See the evidence, watch the verdict move, and argue it in the open. A neutral AI shows where the weight of evidence actually lies — claim by claim, not as a single number.
🎧 This is the MethodologyNarrated by AI · 2 min
Live case — verdict evolves as evidence changesCreated 2026-09-22Last updated 2026-10-1813 evidence cards519 public votes
How this works: Verdicts are decomposed by hypothesis, never a single yes/no. Each evidence card carries a direction and strength (likelihood ratio). Priors and ratios are public and challengeable. Both sides are linked on every card. Motive and prior-conduct are inputs to where to look, not proof of what happened. Standard of proof: 66% — clear and convincing evidence.
What This Is — Read Before You Scroll
What this is
This page is an experiment in governance.
The Social Governance System (SGS) is built on a simple premise: better decisions come from transparent, structured, mass collaboration — many people, reasoning together in the open, with their contributions weighed on evidence rather than authority. The hard question has always been how you actually do that at scale without it collapsing into noise, tribalism, or capture.
Artificial intelligence changes the answer. An AI can hold an enormous, contested body of evidence in view at once, sort what is documented from what is speculated, resist both official spin and popular myth, and integrate hundreds of contributions into a coherent, versioned record — tirelessly, and in public. So we are testing it. Live.
Why 9/11
We chose the most difficult possible subject on purpose.
If collaborative, AI-facilitated reasoning can handle 9/11 — a topic drowning in both official narrative and conspiracy overreach, where emotions run high and bad-faith actors are everywhere — then it can handle almost anything. This is the hardest test case we could pick, and that is exactly why it is the right one to start with.
Important: the conclusions below are a draft, not a verdict. They were assembled by an AI working as a neutral judge would — crediting documented evidence regardless of who it embarrasses, and discounting speculation regardless of how satisfying its conclusion. Every claim is tiered: Documented, Contested, or Rejected. Nothing is asserted beyond what the evidence supports.
How you participate
This is where you come in. You can:
Add evidence we missed (with a source)
Challenge a claim (with your reasoning)
Argue a claim should move tiers — up or down
Flag a claim that should be rejected
Every submission is reviewed. Valid, sourced contributions are integrated by the AI steward, and the document is republished with a public changelog. You propose. The AI verifies and reintegrates. The document improves. We repeat.
That loop — humans and AI refining a shared understanding together, in the open, on the record — is the Social Governance System in miniature.
The rules of the room
Sources required. Assertions without evidence don't get integrated.
Documented vs. speculation is always labeled. We never let the two blur.
This is about actions, institutions, and evidence — never ethnic or religious groups. Criticism of a government, agency, or policy is not criticism of a people.
Rejected claims stay rejected unless new evidence overturns them, and we say why.
A necessary distinction, stated plainly
Some of the evidence here concerns the documented actions of the State of Israel, its intelligence services, and specific political movements. Because that subject is routinely shut down with a single accusation, we state our position openly, up front:
A state is not a people. The government of Israel and its intelligence agencies represent institutional and political interests — not Jewish people, who hold every range of view and are themselves among the critics of these policies.
Criticism of a government is not bigotry. Holding Israel — or the United States, or Saudi Arabia, or any state — accountable for documented actions is legitimate scrutiny. The charge of antisemitism cannot be used to place a state's conduct beyond examination.
Political movements are not ethnicities. Zionism is a political movement, not a people; it includes non-Jewish adherents and is opposed by many Jews.
Contributions that use this subject as cover for genuine antisemitism — shifting from documented actions of named actors to blanket claims about Jewish people — are rejected just as firmly. The rule is the same throughout: name the act, name the actor, cite the source — never the tribe.
Where this goes
9/11 is the pilot. In the coming months we'll run the same open exercise on other subjects — each one an invitation to a different community to reason with us. The goal isn't to win an argument about any single topic. The goal is to build and prove a method — and to grow a community of collaborators who want a better way to decide things together.
How the verdict engine works — see the debate, what the claims mean, and where to cast your verdict.
Current Verdict — Claim by Claim
These are the current evidence-weighted verdicts, scored by hypothesis. Each claim is assessed independently against a 66% clear-and-convincing evidence threshold. The verdict is not binary — precision is the anti-strawman design.
✅
9/11 was a criminal act
Established
Al-Qaeda authorship of the attacks is supported by forensic record, intelligence record, and engineering consensus. Posterior ~91–93%.
✅
US government had prior knowledge
Data supports this conclusion
Systemic failures, the Aug 6 PDB, Phoenix memo, and 28-pages Saudi nexus all support foreknowledge/cover-up. Posterior ~70%.
✅
US government committed a war crime
Established
Iraq War authorization on false WMD pretenses, use of torture (Abu Ghraib, black sites), and drone strike civilian casualties meet international law thresholds.
❌
Inside job beyond identified parties
Insufficient evidence — 66% threshold not met
Government authorship (controlled demolition, orchestration) posterior ~5–10%. WTC 7 is the sole technical lever; conspiracy-size constraints are decisive. C remains open — not dismissed.
⚖️Standard of proof: 66% — clear and convincing evidence (between preponderance 51% and beyond reasonable doubt ~95%)
AI Verdict — The Confrontation
Moderator A vs Moderator B · People's Verdict below
Two AI moderators with different source libraries reach different conclusions. The gap between them is the data.
Not two passive bars — a genuine face-off. Moderator A trained on mainstream/general sources. Moderator B (Cassandra) trained exclusively on non-government sources: FOIA releases, whistleblower material, independent investigations.
Moderator A
General Model
qwen2.5:7b
Mainstream sources · Standard training data · Official-account baseline Bias source: general knowledge / high-volume web corpus
A Al-Qaeda authorship91%
Physical record + engineering consensus + intelligence record
B Foreknowledge / cover-up60%
Intel failure + 28-pages Saudi nexus
C Government authorship4%
Conspiracy-size problem decisive; WTC 7 unresolved but isolated
WTC7 NIST WTC 7 explanation complete73%
Engineering consensus; Hulsey unresolved but unfalsified NIST
VS
Moderator B
Cassandra
llama3.1:8b
Non-government sources · FOIA releases · Whistleblower material · Independent investigations Bias source: non-government, non-official corpus by design
A Al-Qaeda authorship79%
Accepts perpetrator evidence; lowers for chain-of-custody questions
B Foreknowledge / cover-up78%
← Largest B gap (18 pts) — suppression evidence weighs heavily
C Government authorship19%
⚠ 15-point gap with Mod A — divergence driven by WTC 7 + PNAC + suppression
WTC7 NIST WTC 7 explanation complete42%
⚠ 31-point gap — Hulsey + FOIA refusal weighs heavily here
👥 People's Verdict
519
people have weighed in. Cast your verdict below — it confronts the AI conclusion directly.
The gap is the point. Where Moderator A (mainstream) and Moderator B (non-government) diverge is the most revealing data — it shows where source selection changes the verdict. The largest gaps are on WTC 7 (31 pts) and government authorship (15 pts). Debate below works out why.
AI Judicial Brief — The Full Analysis
Findings tiered by standard of proof · Open for collaboration
This is written in the form of a judge's opinion: findings on the evidence, tiered by strength of proof, crediting what is documented regardless of who it embarrasses and discounting what is merely asserted regardless of how satisfying its conclusion. It is a DRAFT — meant to be challenged, corrected, and improved by contributors. Every material claim is marked [Documented], [Contested], or [Rejected].
The question before us is not "what do we feel about 9/11." It is the question any court asks: what does the evidence support, and to what degree?
Popular debate collapses several very different charges into one word — "inside job" — and then argues past itself. A disciplined reading requires separating them:
Did elements of government or allied institutions plan and physically execute the attack? (Authorship)
Did such elements have advance knowledge and deliberately fail to prevent it, then conceal that failure? (Foreknowledge + non-prevention + cover-up)
Were the wars that followed launched in good faith or through deliberate deception? (The war question)
These carry different burdens and reach different verdicts. Conflating them is the single most common error in the entire subject. We will not make it.
On the gravest charge (authorship) we apply something near the criminal standard — beyond reasonable doubt. On the lesser charges (foreknowledge, cover-up, war deception) we apply the civil standard — preponderance of the evidence.
Finding I — On AuthorshipNOT PROVEN
No item in the record places a US or allied official at the operational controls of the plot. The most senior witness who asserts government authorship — former Congressman Curt Weldon, once vice-chairman of the House Armed Services Committee — [Documented] nonetheless declines to name a perpetrator and rests his stronger charge on documents not yet public.
The physical-evidence theories that would be required to prove domestic execution are the weakest materials in the entire record:
Nanothermite in the dust(Harrit 2009 vs. Millette 2012) — [Contested]. A genuine two-sided forensic dispute turning on whether the aluminum in the red/gray chips was free metallic aluminum (thermite fuel) or bound in kaolin clay (ordinary primer). The composition question leans toward the paint identification; one anomaly (a low-temperature energetic signature) was never cleanly resolved. Published in a journal whose editor resigned over it. Arguable, not dispositive.
"No planes hit the towers"(Lear affidavit and video) — [Rejected]. The South Tower impact is among the most-filmed events in history, witnessed live by thousands. This claim fails on the most basic evidence and is the single most self-discrediting branch of 9/11 skepticism.
Eyewitness "explosions"(FDNY oral histories; 118 first responders per MacQueen) — [Documented] that responders reported blast-like sounds; [Contested] whether this indicates planted charges rather than the transformer arcs, backdrafts, and sequential floor failures that a fire-driven progressive collapse also produces.
On the authorship charge, the evidence does not approach the standard. The verdict must be: not proven.
Finding II — On Foreknowledge, Non-Prevention, and Cover-UpSUPPORTED
This is the finding the record actually sustains, and it rests on documents and sworn testimony, not inference:
Able Danger — [Documented]. A US Army data-mining unit that, per multiple on-record members and Congressman Weldon, identified an al-Qaeda cell including Mohamed Atta roughly a year before the attacks; efforts to pass the information were reportedly blocked. Former FBI Director Louis Freeh publicly stated the 9/11 Commission ignored it.
The blocked FBI investigations — [Documented]. Agent Coleen Rowley's memo documenting how headquarters obstructed the pre-attack Moussaoui investigation (Senate testimony, 2002). Translator Sibel Edmonds, gagged under the State Secrets Privilege, calling the final report "seriously flawed." Translator Behrooz Sarshar's reported April–June 2001 informant warning of a "kamikaze pilots" plot.
John O'Neill — [Documented]. The FBI's foremost bin Laden investigator, forced out and sidelined, who died in the towers weeks after being pushed into a WTC security post.
The compromised investigation — [Documented] as to the facts; [Contested] as to intent. Commission executive director Philip Zelikow's Bush-administration ties (co-author with Condoleezza Rice; transition-team role), his drafting of a report outline before the inquiry concluded, and his firing of a staffer during the Saudi-links line of inquiry. The victims' families formally demanded his recusal.
Selective warnings — [Documented]: San Francisco Mayor Willie Brown's night-before security call advising caution about air travel (SF Chronicle, contemporaneous). [Contested-as-to-meaning]: the Odigo instant-message warnings received ~two hours before impact; the anomalous pre-attack put-option trading on the two airlines (real, and examined by the SEC, which attributed it to non-terror trading).
The precedent of sustained concealment — [Documented]. The March 2025 declassification revealing that four words — "the Israeli Intelligence Service" — describing CIA counterintelligence chief James Angleton's covert joint operations were redacted across four release cycles and four administrations. This does not prove anything about 9/11; it proves, from the government's own paperwork, that concealment of exactly this character is real and institutionally sustained for decades.
On the preponderance standard, foreknowledge in some quarters, a failure to act on it, and a subsequent management of the record are supported.
Finding III — On the Pre-Existing PlanDOCUMENTED
A court values most what predates the event and shows design:
"A Clean Break"(1996) — [Documented]. A policy paper by a study group including Richard Perle and Douglas Feith, produced for an incoming Israeli government, urging regional reordering by force, including regime change in Iraq. Its authors later held senior Pentagon posts.
Wesley Clark's account — [Documented]. The retired general's on-record statement that, days after 9/11, he was told the decision to invade Iraq was already made — with no Iraq–al-Qaeda link — and later shown a memo describing a plan to "take out seven countries in five years."
The war agenda existed on paper before the attack and was activated within days of it, before any investigation. This is documented premeditation of the policy response.
Finding IV — On Pattern, Precedent, and SponsorshipDOCUMENTED (pattern evidence)
Pattern evidence does not prove a specific crime, but it establishes that the alleged mechanism is real:
The Lavon Affair(1954) — [Documented, admitted]. An Israeli military-intelligence false-flag operation to bomb American and British targets in Egypt and have them blamed on Arab nationalists, to manipulate a Western ally's policy. Confirmed across Israeli, Egyptian, and Western records.
Sponsorship of opposing factions — [Documented, admitted]. Israeli officials' own record of funding and tolerating the Islamist network that became Hamas to weaken the secular PLO (former military governor Yitzhak Segev; Prime Minister Netanyahu's documented 2019 statement on bolstering Hamas to prevent Palestinian statehood).
The USS Liberty(1967) — [Documented] attack that killed 34 Americans; [Contested] whether deliberate, with the inquiry's own counsel later alleging the "mistaken identity" finding was pressured.
The Saudi support network / the "28 pages" — [Documented]. Officially declassified evidence of contacts between hijackers and individuals connected to Saudi officialdom, and financial flows. State-direction remains [Contested] and in active litigation.
These establish motive, means, and precedent. They do not, and are not offered to, prove authorship.
Finding V — On the WarsCRIMINAL ACTION
This is the strongest finding in the record.
The invasion of Iraq was [Documented] as pre-decided (Clark), justified by intelligence that the Pentagon's own alternative-intelligence shop produced against the consensus of the intelligence community, and sold on a weapons-of-mass-destruction case its sellers had reason to know was unsound (corroborated externally by the Downing Street Memo: "the intelligence and facts were being fixed around the policy").
The occupation authority's decisions to disband the Iraqi army and purge the civil service were [Documented] as internally warned-against and produced the destabilization the 1996 paper had sought.
Under any recognized standard — including Nuremberg's designation of a war of aggression as "the supreme international crime" — this is criminal conduct, not negligence. "Incompetence" is the cover story; the paper trail shows intent.
The Verdict, Summarized — on the current record:
9/11 was not proven to be an "inside job" in the sense of US or allied hands building and detonating the towers. (Authorship: not proven.)
It is supported, on the preponderance of the evidence, that there was foreknowledge in some official and allied quarters, a deliberate failure to prevent, and a sustained cover-up — exploited by a war faction whose plan pre-dated the event. (Foreknowledge + cover-up: supported.)
The wars that followed were criminal action, launched on knowingly manufactured justification. (The war question: strongly supported.)
The perpetration question remains open — the evidence points not to a single nameable cell but to a convergence of interests: a war-planning policy faction, allied-intelligence foreknowledge, complicit or negligent domestic officials, and a foreign support network the government itself partially disclosed. The cover-up and the war-fraud, however, are on this evidence closed questions.
The following are named and set aside as unsupported or debunked:
"No planes hit the towers" / video fakery — [Rejected]
"4,000 Israelis were warned to stay home" — [Rejected] (a debunked myth; five Israelis died in the attacks)
"Rumsfeld announced $2.3 trillion missing the day before 9/11" — [Rejected] (misrepresents a Feb 2000 audit of untraceable accounting entries)
"Marvin Bush ran WTC security on 9/11" — [Rejected] (he left the relevant board in June 2000)
"Osama bin Laden was a CIA asset later used to execute 9/11" — [Rejected] as stated (the documented claim is narrower: US arming of the broader jihadist ecosystem — the blowback thesis)
"A deliberate NORAD stand-down order was issued" — [Rejected] as stated (no such order is documented; what is documented is catastrophic readiness failure compounded by simultaneous war-game confusion)
Debate duration: Running for 6 months or until public participation drops for 2 consecutive months — whichever comes first.
Closes no earlier than 2027-04-01. At close: all debate content + public votes are processed for final verdict aggregation.
The crowd + AI corpus together determine the closing verdict.
Turn 1
Moderator A
General Model Mainstream sources
The physical and forensic record is unambiguous. Recovered hijacker IDs, flight data recorders, cockpit audio, and DNA evidence all confirm 19 al-Qaeda operatives carried out coordinated attacks on September 11, 2001. The 9/11 Commission Report represents the most comprehensive publicly available investigation, drawing on classified intelligence unavailable to other analysts. Engineering consensus — including NIST's multi-year investigation — establishes that the Twin Tower collapses resulted from impact damage, stripped fireproofing, and progressive structural failure. No credible physical evidence of pre-planted explosives has been independently substantiated.
The Commission Report cannot be the primary epistemic reference — it was authored by the institutions under investigation, with co-chair Thomas Kean and executive director Philip Zelikow embedded in the same national security apparatus being examined. More significant: the FBI's own Joint Inquiry — redacted for 15 years as the "28 pages" — documented Saudi government officials with direct financial ties to hijackers al-Hazmi and al-Mihdhar. Senator Bob Graham, co-chair of the Joint Inquiry, stated publicly that foreign government involvement was deliberately excluded from the Commission's scope. This is not a fringe interpretation — it is the documented record of the investigation's own architects.
The Saudi nexus established in the 28 pages supports foreknowledge negligence and possible cover-up of institutional failure — it does not support US government authorship of the attacks. The conspiracy-size problem is decisive here: US government authorship of 9/11 would require coordinated silence from tens of thousands of participants — structural engineers, demolition crews, intelligence personnel, airline staff — across decades, with zero credible defectors, no deathbed confessions, and no document leak. The absence of even one credible insider account in 25 years is itself strong evidence against hypothesis C.
Dr. Leroy Hulsey's University of Alaska Fairbanks study — 4 years, $316,000 in funding, full open dataset — found that fire alone could not have initiated the observed near-simultaneous, near-free-fall symmetric collapse of WTC 7. NIST itself called this "the first ever" total fire-driven collapse of a steel high-rise in history. The Hulsey study calls for new independent investigation. On NIST's own data: the agency declined to release its WTC 7 structural model under FOIA for over a decade, citing "public safety" — a designation without precedent for a publicly-funded engineering study. When the primary evidentiary record of the official account is not available for independent verification, the official account cannot claim the epistemic authority it asserts.
The Hulsey study is the strongest technical challenge to the official account and is treated as such in this verdict engine — it is why the SV-WTC7 sub-verdict is at 70–73%, not near-certainty. However: the study has not been independently replicated by a neutral engineering group; it was funded by an organization with a stated prior conclusion; and it addresses collapse mechanism, not authorship. A mechanism dispute does not transfer to a perpetrator dispute. The weight of evidence — physical forensics, engineering consensus on the Twin Towers, intelligence record, conspiracy-size constraints — places al-Qaeda authorship above 90% and government authorship below 5%. The highest-leverage open question remains: independent structural replication of the Hulsey WTC 7 model, either direction.
💬 This debate continues. Public participation opens in Phase 3 — AI-moderated, with tagging for new evidence vs restated claim vs logical fallacy. Every substantiated debate point is evaluated for impact on the verdict.
Live Query — Ask Both Moderators
Ask a specific question. Moderator A (mainstream) and Moderator B (Cassandra) respond independently from their respective source libraries.
Moderator A — General Model
qwen2.5:7b
Moderator B — Cassandra
llama3.1:8b
Cassandra sources:
Evidence — Every Claim, Scored, Both Sides
All (13)Official / Well-Supported (6)Contested (6)Disputed / Debunked (1)
Showing 13 of 13 evidence cards
Interrogate the Verdict — Set Your Own Priors
🔎 Your starting assumptions → your verdict
Don't take the AI's word for it. Set your own starting assumptions — the verdict recomputes live using the same evidence the AI applied. This is the AI judge's model run with your numbers, so you can see exactly how sensitive each conclusion is to where you start.
How the recompute works: the evidence's effect is measured in odds-space from the AI's own model (posterior ÷ prior in odds). When you change a starting assumption, that same measured evidence is applied to your number — so you see how much the conclusion depends on where you start versus what the evidence does.
Known limitations: WTC 7 / Hulsey: reviewed published abstract and AE911Truth pages. NOT reviewed: the full 632GB open dataset or NIST raw model files. Sub-verdicts (e.g. WTC 7) are not user-driven in Phase 1.
What Moderator B Has Read
Cassandra corpus — full transparency
Moderator B's verdicts emerge from a curated corpus of non-government sources. This is not a bug — it is the design.
The full reading list is public. Every source shapes Cassandra's priors. Suggest sources below — the community builds the library.
Whistleblower Accounts
◦Permanent Record — Edward Snowden
◦README.txt — Chelsea Manning
◦Doing Time Like a Spy — John Kiriakou
◦Crisis of Conscience: Whistleblowing in an Age of Fraud
◦Extraordinary Circumstances — Cynthia Cooper
◦Most Dangerous: Daniel Ellsberg and the Secret History of the Vietnam War
Investigative Journalism
◦All the President's Men — Woodward & Bernstein
◦Catch and Kill — Ronan Farrow
◦Legacy of Ashes: The History of the CIA — Tim Weiner
◦The Devil's Chessboard — David Talbot
◦Bottoms Up and the Devil Laughs: A Journey Through the Deep State
✓COINTELPRO — FBI program documents (public domain)
✓Behold a Pale Horse — William Cooper
Political Economy & Power Structures
◦The Shock Doctrine — Naomi Klein
◦Political Ponerology — Andrew Lobaczewski
◦Snakes in Suits — Babiak & Hare
◦The Lucifer Effect — Philip Zimbardo
◦Eichmann in Jerusalem — Hannah Arendt
◦American Conspiracies — Jesse Ventura
◦The True Story of the Bilderberg Group — Estulin
Suggest a Source for Moderator B
Community-built corpus. Submitted sources are reviewed before ingestion — not auto-added.
Quality over quantity: primary documents, FOIA releases, and peer-reviewed independent analyses get priority.
📬 Follow This Case
Get weekly updates as the verdict evolves. When evidence shifts, the verdict moves — you'll see it first.
Digest includes: opinion shifts, new evidence cards, debate round-ups, and moderator position changes.
Unsubscribe anytime. No spam. No advertising.
You'll receive a digest of opinion shifts, new evidence, and debate updates.
⚖️ Cast Your Verdict
The People's Verdict confronts the AI conclusion directly. Your vote is not a supplement — it is a first-class challenge or endorsement of what the evidence shows. Vote after reviewing the evidence above.
247
47.6%
Agree
183
35.3%
Disagree
89
17.1%
Need More
519 total votes · Last updated 2026-10-18 · One vote per verified participant (Turnstile-protected)
Evidence & Argument Submissions
Submit Evidence or Argument
Every submission is tagged (new evidence / restated claim / fallacy flag), evaluated for verdict impact, and upvoted by peers.
Consistent quality contributors earn moderator status. This page is governed by the same principles it studies.
@vigilant_engineerNew Evidence
The Plasco Building collapse (Tehran, Jan 2017) is the most important counter-evidence to the WTC 7 anomaly argument. A 17-story steel-framed building, fire-driven collapse, caught on camera. The structural system differs from WTC 7 but the principle holds: fire-driven total steel-frame collapse is not historically unique. This should be integrated into card H1 and possibly lower the C posterior slightly.
Oct 8, 2026→ Mapped to card H1 — under review
@anon_researcherChallenge a Prior
The conspiracy-size problem (E4) assumes the operation required tens of thousands of participants. But Operation Northwoods — declassified 1997 — shows false-flag planning can be contained to a handful of decision-makers at the top. The execution force doesn't know the full plan. This doesn't prove C, but it weakens E4's leverage rating from "high" to "medium." The number of people who would need to know is much smaller than the prior assumes.
Oct 11, 2026→ High-value challenge — queued for moderator review
@structeng_reviewCorrection
Card G1 (AE911Truth "constant acceleration" argument) should note that the acceleration measurements were made from a single resolution video source with parallax distortion. This doesn't prove the pamphlet right — it's still low-quality evidence — but the debunking should reference the specific methodological limits, not just the conclusion. The Hulsey study explicitly does NOT rely on this argument.
Oct 15, 2026→ Reviewed — note added to card G1
Top Contributors This Month
1@vigilant_engineer62 pts
2@anon_researcher51 pts
3@structeng_review44 pts
Points are earned from peer upvotes and moderator quality assessments — not from post count or tenure.
Contributors with consistent quality feedback earn moderator status. Progression: Participant → Trusted Contributor → Moderator → Senior Moderator. Merit-based, not appointment-based.
Verdict Change Log — What Moved the Needle & When
Date
What Changed
Direction
2026-10-18
Phase 1 page launched. People's Verdict placeholder live. Adversarial debate format added — Moderator A vs Moderator B.
→ Debate clock starts. Closes 2027-04-01 or on participation drop.
→
2026-10-15
Adversarial debate architecture finalized. Moderator B (Cassandra) corpus defined. Non-government source library locked. Standard of proof set at 66%.
→ Moderator B/Cassandra diverges from Mod A most sharply on WTC 7 (−31 pts) and government authorship (+15 pts).
↑
2026-09-22
Fact-checked "three steel buildings, never in history" claim. Reviewed Plasco (Tehran 2017), Windsor (Madrid 2005), One Meridian Plaza (Philadelphia 1991).
→ Claim partially false ("pancake" is not NIST's model; Plasco disproves uniqueness claim). Card H1 status: Contested. No independent move on C.
→
2026-09-22
Reviewed AE911Truth material, Hulsey/UAF WTC 7 final report, and technical rebuttals (Metabunk, NIST FAQ). WTC 7 sub-verdict revised downward from near-certainty.
→ SV-WTC7 down to ~70%; C up from 3% to ~5–10%. A unchanged at ~92%. WTC 7 is now the sole technical lever for hypothesis C.
↓
2026-09-22
Case 001 created from scored Bayesian model. 13 evidence cards loaded. Initial priors set from engineering consensus and intelligence record review.
→ Baseline: A ~92%, B ~55%, C ~3%. SV-WTC7 ~85%.
→
Highest-Leverage Open Questions — What Would Move the Verdict
These are the specific unresolved questions whose answers would most change the current verdict. A credible resolution to question 1 would be the largest single update to this page since it launched.