Corrections Log
A public record of factual corrections made to content on this site. Nothing builds trust faster than demonstrating a willingness to be wrong.
Our Corrections Policy
What gets corrected
Factual errors: wrong dates, misattributed quotes, incorrect document references. Broken source links. Misidentified individuals.
What doesn't
Editorial disagreements, differing interpretations of ambiguous evidence, differences of opinion on confidence ratings. These are addressed through our methodology framework, not corrections.
Corrections Issued
28 corrections to date, newest first. Each one names the page, what was wrong, and what it says now. This list is generated from the update records attached to each page, so a correction can't ship without appearing here.
Correction: the Library Rated Schwinger-Limit Energy-Extraction Claims Definitive and Strong
Five generated claims in the knowledge library asserted that time manipulation is achieved by breaking the Schwinger limit, that breaking it extracts virtual photon pairs or energy from the vacuum, that a propulsion technology involves doing so, and that the orbs build field strength to break it. One was rated definitive and four rated strong or medium, while this guide and the Salvatore Pais page both treat the same propositions as unproven. All five are now rated speculative in master_reference.json and in the per-video analysis files that also carried them. The claim that the Schwinger limit describes vacuum breakdown under an intense field is correct physics and remains definitive.
Correction: The Site Published a Person Page for a Name Ashton Forbes Invented on Air
The evidence page described a "'Chris Cernohav' 119-page Russian engineering specification for fusion-powered electromagnetic propulsion", and the library carried a matching person entry, a claim rated definitive, and a scientific reference. Four things were wrong. The engineer is Valerijs Cernohajev (1958-2019) and he was Soviet-Latvian, not Russian. "Chris" is Forbes' own placeholder: on the tape he says "How am I supposed to say this guy's name? Can we just give him an English name? I'm just going to call him Chris", and the auto-captions rendered the surname as Cernohav and Sernohav. The pages are personal handwritten works dated 1980 to 2007, not an engineering specification, and only about half concern propulsion. And the card cited two videos as separate sources when B4bcNs_kFAc is a re-upload of vcrzxWCu-WU, both 13:19. The claim's confidence drops from definitive to speculative, the person page moves to valerijs-cernohajev, and the card now carries the provenance finding: chain of custody runs entirely through the author's family, who also publish the books and the analysis, with no third-party authentication.
Correction: The Gaetz Card Attributed a Claim to His Clearance That He Sourced to Books
The evidence page carried Matt Gaetz's remark about plasma propulsion being sent to Lockheed Martin as a single-source Speculative item citing a Forbes video, and noted that a direct citation would strengthen it. The primary has now been read: Benny Johnson's interview of 31 March 2026 (rgFYRfEltok). In it Gaetz attributes the Lockheed reverse-engineering account to "the commentary of people who discussed the reverse engineering theories" and adds "that's been chronicled in a number of books and I studied that literature", so it is not firsthand. What he does attribute to his own classified briefings is stronger and had been missed: radar of multiple craft merging into a single return and separating repeatedly, and the same behaviour in formation. The card now carries that claim, rated Established, cited to the interview at 4:19 and 5:50.
Correction: The Iran Rescue Capability Card Gave the Wrong Date and the Wrong Newspaper
The "Capability Reveals" card on the evidence page dated the Iran rescue to 2026-04-07 and credited the Ghost Murmur disclosure to the Wall Street Journal. Both are wrong. The F-15E was shot down on Friday 3 April 2026, the pilot was recovered the same day, and the weapons systems officer was recovered late on Saturday 4 April (EST); the White House briefing where Trump described the camera and CIA Director John Ratcliffe cited "exquisite technologies that no other intelligence service in the world possesses" was Monday 6 April. The Ghost Murmur account was published by the New York Post on 7 April, which is where the wrong date came from. The Wall Street Journal was cited in this story for MQ-9 Reapers firing on Iranian forces near the hide site, a different claim. The card is now split in two: the on-the-record capability statements, and the separately sourced and contested biomagnetic claim.
Correction: the Iran rescue was 3 to 5 April 2026, not 7 April
The page dated the rescue to 7 April 2026 in its timeline, its section heading and its opening paragraph. That is the date the New York Post published the Ghost Murmur story. The F-15E was shot down on Friday 3 April, the pilot was recovered the same day, and the weapons systems officer was recovered late on Saturday 4 April (EST), with the White House briefing following on Monday 6 April. The timeline now separates the operation from the reporting. A second correction: the page stated as fact that the rescue used Gorgon Stare imagery and Palantir Maven filtering, which no official has confirmed; that passage now reads as the inference it is.
Correction: The Vanishing EP01 Overstated the MH370 Radar Altitude Evidence
At 3:50, EP01 (youtube.com/watch?v=HuRLNI0cmkY) says the military radar showed a climb to 45,000 ft and a descent to 23,000 ft, and that "no explanation has ever been offered for this profile", then argues from that against pilot suicide and mechanical failure. Two explanations had been offered. Angus Houston, who led the search, stated on 23 June 2014 that "the primary radar data pertaining to altitude is regarded as unreliable" because Malaysian radar was not calibrated precisely enough to determine altitude; Malaysia's 2018 Safety Investigation Report (p. 4) says the blip-to-blip altitude and speed changes were "well beyond the capability of the aircraft" and that "the only useful information obtained from the Military radar was the latitude and longitude position", adding on p. 6 that the profile was "not achievable even after repeated simulator sessions". The 45,000 and 23,000 figures come from March 2014 press reporting of leaked raw radar and appear in neither official report. The same segment also says the aircraft was "last detected near the island of Penang at 2:22 a.m.", which welds the time of the final radar return to the location of an earlier one: Penang was 01:52:31 MYT, and the last return was 02:22:12 MYT, 10 nm past waypoint MEKAR and about 195 nm from Butterworth. The video's source script and the altitude-graph composition are corrected so a re-render cannot reproduce either error.
Correction: Flight Path Page Stated Radar Altitudes the Official Reports Discarded
The flight-path page presented MH370's military-radar altitude profile (a climb to roughly 45,000 feet and a descent to 23,000 feet) as recorded fact. Neither investigation used those values: Malaysia's Safety Investigation Report calls the blip-to-blip altitude and speed changes "well beyond the capability of the aircraft" and says only latitude and longitude are reliable, and the ATSB built its route analysis on the positions. The figures themselves come from March 2014 press reporting of leaked raw radar and appear in neither official report; Angus Houston, who led the search, stated on 23 June 2014 that "the primary radar data pertaining to altitude is regarded as unreliable". The paragraph now gives the reported range with the reports' own assessment of it. The same paragraph also attributed to Ashton Forbes the view that this altitude profile indicates evasive manoeuvring or an interception; his recorded position is close to the opposite. He says of the radar figures, "I don't really believe the radar data though" (22 January 2024), reads the altitude change as an emergency descent following a cargo fire rather than an evasive climb, and argues against the dramatic-climb framing directly ("30,000, 32,000 feet is not high altitude"). That attribution has been removed. Same correction applied to the site timeline entry, which described the military track as an unidentified aircraft when section 1.1.4 of the Malaysian report states the air force recognised it as MH370 on the day and chose not to intercept it.
Five pipeline entries on the "Ion Pumpy Thing" described magnetic nozzles the source never mentions, and cited a paper that does not exist
The glossary, two theory entries and one claim entry all attributed "magnetic nozzles" to the schematic. The word "nozzle" appears zero times in the transcript of the only source video (70M-6hMUZEk), in any spelling. Separately, a scientific-reference entry cited "George Miley, 'A Neutronic Fusion Spacecraft Architecture' (2001)" and marked it fringe. No such paper exists. It is a garbled rendering of "Aneutronic Fusion Spacecraft Architecture", a NASA Innovative Advanced Concepts Phase I study whose principal investigator was Alfonso Tarditi, with John Scott and Miley as collaborators, final report September 2012. All five corrected in master_reference.json while tracing the schematic for the section above.
Withdrawing the 2026-05-12 Lamoreaux correction: it was wrong, and it replaced a correct fact
On 2026-05-12 this page changed Lamoreaux's 1997 affiliation from University of Washington to Los Alamos National Laboratory, and the correction notice said it did so "per the published PRL paper byline". The byline says the opposite. Physical Review Letters 78, 5 prints "S. K. Lamoreaux*, Physics Department, University of Washington, Box 35160, Seattle, Washington 98195-1560", and the starred footnote reads "*Present address: Los Alamos National Laboratory, H803, Los Alamos, NM 87545". Los Alamos was where he was moving, not where the torsion-pendulum measurement was done. The page now says University of Washington and explains the footnote so this doesn't get "corrected" a third time. Found by the independent fact-check pass on the 2026-08-14 Dynamic Casimir Effect expansion. Note for anyone auditing the channel: videos generated between 2026-05-12 and this date may repeat the Los Alamos attribution.
The page said the 2011 experiment showed vacuum energy "can be extracted". It does not show that
The Dynamic Casimir Effect section closed with "This shows the energy isn't just theoretical; it can be extracted under extreme conditions." That reads the physics backwards. Wilson et al. drove a SQUID with microwave power, and the emission is parametric down-conversion: a drive photon splits into a pair whose frequencies sum to the drive frequency, at efficiency far below unity. The paper states no energy budget, but it is explicit that the screening currents "act against the vacuum fluctuations", which is the division of labour that matters: the drive supplies the energy and the vacuum supplies the fluctuations. Nothing is extracted from the vacuum. The claim is removed and the section now states the boundary explicitly, because the zero-point material elsewhere on the site depends on this distinction being drawn honestly.
The 2011 experiment did not move a mirror, and the page overstated how fast its boundary moved
The page described the Chalmers result as moving "a mirror fast enough (a significant fraction of the speed of light)". No matter moved: the team modulated the inductance of a SQUID terminating a coplanar transmission line at roughly 11 GHz, changing the line's electrical length. The paper's abstract gives the rate as "a few percent of the speed of light", about v/c = 0.05. The effect is genuinely relativistic and the authors call that velocity very large, so the fix is to quote the figure rather than to deny the physics. Corrected against Wilson et al., Nature 479, 376-379 (2011), DOI 10.1038/nature10561.
The page said the Dynamic Casimir Effect is "literally extracting energy from the vacuum". It isn't
The zero-point energy section described a boundary "accelerated to relativistic speeds" converting virtual photons into real pairs, "literally extracting energy from the vacuum", in the site's own voice rather than as an attributed claim. The extraction half is wrong: Wilson et al. drove a SQUID at roughly 11 GHz with microwave power, and the emission is parametric down-conversion at efficiency far below unity. The vacuum supplies the fluctuations and the drive supplies the energy. The rate is also overstated: the paper's abstract gives the boundary as changing at "a few percent of the speed of light", about v/c = 0.05, though the effect itself is genuinely relativistic and the authors call that velocity very large. Corrected against Wilson et al., Nature 479, 376-379 (2011), DOI 10.1038/nature10561, and now links the Casimir 101 guide's full treatment. Found by the Phase 1.4 regression check on inbound links while expanding that guide.
Spacetime stiffness entry stated the relation backwards
The Frequency-Dependent Spacetime Stiffness entry said that spacetime's Young's modulus "decreases as the frequency of gravitational waves increases, making high-frequency manipulation feasible for teleportation." That is inverted. Tello and Strong, "Breaking spacetime", European Journal of Physics 46(5) 055602, 2025 (doi 10.1088/1361-6404/ae04b4), treats the modulus as frequency dependent at roughly 10^20 times that of steel at 100 Hz and concludes that reaching the stiffness required would need frequencies beyond the Planck frequency. Stiffness rises with frequency, so high-frequency manipulation is harder, not easier. The entry now states Forbes' claim as his claim and gives the published relation alongside it. The arithmetic in the source video was also wrong: 100 Hz to 1 GHz is a factor of 10^7, which squares to 10^14, not the 10^18 asserted. Entered the site through the S6noGwFjJic video ingest and was live as a hidden stub. Found by the August 2026 roadmap audit.
The same inverted stiffness claim survived in the claims layer
The correction above fixed the theory entry but not the underlying claim entity, which carried the identical inverted physics and was linked into 43 theory records. It read "Spacetime stiffness can be reduced by increasing the frequency of gravitational waves, making teleportation physically possible at lower energy thresholds." The first pass corrected data/pipeline/analysis/S6noGwFjJic_analysis.json and missed the copy in master_reference.json, which is the file the claim entity is generated from, so the wrong version kept regenerating on every build. Both now carry the same wording: Forbes' claim stated as his claim, with the published relation from Tello and Strong alongside it. Also corrected here: the evidence line credited the GASER concept to "Gary Stevenson", which is both the wrong spelling and the name of a British economist. It is Gary V. Stephenson, and the GASER originates with Giorgio Fontana.
Four source citations in the surveillance section displayed no text
The Sources block under the underwater-surveillance section passed its four citations with a 'text' field, but the component reads 'name'. Each one rendered as an empty link: a bullet, an external-link icon, and no visible label. The URLs were intact and still clickable, so nothing stated on the page was untrue, but a reader could not see what was being cited and the section appeared to be unsourced. The four sources are SOSUS, India's Indian Ocean sensor network, a 2022 arXiv paper on submarine navigation using neutrinos, and an ASPI analysis of submarine-detection prospects. Live from 2026-03-16, when the section was added, until this fix. Found by running 'astro check', which had been reporting it as a type error the whole time; the check was not in the build chain, so nothing surfaced it. It is now.
Correction: the FBI tip is not in the court record
This page stated that "court records indicate the investigation was triggered by an FBI tip about Lin's foreign contacts." That was wrong, and had been live since February 22, 2026. The 57 pages released under FOIA 2026-NavyFOIA-005116 do not mention a tip anywhere. The detail comes from the Navy's insider-threat case study on Lin, which says "a tip to FBI in early 2014" led to a joint FBI/NCIS investigation. The page now attributes it there and says plainly that the released court documents don't carry it. The joint FBI/NCIS operation itself is separately confirmed in the release: the appellate opinion has NCIS and FBI agents making the arrest, and the Stipulation of Fact has Lin describing an undercover FBI agent. Found during an independent fact-check of the August 2026 FOIA update.
Skeptic's Case Refreshed Against New Research
Updated the eight-argument skeptic's page with findings from the Lin FOIA records, the debris barnacle analysis, the satellite video forensics deep dive, and the concluded 2026 ocean search. Corrected the stale reference count and the Lin conviction wording.
Iran Clip Section Corrected After Frame-by-Frame Analysis
Section 5 (DOW-UAP-PR050) revised. Earlier text said all four contrasts trailed wakes; frame-by-frame analysis shows three objects interacting with the surface (wakes) plus one wakeless object crossing far faster (enters lower-left ~2s, gone by ~3.5s) that can't be resolved at 640x360. The official 'four UAP formation' count is not supported: three clear objects, not four. We stopped short of calling the three boats: a surface wake alone doesn't prove a hull, since any moving surface disturbance (including a low craft's downwash) makes the same wake. The robust finding is that three interact with the water and the fourth does not. Updated the evidence table, the signature-vs-mechanism row, and the source citations.
DOI corrected to resolvable form
Replaced the non-resolving citation 10.1103/PhysRevResearch.8.013264 with the live DOI 10.1103/l8y7-r3rm for White et al. (2026), Physical Review Research 8, 013264, and linked it.
Elizondo resignation dates clarified, Wilson-Davis meeting date sourced
Timeline entry now distinguishes Elizondo's June 2017 internal notice of intent from his October 4, 2017 effective resignation per Defense Department records cited in Politico and NYT reporting. Eric Davis key-people entry adds the October 16, 2002 Wilson-Davis meeting date and EG&G Special Projects (Las Vegas) location per the leaked Davis longhand notes hosted at TheBlackVault. Triggered by NotebookLM video QA audit 2026-05-12.
Pais Patent Date Corrections (US10,144,532 / US20190058105A1 / US20190295733A1)
paisPatents array corrected against Google Patents. US10,144,532 (inertial mass reduction craft) expiration corrected from Jan 9, 2023 to Dec 4, 2022 (four-year maintenance window from Dec 4, 2018 grant). US20190058105A1 (room-temperature superconductor application, abandoned) filing date corrected from 2018 to Aug 16, 2017 (published Feb 21, 2019). US20190295733A1 (plasma compression fusion application, abandoned) filing date corrected from Mar 2019 to Mar 22, 2018 (published Sep 26, 2019). Triggered by NotebookLM video QA audit 2026-05-12.
Pais Patent Date Corrections (US10,144,532 / US20190058105A1 / US20190295733A1)
Patent status table corrected against Google Patents. US10,144,532 (inertial mass reduction craft) expiration corrected from Jan 2023 to Dec 4, 2022 (four-year maintenance window from Dec 4, 2018 grant). US20190058105A1 (room-temperature superconductor application, abandoned) filing date corrected from 2018 to Aug 16, 2017 (published Feb 21, 2019). US20190295733A1 (plasma compression fusion application, abandoned) filing date corrected from Mar 2019 to Mar 22, 2018 (published Sep 26, 2019). Triggered by NotebookLM video QA audit 2026-05-12.
Lamoreaux 1997 affiliation changed to Los Alamos (SUPERSEDED, see 2026-08-14)
Steven Lamoreaux's 1997 Casimir force measurement affiliation changed from University of Washington to Los Alamos National Laboratory, citing the published PRL paper byline (DOI 10.1103/PhysRevLett.78.5). Triggered by NotebookLM video QA audit 2026-05-12. This change was itself incorrect and was reversed on 2026-08-14; the entry is kept because the wrong version was live for three months.
TAE and Helion fuel cycles split apart; company name corrected
The page treated TAE Technologies and Helion Energy as pursuing the same fuel cycle. TAE runs proton-boron, Helion runs deuterium-helium-3, and the two were separated. The TAE company name was also corrected. Part of the fusion-cluster currency audit.
Helion fuel cycle corrected; direct-conversion claim hedged; Pais patent status fixed
Same Helion D-He3 correction as the fusion hub. The direct-energy-conversion claim was hedged to match what the companies actually publish, since HB11's current site describes a conventional steam cycle rather than direct alpha-particle conversion. The 'as recently as 2018' framing of Navy patent activity was replaced with the 2019 Pais plasma-compression fusion application, noting it has since been abandoned. A note was added to the Polywell glossary entry recording that EMC2 is still active.
Helion's fuel cycle corrected: D-He3, not p-B11
The page grouped Helion Energy with TAE Technologies under proton-boron fusion. Helion runs a deuterium-helium-3 cycle, and its Polaris machine was running D-T as of February 2026 to validate scaling. Corrected during the fusion-cluster currency audit that followed a viewer correction on the TAE/HB11 video.
Primary-paper citation corrected: wrong DOI and wrong third author
The page cited Miley, Murali and Hora in Laser and Particle Beams 34(4) under DOI 10.1017/S0263034616000604, which resolves to an unrelated Bremsstrahlung X-ray paper by Rhee et al. in the same issue. Corrected to Miley, Hora and Kirchhoff, Journal of Physics: Conference Series 717 Article 012095 (IFSA 2015 proceedings), DOI 10.1088/1742-6596/717/1/012095, arXiv 1511.03231. A second citation, 'Hora/Miley et al. 2023 Applied Sciences', was corrected to Margarone et al. 2022 (DOI 10.3390/app12031444) plus the McKenzie, Batani and Hora 2023 review in Journal of Fusion Energy. Both errors were introduced at the writing stage and appear nowhere in Forbes' video. Found by the retroactive ARTICLE-PROCESS audit of the April 13 batch; live for 18 days.
Solve for X date corrected; Lockheed CFR section expanded
Corrected Solve for X date (Feb 7, 2013, not 2014), added US 9,959,942 patent and Trimble's Aug 28, 2023 byline, and surfaced Chase's UnLAB resume describing the program as 'breathing on its own' against Lockheed's cancellation framing.
Claims That Went Stale
These aren't errors. The page was accurate when it published and the outside world moved: a company changed its reactor design, raised another round, or shut a programme. We log them separately because filing them as corrections would both overstate how often we get things wrong and blur what a correction means. Time-sensitive third-party claims carry a currency sidecar and get re-checked on a schedule.
Fourth PURSUE Release and the FY2025 AARO Report Added, Plus the Counter-Reading
The disclosure timeline stopped at the 12 June 2026 third tranche and had gone two months stale. Added the fourth PURSUE tranche of 10 July 2026 and AARO's FY2025 annual report of 21 July 2026: 319 UAP reports received between 2 June 2024 and 30 May 2025, no verifiable evidence of extraterrestrial technology, and an open maritime case off the Virginia coast involving roughly 100 airborne objects. Also added a closing paragraph acknowledging the opposing reading of the same record, since the 1953 Robertson Panel file that arrived inside a transparency release is itself evidence that the United States has run a public-perception programme on this subject before.
TAE funding figure refreshed to $1.5 billion
TAE Technologies' own materials now state more than $1.5 billion in private capital, up from the $1.3 billion figure this page carried since the May 2026 currency pass. Figure updated and re-date-stamped. Found during a triage of 17 stale currency rows; the shallow regex check had been passing because it matched the company name rather than the number.
TAE funding figure refreshed from the 2024 number
The page cited 'as of 2024 it has raised over $1.2 billion'. TAE's own materials now state more than $1.5 billion in private capital. Figure updated and re-date-stamped to August 2026. This page had no currency sidecar, so nothing was watching the claim; it was caught by a manual sweep for stale funding figures rather than by the automated check.
Series G funding update ($465M, $15.5B valuation)
Updated the funding figures for Helion's June 4, 2026 Series G: $465M raised at a $15.5B post-money valuation, led by Thrive Capital, total raised roughly $1.5B, with Orion (first power plant) under construction at Malaga, WA. Prior figures (Series F, $425M, $5.4B, Jan 2025) retained as history. All funding figures now carry explicit as-of dates so they read correctly once superseded, and a currency-claims.json sidecar was added to flag the next round automatically. Verified against Helion's June 4 2026 newsroom release.
EMC2 Fusion is still active; "ceased around 2019" was wrong
The page described EMC2 Fusion as having ceased operations around 2019. The company is still active: a TOFE 2024 presentation, ongoing DOE FPNS work, and emc2fusion.com carrying news dated October 2025. Corrected during the fusion-cluster currency audit.
TAE programme status and funding refreshed against the company's current materials
TAE Technologies dropped the Copernicus device in November 2025 and repositioned around Norm and Da Vinci, so the page's Copernicus framing no longer described the programme. Funding updated from $1.2 billion to $1.3 billion plus the December 2025 TMTG merger at $6 billion. Triggered by a viewer correction on the TAE/HB11 video, which found HB11 described as direct alpha-particle conversion when the company's current site describes a conventional steam cycle.
Report an Error
Found a factual error on the site? We want to know. Email us the page URL and a description of the error, and we'll review it.
Email a correction