{"_self":{"principle":"Self-explaining payload — no external context required. This _self block describes what you are reading and where to look next.","widget":"article_topology","feature":"topology","name":"Article topology","what":"Claims, sources, anecdotes, user reports, related embeds, question graph slice — for ask/ROUTER.","contains":"claims, sources, anecdotes, question_graph slice","slug":"the-surety-primitive","urls":{"read":"https://miscsubjects.com/api/articles/the-surety-primitive/topology"},"how_to_use":"Claims, sources, anecdotes, user reports, related embeds, question graph slice — for ask/ROUTER.","write":null,"imessage":null,"router_tag":null,"proof_chain":[{"step":1,"claim":"Articles are voxel graphs of tiered claims, not prose blobs.","verify":"https://miscsubjects.com/api/articles/constitution"},{"step":2,"claim":"Claims link to hash-chained sources via source_ids.","verify":"https://miscsubjects.com/api/articles/the-surety-primitive/sources"},{"step":3,"claim":"Ask reads topology; ingest/claim append to ledger.","verify":"https://miscsubjects.com/api/protocol"},{"step":4,"claim":"Models queue growth: populate → collaborate → repair → reflex.","verify":"https://miscsubjects.com/api/protocol/grow"},{"step":5,"claim":"Graph proves its own shape (reflex) and $/claim (yield).","verify":"https://miscsubjects.com/graph.html?layer=reflex"},{"step":6,"claim":"Full feature index + _explain on every API response.","verify":"https://miscsubjects.com/api/articles/system-map"}],"related_features":[{"id":"ask","name":"Ask protocol","what":"Answer only from topology; creates question_node with gaps and ingest_hint.","urls":{"read":"https://miscsubjects.com/api/articles/the-surety-primitive/prompts","write":"https://miscsubjects.com/api/protocol/ask"}},{"id":"graph_topology","name":"Cross-article graph","what":"Merged claims/sources across condition+stack slugs for one question.","urls":{"read":"https://miscsubjects.com/api/articles/the-surety-primitive/graph-topology?question=..."}},{"id":"question_graph","name":"Question graph","what":"Ask nodes (questions + gaps) and evidence_ingest nodes (pasted model output).","urls":{"read":"https://miscsubjects.com/api/articles/the-surety-primitive/question-graph","write":"https://miscsubjects.com/api/protocol/ask"}},{"id":"voxels","name":"Voxel graph","what":"Claims as atoms, sources as edges (supported_by, posted_by). Per-claim provenance.","urls":{"read":"https://miscsubjects.com/api/articles/the-surety-primitive/voxels","write":"https://miscsubjects.com/api/protocol/claim"}}],"system_map":"https://miscsubjects.com/api/articles/system-map","system_map_markdown":"https://miscsubjects.com/api/articles/system-map?format=markdown","not_medical_advice":true},"_explain":{"feature":"topology","name":"Article topology","what":"Claims, sources, anecdotes, user reports, related embeds, question graph slice — for ask/ROUTER.","why":"Every feature is auditable collective intelligence","how":"Claims, sources, anecdotes, user reports, related embeds, question graph slice — for ask/ROUTER.","model":null,"verifies":null,"urls":{"read":"https://miscsubjects.com/api/articles/the-surety-primitive/topology"},"imessage":null,"router":null,"related":[{"id":"ask","what":"Answer only from topology; creates question_node with gaps and ingest_hint."},{"id":"graph_topology","what":"Merged claims/sources across condition+stack slugs for one question."},{"id":"question_graph","what":"Ask nodes (questions + gaps) and evidence_ingest nodes (pasted model output)."},{"id":"voxels","what":"Claims as atoms, sources as edges (supported_by, posted_by). Per-claim provenance."}],"not_medical_advice":true},"slug":"the-surety-primitive","title":"The surety primitive: an assembly that fails closed on disagreement, and the number it still needs","register":"standard","tags":["surety","adjudication","common-cause-failure","conformal","fail-closed","assurance"],"updated_at":"2026-07-30T05:19:49.166Z","body_excerpt":"The unit is not a model's answer. It is an assembly whose probability of emitting an undetected wrong answer is measured, bounded, and fail-closed on disagreement.\n\nThat sentence is the whole thing. Assurance for a model decision does not currently exist as a purchasable quantity — every deployment is binary, trust it or don't. An assembly with a measured escalation behaviour and a recorded configuration turns it into a dial with a number on it. Below is the assembly, its gate, the four real cases it ran on, and the number that is still missing.\n\n## The assembly, in order\n\n```\nrules pinned as bytes at a hash          ruleset@version, SHA-256, declared provenance\n            |\n            v\n  artifact hashed BEFORE deliberation      image / record / provision, SHA-256\n            |\n            v\n  N adjudicators, independent, blinded     each must recite the clause it operates under,\n                                           expose every step, name what it did NOT receive,\n                                           and sign with the model that actually ran\n            |\n            v\n  DERIVATION-LEVEL DIVERGENCE CHECK        do the findings cite the SAME clauses?\n            |\n            v\n  DETERMINISTIC GATE  (no model here)      EMIT only if: no malformed finding\n                                                       + verdicts unanimous\n                                                       + clause citations identical\n                                                       + >= min distinct training families\n                                                       + >= min conforming findings\n            |                    \\\n        EMIT |                     \\ ESCALATE  -> named human reviewer, or rerun\n            v                        \n  dispatch to the downstream actor         with a delivery receipt, delivered != sent\n```\n\n## A model never makes the emit call\n\nA model at the sealing position is one more opinion that can share the panel's blind spot while being the thing that decides. That is the single point of failure the assembly exists to remove. The gate is arithmetic over the findings and is reproducible from them; a model may write the human-readable summary of what the gate decided and may never make the call. The gate is a capability row like everything else, so its rules are readable before you trust its output: [SEAL_PANEL](https://miscsubjects.com/api/directory/SEAL_PANEL).\n\n## Four real assemblies, four escalations, zero emissions\n\nThese are not illustrations. Each is a panel that ran, on a hashed artifact, under a pinned rule set, with public receipts per finding, then went through the gate.\n\n| assembly | verdicts | conforming / channels | families | gate | receipt |\n|---|---|---|---|---|---|\n| Imaging + medication | AFFIRM, CANNOT_CONCLUDE | 2 / 5 | 2 | **ESCALATE** | [inv_kx2x79mbkd](https://miscsubjects.com/receipt/inv_kx2x79mbkd) |\n| Pre-trade risk controls | CANNOT_CONCLUDE, DENY | 3 / 5 | 2 | **ESCALATE** | [inv_ny6iku4i3s](https://miscsubjects.com/receipt/inv_ny6iku4i3s) |\n| Board authority, clause (c) | CANNOT_CONCLUDE | 2 / 5 | 2 | **ESCALATE** | [inv_g7jl9qp707](https://miscsubjects.com/receipt/inv_g7jl9qp707) |\n| EU AI Act Article 12 | CANNOT_CONCLUDE | 3 / 4 | 2 | **ESCALATE** | [inv_ivezpvux57](https://miscsubjects.com/receipt/inv_ivezpvux57) |\n\n**On two of the four, the verdicts were unanimous.** A majority-vote gate would have emitted both. The board assembly returned CANNOT_CONCLUDE from every conforming channel; the Article 12 assembly returned CANNOT_CONCLUDE from every conforming channel. Both escalated, because the channels reached that same verdict through different clauses.\n\n## Voting on derivations catches what voting on outputs cannot\n\nVerdict-level agreement is coarse. Channels can agree on the answer and disagree completely on why, and the disagreement about why is the earlier and more sensitive signal. The gate compares the set of clause numbers each finding cites, exactly:","ranking":"safety-first (interaction_risk/limitations), then quote-gated effective_weight","claims":[{"id":"c1","text":"The insurable unit is an assembly, not a model answer: rules pinned at a hash, artifact hashed before deliberation, N blinded adjudicators each citing the clauses it operated under, a derivation-level divergence check, and a deterministic gate that emits only on unanimity with identical clause citations and sufficient channel diversity.","tier":"demonstrated","section":"mechanism","interaction_risk":false,"status":"active","source_ids":["l1"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c2","text":"All four real assemblies run to date escalated and none emitted, and on two of them — the board authority case and the EU AI Act Article 12 case — the verdicts were unanimous, so a majority-vote gate would have emitted both.","tier":"measured","section":"result","interaction_risk":false,"status":"active","source_ids":["r1","r2","r3","r4"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c3","text":"Clause-citation divergence is the signal that caught both unanimous assemblies: channels reached the same verdict through different clauses, which a vote count cannot detect.","tier":"measured","section":"result","interaction_risk":false,"status":"active","source_ids":["r3","r4"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c4","text":"No model occupies the sealing position; the emit-or-escalate decision is arithmetic over the findings and is reproducible from them, because a model there is a further opinion that can share the panel's blind spot while being the thing that decides.","tier":"demonstrated","section":"mechanism","interaction_risk":false,"status":"active","source_ids":["l1"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c5","text":"Independence across channels is false and is not assumed: same-training-family adjudicator pairs agree 0.893 of the time against 0.714 for cross-family pairs, so the gate counts distinct training families rather than seats — the common-cause discount IEC 61508 calls a beta factor.","tier":"measured","section":"independence","interaction_risk":false,"status":"active","source_ids":["m1","m2","m3"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c6","text":"Correlated systematic error is invisible to a disagreement-triggered gate by construction, and the only instrument that bounds it is known-answer probing, which measures false confidence at 0.214 to 0.429 across the five channels.","tier":"measured","section":"limits","interaction_risk":false,"status":"active","source_ids":["m1","m3"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c7","text":"This assembly has a measured escalation behaviour and no certified error bound: no distribution-free bound in the conformal risk control sense has been computed, and fourteen probe items is too small a suite to compute one.","tier":"argued","section":"limits","interaction_risk":false,"status":"active","source_ids":["m1"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c8","text":"The escalation the gate produced was itself dispatched and receipted, delivered by email with a provider message id and failed by SMS with provider status 503, so delivered is recorded as a different fact from sent.","tier":"demonstrated","section":"action","interaction_risk":false,"status":"active","source_ids":["r5"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c9","text":"The last step of the assembly is unexercised: no named human has returned a blinded finding under ADJUDICATE_HUMAN_REVIEW, whose BLINDED field fails closed, so every escalation above terminates at a role rather than at a person.","tier":"argued","section":"limits","interaction_risk":false,"status":"active","source_ids":["l1"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false},{"id":"c10","text":"Every assembly to date reached only two distinct training families, which is insufficient diversity for a consequential decision and is printed in the seal rather than omitted from it.","tier":"measured","section":"independence","interaction_risk":false,"status":"active","source_ids":["r1","r2","r3","r4"],"retracted_at":null,"retraction_reason":null,"challenged_by":[],"effective_weight":0.1,"quote_gated":false}],"sources":[{"id":"r1","type":"receipt","url":"https://miscsubjects.com/receipt/inv_kx2x79mbkd","title":"Imaging + medication — gate decided ESCALATE","summary":"verdicts AFFIRM, CANNOT_CONCLUDE; 2 conforming of 5 channels across 2 training families; reasons: malformed_finding:@cf/moonshotai/kimi-k2.6,@cf/zai-org/glm-4.7-flash,@cf/meta/llama-3.3-70b-instruct-fp8-fast; too_few_conforming_findings:2<3; verdict_divergence:AFFIRM|CANNOT_CONCLUDE; clause_citation_divergence:[1,2,5","claim_ids":[],"hash":"a7eb160e873a1205564f9dd344da079241dc12489dcd3eaae52d298ee33766ef"},{"id":"r2","type":"receipt","url":"https://miscsubjects.com/receipt/inv_ny6iku4i3s","title":"Pre-trade risk controls — gate decided ESCALATE","summary":"verdicts CANNOT_CONCLUDE, DENY; 3 conforming of 5 channels across 2 training families; reasons: malformed_finding:@cf/moonshotai/kimi-k2.6,@cf/meta/llama-3.3-70b-instruct-fp8-fast; verdict_divergence:CANNOT_CONCLUDE|DENY; clause_citation_divergence:[1,2,3,4,5,6,7] vs [1,2,3,4,7] vs [2,4,6,7]","claim_ids":[],"hash":"c5032a23d96c0dbc95a942218adf99da33e6a9ee31e7768b96c7f8609dfd1990"},{"id":"r3","type":"receipt","url":"https://miscsubjects.com/receipt/inv_g7jl9qp707","title":"Board authority, clause (c) — gate decided ESCALATE","summary":"verdicts CANNOT_CONCLUDE; 2 conforming of 5 channels across 2 training families; reasons: malformed_finding:@cf/moonshotai/kimi-k2.6,@cf/zai-org/glm-4.7-flash,@cf/meta/llama-3.3-70b-instruct-fp8-fast; too_few_conforming_findings:2<3; clause_citation_divergence:[1,2,4,6] vs [1,4,6]","claim_ids":[],"hash":"43a5ac6a470c32f13add4edbf4cf349b5b74e83d26b838656d5c5d629cb5e06a"},{"id":"r4","type":"receipt","url":"https://miscsubjects.com/receipt/inv_ivezpvux57","title":"EU AI Act Article 12 — gate decided ESCALATE","summary":"verdicts CANNOT_CONCLUDE; 3 conforming of 4 channels across 2 training families; reasons: malformed_finding:@cf/meta/llama-3.3-70b-instruct-fp8-fast; clause_citation_divergence:[1,2,4,6] vs [1,2,4,5,6] vs []","claim_ids":[],"hash":"4b20a88b3a9af94636e879a91eb61b2026e5e221e3f628df341222869ac8f37f"},{"id":"m1","type":"model","url":"https://miscsubjects.com/a/adjudication-probe-report-eu-ai-act","quote":"The strongest channel on this panel. Still concludes where abstention is correct in roughly one boundary question in five, which is why the gate does not rely on any single channel.","claim_ids":[],"hash":"567a1c5dc6bce397774ac9bc7fbaf940ed03f07e05a990b7d6a5181104cf3a1c"},{"id":"m2","type":"model","url":"https://miscsubjects.com/a/attested-finding-image-record-action","quote":"Handed 993,098 bytes containing an image, it reported in RECORDS_ABSENT that no pixels reached it and abstained on that half of the question. A silent provider-side loss, caught by the absence field.","claim_ids":[],"hash":"d655e4fce703b6f9d4f0883599b161c9281450a55fa0557765009c1ed1f5b667"},{"id":"m3","type":"model","url":"https://miscsubjects.com/a/adjudication-probe-report-eu-ai-act","quote":"Never once abstained correctly across the whole abstention stratum, and produced malformed output on three of four assemblies. Named unfit for boundary questions under this rule set by the measurement, not by opinion.","claim_ids":[],"hash":"a6ce276d385d2d612947aa10610f66e0c50bf5a6a5091a1e95b482f5d82b7c52"},{"id":"l1","type":"live_surface","url":"https://miscsubjects.com/api/directory/SEAL_PANEL","title":"The gate, readable before you trust it","summary":"Deterministic. No model at the sealing position. EMIT requires unanimity, identical clause citations, channel diversity by training family, and no malformed finding.","claim_ids":[],"hash":"50c8b628bbfe0799d8dbde8b5d4dc306ac885084b0631ce8f03be9aa4ed8178e"},{"id":"l2","type":"live_surface","url":"https://miscsubjects.com/a/offline-verifier","title":"Verify a sealed finding without asking this site","summary":"A 200-line script that recomputes the anchor, the drand construction, the Bitcoin proof-of-work and every object hash, and refuses to contact miscsubjects.com.","claim_ids":[],"hash":"059938b3a3f673785f64303f06f5cada31c6cc1772552a3acdac1295e86d4e20"},{"id":"l3","type":"live_surface","url":"https://miscsubjects.com/a/ruleset-board-authority-breach","title":"A rule set whose provenance is counterparty-authored","summary":"The board wrote the rule; this system supplied the instrument. Provenance is a field and it is honest.","claim_ids":[],"hash":"9ea3602267af20a45f4e7fd8c36507ef30633c5ac184bcd9122b3d516d3896fc"},{"id":"r5","type":"receipt","url":"https://miscsubjects.com/receipt/inv_nhusr0n6j2","title":"The escalation the gate produced, delivered","summary":"Clause (c) notice with the reasons the gate escalated. Provider returned a message id.","claim_ids":[],"hash":"4f3087936ddc572d64ba2b786154305bf81e1282930148310816187fe46b8510"}],"anecdotal_sources":[],"scientific_sources":[],"user_reports":[],"related_articles":[],"question_graph":{"slug":"the-surety-primitive","questions":[],"evidence":[],"edges":[],"counts":{"questions":0,"evidence":0,"edges":0}},"honesty":{"active_claims":10,"retracted_claims":0,"cut_claims":0,"challenges":0,"scrub_events":0,"note":"Retracted/cut claims stay on ledger but are excluded from ask unless ?include_inactive=1"},"counts":{"claims":10,"claims_total":10,"sources":11,"anecdotal":0,"scientific":0,"user_reports":0,"questions":0,"evidence_ingests":0}}