The Dahling Index
Which public claims hold up?
Every verdict on this site, counted. The figures update as essays are published and corrected.
of the claims checked in 2026 don't hold.
of all 118 verdicts found the claim holds up.
of popular claims don't hold, the highest share of any kind of claimant with five or more checked claims.
118 essays · 118 claimants named · data as of 1 November 2026
The 2026 Q4 report
Whose claims held up across 118 sourced essays?
This index covers 118 essays, 107 of them published in 2026. Each essay tests a single claim, and the editor confirmed whether the claimant's own claim held. Of the 118 claims, 44 held, 20 partly held, 53 did not hold, and 1 remains unproven.
- 45%
53 of 118 claims did not hold, which is more than the 44 that did.
did silicon valley banks 2022 capital ratios count its bond losses · did memberships generate all costcos operating profit in 2024 · was macys 151 million correction really a one employee problem · oura sleep tracking lawsuit accuracy claims · what evidence did donotpay have that its ai could replace a lawyer
- 9 of 13
Claims made by agencies and regulators held in 9 of 13 essays, and only 1 did not hold.
did carl icahn pledge most of icahn enterprises without telling investors · did avast sell the browsing histories it promised to protect · did adobe make annual subscriptions look like monthly commitments · why did apple lose its 13 billion irish tax case · can a bank outsource its technology without outsourcing responsibility for an outage
- 11 of 14
In Beauty, 11 of 14 claims did not hold, the highest share of any subject with more than two essays.
why american sunscreen is stuck fda monograph · do ai skin analyzers actually diagnose anything · menopause skin collagen what the evidence supports · who decides whether body oil is better than lotion · flawless base makeup sensitive reactive skin
- 45 of 80
Popular claims did not hold in 45 of 80 essays, or 56 percent.
does the genius act guarantee that stablecoin holders will be repaid in · first do no harm hippocratic oath · the real ai threat is persuasion not code · what wellness culture gets dangerously wrong · expired cards still work zombie card contactless
Written from 118 essays, as of 25 October 2026.
Verdicts by subject
Where claims hold up
Verdicts by who made the claim
Whose claims hold up
Over time
Share that doesn't hold, by quarter
Quarters with fewer than 5 verdicts are drawn dashed.
Claimants
Who made the claims
- Popular claimOther · 80 claims
- FTCAgency or regulator · 4 claims
- SECAgency or regulator · 2 claims
- OuraCompany · 2 claims
- London Metal ExchangeCompany · 1 claim
- Fifth CircuitAgency or regulator · 1 claim
- Generative AI at Work study authorsResearcher or study · 1 claim
- CongressElected official · 1 claim
- EYCompany · 1 claim
- FTC and Justice DepartmentAgency or regulator · 1 claim
- Warren BuffettOther · 1 claim
- BlockCompany · 1 claim
- European CommissionAgency or regulator · 1 claim
- PalantirCompany · 1 claim
- AmazonCompany · 1 claim
- IntelCompany · 1 claim
- Macy'sCompany · 1 claim
- StarbucksCompany · 1 claim
- WorkdayCompany · 1 claim
- AppleCompany · 1 claim
- RealPageCompany · 1 claim
- DoNotPayCompany · 1 claim
- TeslaCompany · 1 claim
- MetaCompany · 1 claim
How this is counted
Each published essay tests one public claim against the primary record and gives one of four verdicts: holds up, partly holds, doesn't hold, or unproven. The Index counts those verdicts. It counts essays, not claims in general. Alicia chooses which claims to check, so these figures describe the claims she has checked, not every claim made in public. Subjects are taken from the essay itself. Claimant and claimant type come from each essay's attribution, and the editor checks them before the Index goes out. Quarters follow publication date. Corrections change the counts the next time this page loads.