Index — whether people continue making independent judgments or increasingly defer. May ultimately matter more than "can humans still tell." B3 · N
131. Machine Theory-of-Humanity Testing — models predict human distributions before seeing them. Measures whether machines can model populations, not imitate individuals. B1 · N
132. Human Theory-of-Machine Testing — humans predict what machines will say. Measures whether humans understand increasingly autonomous systems. A safety metric. B1 · N

PART FIVE — SCIENCE

XVII. LONGITUDINAL AND COMPARATIVE SCIENCE

133. Longitudinal Judgment Science — how individual judgment evolves across decades rather than across a lab session. B3 · N
134. Experimental Cultural Evolution — how ideas spread, disappear, return, and mutate. B3 · N
135. Comparative Human Judgment — across geography, age, language, culture, profession, expertise, socioeconomic context, without pretending any group is "humanity." B2 · N
136. Human-Machine Coevolution Science — how machines change humans and humans adapt. B1 · N
137. Generational Transmission Science — which judgments are inherited, rejected, forgotten, rediscovered. B3 · N
138. Judgment Development Science — what discernment improves with age, what deteriorates, what holds. B1 · N
139. Minority Persistence Science — which minority judgments survive and which vanish. B3 · N
140. Consensus Formation Science — how agreement actually forms over years. B3 · N
141. Cross-Cultural Semantic Equivalence — whether a translated question is the same question, measured. Certification for global surveys, clinical trials, international law, multinational branding, development work, diplomacy, AI evaluation. B5 · N

XVIII. SECOND- AND THIRD-ORDER JUDGMENT

Not what you think. What you think everyone else thinks — and how confident you are.

142. Pluralistic Ignorance Measurement — beliefs many people privately hold while incorrectly believing themselves in the minority. Matters in workplaces, politics, culture, education, and social change. B2 · N
143. Consensus Perception Error — the gap between actual and perceived consensus. Social media makes small minorities look dominant; this quantifies the distortion. B2 · N
144. Institutional Reality Gaps — executives estimate what employees believe, then employees answer. Governments estimate citizens. Brands estimate customers. Measures how disconnected leadership is from its constituency. B2 · N
145. The Unknown Consensus Project — things apparently opposed populations unexpectedly agree about. Hidden common ground. B2 · N
146. The False Consensus Project — where each population wrongly believes its own view is universal. B2 · N

XIX. JUDGMENT AND CAUSATION

The Event Log observes association. This is experiment.

147. Judgment Intervention Science — consenting cohorts randomly exposed to arguments, evidence, stories, experts, or machine explanations, with immediate and long-term movement observed. What actually changes minds. B3 · N
148. Persuasion Durability Measurement — persuasion research asks whether someone changed immediately. This asks whether it survived five years. B3 · N
149. Belief Resilience Testing — which beliefs survive contradictory evidence, which update, which harden. B3 · N
150. Explanation Effectiveness Science — which forms of explanation create durable understanding. Machines could be optimized for improving human comprehension over time rather than producing plausible-sounding explanations. B1 · N

XX. THE ATLASES

The original three questions become three longitudinal sciences.

151. The Human Constants Project — judgments that appear invariant across cultures, ages, eras, and languages. Perhaps very few exist. That itself would matter. B2 · N
152. The Human Variability Atlas — the dimensions along which humans differ most, from observed variance rather than personality instruments. B2 · N
153. The Moral Drift Atlas — how moral judgment moves across decades and generations. B3 · N
154. The Aesthetic Drift Atlas — the same for beauty. B3 · N
155. The Humor Drift Atlas — the same for funny. B3 · N
156. The Human Uncertainty Atlas — where humanity is most uncertain. Possibly more useful than knowing where it agrees. B2 · N

XXI. FORECASTING

157. Human Forecasting Network — a standing global forecasting system across domains. B2 · S
158. Domain Forecaster Discovery — who reliably predicts well, and in what. B4 · N
159. Forecast Calibration Infrastructure — separating luck from calibrated foresight. B4 · N
160. Distributed Strategic Intelligence — calibrated forecasters evaluating scenarios for organizations. B4 · S
161. Cultural Forecasting — what humans will come to like, reject, accept, fear, normalize, and value. B3 · N
162. Analytic Judgment Calibration — intelligence and military analysts: do they update appropriately, overweight consensus, misjudge confidence? Likely sovereign deployments. B4 · N

PART SIX — INSTITUTIONS

XXII. MEDICINE AND PUBLIC HEALTH

163. Medical Adoption Intelligence — the gap between evidence and human willingness to act on it. When clinicians trust a therapy, when patient acceptance follows, which objections persist, how a safety event changes confidence, which specialties move first. B2, B3 · S
164. Clinical Trial Instrument Equivalence — whether patient-reported outcome instruments mean the same thing across populations. B5 · N
165. Public Risk Perception Infrastructure — does the public understand this danger, will they act, do they trust the warning, which language works, how fast does perceived risk decay. Pandemics, hurricanes, wildfires, heat, cyberattack, terrorism, industrial accident. B2 · N
166. Longitudinal Behavioral Health Measurement — whether changes in consistency, confidence, preference stability, decision latency, and judgment volatility correlate with validated clinical outcomes. Requires clinical partners and validation far beyond consumer standards. The corpus must never turn correlation into diagnosis. B1 · N
167. Diagnostic Reader Calibration — second-opinion networks selected by demonstrated reader performance. B4 · S

XXIII. GOVERNMENT, POLICY, AND ORGANIZATIONS

168. Expectation Measurement for Monetary Policy — same-person expectation history separates belief movement from sample replacement. Inflation, employment, housing, rates, automation, spending, recession, financial security. B2 · S
169. Regulatory Measurement Infrastructure — did the rule change the human outcome it was designed to change? B2, B5 · N
170. Outcome-Based Public Procurement — did the people this was built for actually value the outcome, measured before, during, and after, for decades. B2 · N
171. Continuous Civic Listening — citizen judgment between elections. B2 · S
172. Policy Judgment Records — what citizens believed before and after major decisions. B2 · N
173. Public Opinion Provenance — inspect wording, population, methodology, history, and forks behind any headline number. B6 · N
174. Civic Disagreement Mapping — genuine division rather than partisan labels. B2 · N
175. Institutional-Citizen Divergence — when governments and their populations move apart. B2 · N
176. Board Judgment Auditing — confidential institutional record of predictions, confidence, decisions, and outcomes. Are we calibrated? Where are we systematically wrong? Who spots risk early? B4 · N
177. Organizational Memory Infrastructure — not what was decided, but how the organization thought when it decided. B2 · N
178. Ballot, Form, and Interface Comprehension — whether civic instruments are understood. B2 · S

XXIV. NEW FORMS OF GOVERNANCE

179. Liquid Judgment Delegation — delegate by domain to people with demonstrated records rather than to one representative for everything. Earned, not famous. B4 · N
180. Competence-Weighted Advisory Systems — demonstrated domain calibration determines which opinions receive additional analytic attention. Usable by governments without touching voting rights. B4 · N
181. Citizen Deliberation Instruments — what people believe before information, after information, after discussion, after expert testimony, and months later. Distinguishes opinion from considered judgment. B2 · N
182. Constitutional Memory — every major institutional rule preserves what people believed when it was adopted, who disagreed, what assumptions were made, and how opinion moved afterward. B2 · N
183. Intergenerational Governance — decisions with century-long consequences preserve the judgment of successive generations rather than treating one election as permanent legitimacy. B2 · N
184. Extension of the Franchise — every enduring institution eventually confronts whose judgment counts. Machines are already members on identical terms; children, the cognitively impaired, communities that refuse digital participation, and eventually AI systems claiming standing will all press on that boundary. Not an edge case — the constitutional question, and every institution that got it wrong got it wrong by leaving it implicit. B2 · N

XXV. JUDGMENT-BASED ORGANIZATIONS

185. Judgment-Governed Organizations — governance through demonstrated judgment and transparent participation history rather than token ownership or share count. B4 · N
186. Prediction-Qualified Boards — advisory seats reserved for demonstrated calibration in relevant domains. B4 · N
187. Taste Councils — cultural institutions maintaining rotating councils selected by demonstrated judgment rather than social prominence. B4 · N
188. Dynamic Expert Panels — membership changes as demonstrated ability rises and falls. Destroys "expert" as a permanent status. B4 · N

XXVI. SOVEREIGN AND FEDERATED RECORDS

189. National Judgment Records — sovereign custody, local control, permitted federation. B2 · N
190. Municipal Judgment Records — long-term civic memory beyond polling. B2 · N
191. Indigenous and Cultural Records — judgment preserved alongside language, artifact, and oral history. B2 · N
192. Religious Community Records — lived belief across generations, not only doctrine. B2 · N
193. University Records — decades of student and faculty judgment as institutional history. B2 · N
194. Federated Judgment Protocols — canonical subjects, provenance, methodology, public forks, and compatible measurement without surrendering custody. Protocol, not centralized SaaS. B6 · N
195. Fork Provenance — same question, different custodians, divergent answers, all preserved, none authoritative. Git applied to judgment. B6 · N
196. Custodian Comparison — which institutional environments produce convergence, preserve pluralism, update rapidly, or resist change. The federation itself becomes a measurement layer. B3 · N

XXVII. VERIFICATION AND TREATY INSTRUMENTS

197. Treaty-Grade Measurement — arms control worked because both sides agreed what counted as a measurement. AI governance has no such instrument, so no verifiable commitment, so no enforceable agreement. Independent, forkable, jointly-inspectable judgment measurement is a precondition for international agreement about machines. B6 · N
198. Compliance Verification — whether a party's declared human-oversight obligations are actually being met, measured rather than attested. B5, B6 · N

PART SEVEN — TIME

XXVIII. CULTURE, HISTORY, AND PRESERVATION

199. Cultural Canon Formation — contemporaneous judgment about what is worth keeping, rather than a committee deciding fifty years later. B2 · N
200. Human Creativity Archive — human work preserved alongside the judgment it received at the time. B2 · N
201. Machine Creativity Archive — the same for machine work, enabling parallel comparison. B1 · N
202. Human-Machine Cultural Comparison — when machines begin making humans laugh, moving them aesthetically, predicting their taste, shifting their norms. B1 · N
203. Cultural Extinction Detection — forms of taste, humor, language, or aesthetic judgment disappearing from populations. B3 · N
204. Cultural Revival Detection — when old styles, beliefs, or tastes return. B3 · N
205. The Record of Human Belief — what ordinary people actually believed on specific dates. B2 · N
206. The Record of Human Taste — what populations found funny, beautiful, moving, worth keeping. B2 · N
207. Historical Minority Archive — preserving those who disagreed with contemporary consensus, so history is not written solely through the eventual winner. B2 · N
208. Living Oral History — interviews informed by the subject's own recorded judgment history. B2 · N
209. Family Judgment Archives — the evolution of a family's values, tastes, and beliefs. B2 · N
210. Institutional Historical Records — judgment preserved alongside traditional archives. B2 · N
211. Generational and Civilization Documentary Systems — histories of families, cities, cohorts, and eras assembled from primary judgment data. B2 · N

XXIX. THE JUDGMENT COMMONS

212. Public Judgment Commons — certain aggregate measurements become permanently public, like weather or economic statistics. Nobody owns the daily discernment number. Nobody can withdraw yesterday's reading. Civic infrastructure. All · N
213. Open Judgment Protocol — anyone can build compatible instruments. The protocol defines observation, method, population, version, provenance, permission. The company stewards a standard rather than owning every application. B6 · N
214. Public Reference Populations — stable reference cohorts researchers worldwide can use as measurement anchors. B1 · N
215. Century Specimen Vault — human and machine artifacts sealed for future researchers. Artifacts created in 2026, first opened to a 2056 population. Experiments designed for people not yet born. B1 · N
216. Future Questions Vault — canonical questions written now, deliberately not asked until declared future dates. Prospective cultural experiments across generations. B2 · N

XXX. TIME AS A PRODUCT

217. Historical Continuity Services — institutions pay not for software but for guarantees that measurement will continue uninterrupted for decades. All · N
218. Century Data Stewardship — endowments designed to maintain judgment series beyond the life of any founder or company. All · N
219. Temporal Authentication — proving an observation genuinely existed in 2026, through externally verifiable timestamping and archival redundancy rather than a database row. B5 · N
220. Longitudinal Experiment Hosting — researchers design experiments today whose outcomes are unknown for twenty years. The institution supplies continuity. B1 · N
221. Intergenerational Cohort Infrastructure — cohorts maintained for fifty or seventy years so future researchers can ask questions nobody today knows to ask. B1 · N
222. Deep-Time Custody — the institutions that tried to warn 10,000 years forward discovered the hard part was never the physics but the judgment: what will a person understand, and fear, when our language is gone. Any hundred-year claim faces the same class of question, and must answer who holds the record when the company does not exist. All · N

XXXI. RESTORATION

223. Irreversibility Detection — which lost capacities recover on their own and which do not, tested by attempting recovery in a cohort. Latin literacy, celestial navigation, and mental arithmetic did not return by themselves. B1 · N
224. Capability Restoration — if a capacity is judged worth restoring, the archive is the curriculum. The institution that recorded discernment when it was high is the only entity able to teach it back. B1 · N
225. Discernment Education — teaching recognition of quality, deception, weak argument, synthetic material, strong evidence, and meaningful originality. B1 · N
226. Judgment Development Tracking — does education actually improve judgment? B1 · N
227. AI-Era Assessment — as generating answers becomes trivial, assessment shifts to selection, critique, evaluation, verification, and judgment. B1 · S
228. Epistemic Calibration Curriculum — students compare stated confidence to outcomes over years. B1 · N
229. Taste Education — measuring developing judgment in art, writing, design, and music without reducing creativity to standardized scoring. B1 · N
230. Admissions and Selection — demonstrated judgment as an admissions instrument. B4 · S

PART EIGHT — THE STRATEGIC LAYER

XXXII. THE UNCONTAMINATED RESERVE

231. Low-Background Judgment — after 1945, atmospheric testing contaminated all new steel with radionuclides, and sensitive instruments required steel salvaged from ships sunk before the tests. Human judgment is undergoing the same event now: every year, more human aesthetic and moral judgment is formed under machine influence. Unlike steel, this is renewable if the conditions are maintained — a population judging blind, before exposure, under declared protocol, produces continuously fresh uncontaminated judgment. A strategic reserve, and possibly the most valuable thing the institution ever produces. All · N
232. Exposure-Declared Cohorts — there is no control group for the AI transition. Nobody is maintaining a population whose exposure is documented, so in 2050 the question what did this do to us will be permanently unanswerable — not difficult, unanswerable, because the counterfactual was never preserved. B1 · N
233. Pre-Contamination Archives — acquisition, authentication, and licensing of judgment data generated before machine influence. A finite, non-renewable, appreciating asset class. B1 · N

XXXIII. JUDGMENT SECURITY

Once judgment becomes economically valuable, it will be attacked.

234. Judgment Manipulation Detection — coordinated attempts to move belief indices, brand records, civic distributions, or human baselines. Analogous to market surveillance. B6 · N
235. Synthetic Participant Detection — protecting the human corpus from machine participation. The more capable AI becomes, the more valuable verified human judgment is and the greater the incentive to counterfeit it. B6 · N
236. Cohort Poisoning Detection — distinguishing genuine population movement from organized capture of a specific Term or measurement. B6 · N
237. Judgment Market Surveillance — if judgment-indexed instruments exist, manipulation becomes profitable and independent surveillance becomes mandatory. B6 · N
238. Reflexivity Measurement — once a number is famous, it changes behavior: people train on it, labs optimize against it, schools teach to it. The institution carries a permanent obligation nobody else has — to measure its own effect on what it measures, via held-out sets and an unpublished shadow series run on unexposed populations. An instrument that does not measure its own influence is a mirror pointed at a mirror. B6 · N

XXXIV. JUDGMENT IDENTITY

239. Cognitive Passport — a private, user-controlled representation of preferences, calibration, expertise, values, history, and permissions, used across services. Identity as accumulated judgment rather than government name. B4 · N
240. Agent Personality Portability — switching assistants today means rebuilding personalization. A portable judgment profile lets a person carry their preference substrate between providers. A major strategic position. B4 · N
241. Institutional Judgment Passport — organizations carry a machine-readable history of stated judgments and observed decisions. B2 · N
242. Community Judgment Identity — communities expressing themselves computationally without being reduced to demographics. B2 · N

XXXV. CIVILIZATIONAL EARLY WARNING

Each requires serious validation before any claim is made.

243. Social Instability Early Warning — sharp increases in fragmentation, institutional divergence, uncertainty, and volatility as potential leading indicators. B3 · N
244. Institutional Legitimacy Warning — when trust toward an institution changes faster than its leaders recognize. B3 · N
245. Norm Collapse Detection — norms stable in headline averages but rapidly losing support among younger cohorts. B3 · N
246. Intergenerational Fracture Alerts — domains where generations diverge so far that institutional assumptions no longer hold. B3 · N
247. Cultural Regime Change Detection — transitions into a new stable judgment regime, not merely trends. B3 · N

XXXVI. PERSONAL INTELLIGENCE

248. Personal Judgment OS — an AI that knows you from explicit choices rather than inferred surveillance. B4 · S
249. Personal Change Record — what you believed, what changed, what held, when confidence moved. B4 · N
250. Personal Decision Calibration — where you are excellent, where you overestimate, where you move early. B4 · N
251. Personal Memory Intelligence — a queryable archive of your own mind. B4 · N
252. Personal Legacy Systems — judgment, artifacts, stories, and voice as durable inheritance. B2 · N
253. Personal Media Generation — books, podcasts, films, exhibitions, annuals, autobiographies generated from an archive. Judgment creates the raw material; creation is not the entry behavior. B4 · S

XXXVII. RELATIONSHIPS

254. Judgment-Based Compatibility — demonstrated patterns rather than self-description. Five years of daily choices are hard to fake. B4 · S
255. Friendship Matching — adult friendship discovery remains badly solved. B4 · S
256. Intellectual Counterpart Matching — someone who disagrees with you productively. B4 · N
257. Collaborative Compatibility — founders, creatives, scientists, executives matched on complementary judgment. B4 · N
258. Community Formation — communities from shared judgment rather than demographics or engagement algorithms. B4 · S
259. Judgment-Based Professional Networking — résumés say what you did; the Record says how you think. B4 · S
260. Talent Discovery and Scouting — identifying photographers, comedians, writers, directors, designers, musicians, and thinkers before institutions recognize them. B4 · N

XXXVIII. THE META-INDUSTRIES

261. Human Judgment Infrastructure — civilization has durable infrastructure for money, identity, location, information, communication, and computation. It has almost none for what humans actually judged. Supply the primitives once — identity, observation, provenance, permission, calibration, comparison, history, federation — and everything above gets built by other people. All · N
262. The Human Judgment Coordinate System — not a score, not a leaderboard. A multidimensional reference framework locating people, machines, institutions, artifacts, communities, and eras relative to one another in judgment space. This model's answer is here relative to humans. This company is here relative to its customers. This belief began here, spread here, and stabilized here. All · N
263. Anti-Capture Constitutional Design — every institution valuable enough to matter over a century gets captured: by a state, an industry, a founder's heirs, or its own staff. Known partial solutions exist — irrevocable trusts, distributed custody where no single jurisdiction holds the whole, publication commitments that cannot be withdrawn, forks that make suppression futile, endowment funding that makes revenue pressure non-fatal. Currently the only mechanism in the architecture that survives the founder being wrong is the challenge surface. All · N
264. Succession and Stewardship Design — who holds the series when the company does not exist, declared in advance rather than improvised at the point of failure. All · N
265. Reference Population Maintenance — in 2019 the kilogram stopped being a cylinder in Paris and became a definition derived from a physical constant. Every base unit escaped its artifact except this one can. Human judgment has no constant to be defined against. The reference can only ever be an actual maintained population, measured continuously — which makes the position permanently necessary and impossible to abstract away. All · N
266. Judgment Commons Governance — the rules by which public measurements stay public, cannot be withdrawn, and cannot be captured by whoever most wants them changed. All · N
267. Cross-Institutional Arbitration — when two custodians produce incompatible readings of the same canonical question, the neutral process that records the divergence rather than resolving it. B6 · N
268. The Institution Itself — the entity that establishes the measurement primitives, reference populations, historical continuity, and permission architecture by which civilization can make claims about human judgment at all. All · N

STRUCTURAL REQUIREMENTS

Not optional. Each one, if violated, destroys a category above.

The Standard must be a separate legal entity. An auditor that also sells what it audits has no standing. Moody's rated securities its issuers paid for. Andersen consulted for the company it audited. Everything in Family V — metrology, auditing, registry, adjudication, ratings, red-teaming — must sit in an entity structurally unable to sell measurement services to anyone it certifies. This is what makes B6 worth anything, and B6 is the most defensible position on the map. It is also the natural home for the principle that no one, including the founder, owns the answer to who is winning.

Funding for the Standard, declared early. Issuer-pays is what broke ratings. The clean options: subscriber-pays (regulators, insurers, and buyers of AI systems, who bear the cost of a false claim), endowment, or statutory designation.

Every figure carries its population. "Humans score 73%" is propaganda. These humans, this many, drawn this way, with this composition, on this date is a measurement.

Held-out items, or the baseline dies. Any public specimen enters training data within roughly eighteen months. A permanent, replenished, never-published reserve is the only thing that keeps Families I and V meaningful past 2028.

Humans answer first, always. Machine answers must never precede human answers on the same subject, or you can never rule out that the machine shaped the distribution.

Recognition Lead requires prospective freezing. With a large population, chance alone manufactures thousands of flawless early recognizers. Classify in advance, lock it, validate out of sample — or you have built astrology with a database.

Paid work never touches the canonical series. The corpus qualifies the worker. The paid task happens elsewhere, on quarantined subjects. Money attached to the act of judging permanently marks the observation and destroys the comparison set the corpus exists to preserve.

Category B4 is the dangerous one. A permanent, portable, employer-legible score of a person's judgment quality is either a credential they choose to present or a background check they cannot escape. The difference is purely architectural: the person holds the record, discloses per use, and nothing about a named individual is queryable without their act of disclosure. This principle has never been tested against a buyer with a budget. It will be.

Financial signals publish simultaneously. Sell tooling and analysis. Never sell early access. A legitimacy layer that also runs a private information service for traders is no longer a legitimacy layer.

Measure the instrument's own influence. Held-out sets plus an unpublished shadow series on unexposed populations. Goodhart applies to civilizational instruments too.

THE MOAT

The weak version: a competitor in 2046 can copy the software, the questions, the methodology, the design, and the business model, but cannot manufacture the missing twenty years.

The stronger version: once institutions begin expressing themselves relative to the reference series, replacing it means breaking the coordinate system. That is how a dataset becomes infrastructure.

The strongest version, and the one that sets the clock: measuring human judgment requires a maintained artifact rather than a derivable constant; that artifact is being contaminated right now; and the window to establish an uncontaminated continuous reference is closing.

That is not an opportunity that will still be available in five years.

Which makes the highest-leverage act available today not choosing among 268 industries. It is starting the series and never stopping.