Comparisons
Redacted vs. Unredacted: What Flight Log Quality Costs
TLDR Vision-language model processing of unredacted flight logs extracted 4,286 flights with 392 unique passenger names at zero errors. The same model on...
#OCR
11 investigations
Technical Deep Dives
The Many Rodgers: How Entity Resolution Handles Scanning Chaos
TLDR A software tool that uses statistics to decide which database records refer to the same person (called Splink) merged eight or more scanning variants of...
Data Quality
The Singleton Crisis: 197,945 Entities That Appear Once
TLDR Nearly one in four entities in the corpus (197,945 out of 821,633) appears in only one document. An entity that appears in only one document — called a...
Corpus & Data
The 16-Script Pipeline
TLDR A pipeline of 27+ Python scripts transforms 2.1 million raw government documents into a searchable PostgreSQL database with 2.38 million extracted...
Notable FedEx Shipments
Dr. Bruce Moskowitz — An OCR Puzzle
TLDR A FedEx third-party shipment lists the recipient as "DR BRUCE LOSKOWTZ" in West Palm Beach. Character substitution analysis suggests...
Observations
OBS-5/6: When Scanning Software Hallucinates Countries From Blank Forms
TLDR Two observations were retracted after visual inspection revealed that scanning software (OCR, or optical character recognition — software that reads text...
Machine Intelligence
OCR Hallucinations
TLDR OCR engines (software that converts images of text into searchable characters) produce phantom entities from blank form labels and repeated document...
Physical Logistics
2,894 Packages: What FedEx Invoices Reveal About Epstein's Logistics
TLDR Custom-built parsing scripts extracted 2,894 FedEx shipments from DOJ-released invoices, spanning July 2000 through October 2005 (PAPER TRAIL Project,...
Physical Logistics
148 Third-Party Shipments: Who Had Access to Epstein's FedEx Account?
TLDR Of 148 FedEx shipments marked "Third Party," 94% still billed to Epstein's personal account -- meaning the designation identifies who initiated...
Network & Temporal
4,286 Flights: How a Vision Language Model Read Handwritten Flight Logs
TLDR Traditional text extraction failed completely on Epstein's handwritten flight logs. A vision language model -- a type of AI that interprets entire page...
Methodology
External Corroboration: How Six Agents Verified the Corpus
TLDR Six parallel research agents independently verified corpus findings against public registries, FAA records, and court filings (PAPER TRAIL Project,...