by Cory Rove
Last updated: Aug 16, 2026
A Series · 4 parts
A document-grounded investigation into the COVID-19 origins question, held to one standard: what the record establishes, what remains genuinely contested, and where the loudest version overreaches. Every source is in the evidence base.
Read the series →Original Reporting
The decisions being made in the name of We the People — reported here from the primary record. Start here.
The COVID origins fight was not Anthony Fauci's first argument with critics who turned out to have a point. It was his fourth decade of them. What changed, this time, is how it ended: with a pardoned man answering a hundred questions by refusing to speak.
In early 2020, a handful of scientists privately weighed whether the virus looked engineered — and then published a paper telling the public the question was closed. The gap between what they said in private and what they printed is the whole story of how a consensus gets manufactured.
The people most certain, most quickly, that no laboratory could be involved in COVID's origin were often the same people whose money and reputations were tied to the laboratory. That is not a conspiracy. It is a conflict of interest, and it is documented in the grants.
For two years, one sentence ended every argument about where COVID came from: the virus shows no signs of engineering, so it must be natural. This is the story of a patent that says otherwise — and of the difference between what that patent proves and what it doesn't.
Ken Paxton may have voted from an address he'd left. The rival who made it a scandal opposed the very laws that would catch it — and stands accused of the same thing. Held to one standard, both stories shrink to what they are: contested, single-instance, and unproven.
Newly released CBP records show a formal targeting action built around Peter Daszak's return from China — killed by an FBI request that never explains why.
The national noncitizen-registration number has a real disclosure problem. New Jersey just proved the mechanism behind it is real. Both of those sentences are true — and most coverage picks one.
The five scientists who told the world a lab origin wasn't plausible spent February 2020 privately estimating it at 50-50.
Puerto Rico's property tax agency patched a hole that exposed roughly one million Social Security numbers, then told the press — and the public — that no hole ever existed.
The murder of a 27-year-old DNC staffer became one of the most weaponized stories of the 2016 cycle — and the documented record is both more settled and more troubling than either side of the debate will admit.
The fight over what SNAP dollars can buy is, at its core, a fight over who controls access to a $100 billion market — and who gets to write the eligibility definition.
ProPublica's reporting on undisclosed Chinese stakes in SpaceX is worth reading carefully — not primarily as a transparency story, but as a live test of whether post-reform CFIUS rules actually close the private secondary-market loophole they were designed to close.
The Record
Subjects the press got wrong or wouldn't examine on the merits — the newest documented item in each. Reporters are scored on whether they engaged this record, not on the conclusion they reached.
Latest
FBI (moderate confidence) and DOE (low confidence) assessed a lab origin as plausible/likely
Declassified agency assessments, 2023more →Latest
Steele dossier sourcing discredited; origins of the 2016 assessment
Durham report, 2023; DNI Gabbard declassifications, 2025more →Latest
Mesa County (CO) voting-system forensic images cited as showing irregularities
Mesa County forensic reports; Tina Peters case recordmore →Every article is read by an analyst model that separates the fact layer (verifiable claims — numbers, dates, named sources, documents) from the narrative layer (adjectives, framing, omissions). From that it extracts: fact density, credibility, opinion ratio, the loaded words used, the perspectives left out, buried leads, and misleading framings.
Quality rewards fact density (30%), credibility (30%), and facts surfaced (15%); it deducts for opinion load (15%), omitted perspectives (10%), misleading framing (10%), and buried leads (5%). Spin is the mirror — opinion (30%), low credibility (20%), loaded language (20%), misleading framing (15%), omissions (10%), low fact density (5%) — and is amplified 25% when the piece carries a partisan lean.
Both are then percentile-rankedagainst every other story that day, so a 90 means it is more complete than 90% of that day’s coverage. The ranking is relative, and comparable, rather than bunched in the middle. The formula is published here deliberately. A reader is owed the method, not just the verdict.
The Daily Edition
The day's reporting, ranked and sourced, in your inbox each morning. No spin, no paywall.
Daily Brief — Deep Prompt
What deserves more attention, what the counter-narrative surfaces, and where the press veers off.
Credible reporting whose most important fact got buried
Buried lead: The confirmation of Ebola in a sixth province
Buried lead: The mental health struggles of sailors aboard the USS Abraham Lincoln.
Buried lead: The UN’s $1.7bn humanitarian appeal for Afghanistan was only 25 percent funded.
Surfacing what mainstream coverage left out — by lean
Left
No left-leaning counter-narrative today.
Center / Wire
Surfaced: Perspectives from the Taliban or Afghan government regarding their stance on women’s organisations and funding.
Right
Surfaced: Perspectives from sailors aboard the USS Abraham Lincoln regarding their mental health and conditions.
The most-spun coverage — ranked, with what each piece leaves out, the ideology it pushes, and the loaded words
No heavily-spun coverage in today's set.