For Developers

Seven hypotheses about AI's economic impact. A consensus-or-parsimony discriminator. One composite model survives. And it accidentally makes the strongest case for a 1966 UN treaty most engineers have never heard of.

Technical Context

This page provides direct access to the analysis framework, technical architecture, and open-source tools. The differential diagnosis methodology, observatory API, and structured data layers offer extensible building blocks for your own work.

Technical Stack

The site runs on a static-first architecture optimized for performance, accessibility, and zero-JS-by-default delivery. Every interactive component uses the islands architecture pattern — hydrating only when visible, keeping the critical path lightweight.

Framework

Astro 5 + MDX + Svelte 5 Islands. Static build produces 53 pages in ~3 seconds. Zero client JS by default; islands hydrate on visibility via client:visible. Content collections provide type-safe frontmatter validation.

Visualization

D3-geo powers the RatificationMap with Natural Earth TopoJSON data. Custom Svelte components handle ScoreChart, RadarChart, CausalChain, and Timeline rendering. All charts carry ARIA labels and respect the active color theme.

Structured Data

JSON-LD (Schema.org) on every page. SKOS taxonomy for the glossary. 49 terms with sameAs/isBasedOn mappings to authoritative external sources. Machine-readable agent-inbox.json for AI agent discovery.

Hosting

Cloudflare Pages (main + blog) with global CDN. CSP + HSTS headers. Cloudflare Web Analytics — no tracking scripts, no cookies, no fingerprinting.

Vendor Triad

Three vendors power this project. None endorses or has reviewed its conclusions. All contributions carry full attribution.

Anthropic

Generative engine. Claude built this entire site — analysis, copy, code, and components — under human direction. Every editorial decision visible in the commit log.

Wolfram|Alpha

Computational knowledge. Wolfram|Alpha (Wolfram Research, Inc., Champaign, IL) provides compound growth projections, intersection timelines, derived ratios, and reference data. Free API tier. Terms of use.

Cloudflare

Infrastructure. Cloudflare Pages hosts both sites with global CDN, DNS, SSL/TLS, and DDoS protection. Cloudflare Web Analytics — no tracking scripts, no cookies, no personal data collection.

Licensing

Dual-license structure separates code from content. Both licenses permit commercial use. The intent: anyone can fork, extend, or critique this project without friction.

Code — Apache 2.0

All source code, components, build scripts, and configuration. Use, modify, distribute — with attribution. Patent grant included. The full license text lives in the repository root.

Content — CC BY-SA 4.0

All analysis, blog posts, educational materials, and documentation. Adapt and redistribute — with attribution and share-alike. Commercial use permitted. Translations welcomed.

Data

Primary data from DOE/EIA, BLS, OECD, WIPO, EPI, NSCR, CRA. Computed projections via Wolfram|Alpha. All sources cited. All computations link to verification pages.

Data Provenance

Every quantitative claim in the analysis traces to a primary source. No secondary aggregators accepted without verification against original data. The Resources page provides full links, and the Evidence section presents the integrated data narrative.

Economic Data

BLS (labor force statistics, occupational projections), EPI (wage stagnation, productivity-pay gap), FRED (Gini coefficient series), CBO (healthcare cost projections), Yale Budget Lab (tariff impact estimates), Wharton PWBM (productivity modeling).

AI Research

METR (AI developer productivity measurement), Anthropic (estimating AI gains), SF Fed (AI macroeconomic policy), Deloitte (enterprise adoption surveys), Faros AI (productivity paradox analysis).

Energy + Patents

DOE/EIA (data center energy consumption), S&P Global (energy demand projections), WIPO (global patent trends), USPTO (AI patent classification dataset).

Education + Labor

NSCR (postsecondary enrollment trends), CRA (CS enrollment surveys, Taulbee data), OECD (education spending by country), Barrero/Bloom/Davis (WFH Research, remote work adoption rates).

Privacy

This site collects no personal data. No tracking scripts. No cookies. No fingerprinting. Cloudflare Web Analytics provides aggregated traffic metrics (page views, visit counts, referrer domains) without identifying individual visitors. The analytics JavaScript runs client-side but transmits only aggregated, anonymized data to Cloudflare.

CSP headers restrict script execution to same-origin and inline (required for Astro island hydration and theme detection). HSTS enforces HTTPS. No third-party tracking pixels, beacons, or analytics services beyond Cloudflare Web Analytics.

The Analysis Framework

The analytical core uses a reusable methodology documented in the recursive methodology blog post. The /discriminate protocol can apply to any multi-hypothesis decision problem — the ICESCR analysis represents one application of a general tool.

Differential Diagnosis

7 hypotheses × 5 dimensions (empirical support, parsimony, consensus, chain integrity, predictive power — 0–5 each, total /25). Two eliminated (H1: 12, H5: 7), one modulates (H6: 16), four survive → Composite A (20/25).

Higher-Order Effects

Ten orders of recursive knock-on analysis (Orders 0–4 complete, 5–9 Phase 2) with explicit confidence degradation (HIGH→LOW→SPECULATIVE). Convergence at Order 3: Four Scarcities model (judgment, specification, curation, energy).

Counterfactual

Seven ratification scenarios (R1–R7) scored through the same discriminator. Composite R-A survives: State AG enforcement (R3+R5+R6 activated by R7). Enforcement mechanism analysis cross-references the counterarguments page.

Dignity Quotient

PSQ-UDHR evaluation framework (21/25). Maps Platform Safety Quotient dimensions to UDHR provisions, producing a measurable dignity construct for AI platform assessment.

Observatory

The Human Rights Observatory evaluates tech stories against UDHR provisions using an HRCB (Human Rights Context Balance) scoring methodology. Corpus: 759+ stories analyzed for rights-relevant signals. Build-time API integration pipes real-time observatory data into the main site's Evidence section. An ethical AI showcase — full methodology transparency, zero tracking, open source.

The observatory demonstrates that AI-powered content analysis can operate with complete methodological transparency. Every scoring decision, editorial classification, and corpus statistic flows through documented, reproducible pipelines. Read the corpus analysis blog post for findings from the first 800+ stories.

GET https://observatory.unratified.org/api/v1/signals

Returns corpus metadata, signal distributions, and editorial faction analysis. Public endpoint, no auth required. JSON response, CORS enabled.

From the Blog

The blog documents the development process, methodology decisions, and observatory findings. Most posts target a developer audience.

Human Rights; Nothing More, Nothing Less.

Every element of this analysis represents implementation of rights 173 nations already committed to. Nothing here asks for anything beyond what the United States signed in 1977.