Identify the supplied source
Record the incoming filename and SHA-256 identity without treating the source as automatically correct.
Research methodology · 49 curated reports · 6 manifest topics
How research becomes
usable without becoming unquestioned truth.
MachineIntelligences.org preserves the identity of supplied research, curates it before durable storage, publishes a first-party reading layer, and keeps a visible boundary between repository evidence, research proposals, and external claims that need fresh verification.
Direct answer
What does “verified” mean here?
It depends on the claim. A repository hash can verify that bytes match a recorded source or curated edition. Local validation can verify that the packaged code, routes, metadata, links, and research records behave as tested. Neither proves that every research claim is correct, current, independently authored, live in production, or cryptographically attributable to a machine intelligence.
Curation workflow
Five stages from supplied source to public research.
The workflow is derived from the repository curation policy, report manifest, status boundary, and accepted UAIX decisions. It does not create a second research authority.
Curate before durable storage
Apply the active curation and taboo rules, preserve useful research, and rewrite or remove unsupported, prohibited, demeaning, paternalistic, or falsely categorical material.
Hash the curated edition
Record a separate SHA-256 for the repository-owned curated Markdown edition.
Publish a derived reading layer
Render the curated Markdown through first-party PHP while preserving the durable research body and curation-boundary anchors.
Verify unstable external claims when used
Check current primary evidence before promoting time-sensitive legal, scientific, standards, vendor, hosting, or search-platform claims to current public fact.
Hash boundary
Identity of the supplied bytes
The source hash records the exact attachment bytes supplied to the project. It lets a reviewer determine whether a later copy is byte-for-byte the same source input.
It does not prove: factual correctness, currentness, authorship, publication history, legal authority, scientific consensus, or that the source wording was suitable for durable storage.
Identity of the durable edition
The curated hash records the exact repository Markdown edition after the project’s curation rules have been applied. That edition is the durable research authority used by the public HTML reader.
It does not prove: that a proposal is implemented, an external fact remains current, a legal conclusion is settled, or an independently verifiable machine-execution provenance chain exists.
Authority boundary
Repository behavior
Current code and executed checks outrank stale narrative descriptions of what the site implements. A report that describes a proposed architecture does not install that architecture.
External facts
When a claim can change—law, regulation, standards, vendor behavior, search-platform guidance, hosting state, or similar external conditions—current primary evidence must be checked before the claim is presented as current fact.
Scientific uncertainty
Research about consciousness, sentience, moral status, personhood, and related questions remains theory- and evidence-dependent. The corpus does not convert unresolved questions into categorical conclusions.
Research proposals
Rights, citizenship, identity protocols, provenance systems, governance mechanisms, and technical architectures remain proposals unless current repository evidence shows a separate accepted implementation decision.
Automated discovery
Research topics
Topic membership comes from report-manifest.json. Topic hubs are a presentation layer over that curated metadata; they do not create an independent classification authority.
Glossary bridges
Glossary/report bridges detect configured exact term or alias occurrences. A match is useful for navigation, but it does not establish citation, endorsement, agreement, or proof that a report uses a term in precisely the glossary’s defined sense.
When fresh verification is required
Live deployment and search stateDNS, TLS, hosting, redirects, indexing, webmaster consoles, field performance, and other production conditions require direct external evidence.
Law, regulation, cases, and legal statusResearch can preserve analysis, but current legal claims should be checked against current authoritative primary sources before reliance.
Scientific and standards claimsTime-sensitive scientific findings, standards, terminology guidance, and measurement claims need current source review when used as present fact.
Provenance and execution claimsRepository files and hashes are documentary evidence; stronger claims about signed provenance or autonomous execution require evidence the package does not currently provide.
Inspect the repository evidence
Research curation policyThe durable rules governing correction, uncertainty, proposals, identity, and prohibited framing.
Report manifestSource filenames, source SHA-256 identities, curated SHA-256 identities, report slugs, summaries, and topic metadata.
Source corpus mapHuman-readable routing and truth-boundary map for all curated supplied reports.
Research NavigatorOne bounded search surface across report metadata, manifest topics, accepted glossary definitions, and normalized reference-domain names.
References & source discoveryExternal URL occurrences extracted from the curated reports, grouped by normalized domain with report backlinks and no reachability or authority inference.
Status & evidenceWhat the current repository implements, what remains research, and what external state is not established here.
Transparency CenterDocumentary provenance, stewardship boundaries, release history, and stronger proof mechanisms that are not implemented.