I'd seen The Big Short maybe twenty times before I finally read the book — and finding time for a book with a two-year-old in the house is near impossible. Then I spent the nights on forums, Substacks and God knows where, trying to work out whether there was something in the meticulousness of that kind of reading: a place where reading everything, the way he did, could change the game.
It wasn't in mortgage data. In private credit, I struck gold.
A corner of lending that still runs on typing. A borrower types up a certificate every month — the receivables, the agings, the payments — and the lender advances against it. Once or twice a year a field examiner comes in, reads a sample, and bills about a thousand dollars a day. The loan agreement often caps it there. Basically, the boring stuff.
Nothing's changed but the keyboard
This is not a new problem. Kautilya's Arthashastra, around 300 BC, lists forty ways an officer can embezzle from the treasury and tells the king how to check the accounts against the stores. In 1494 Pacioli printed the first description of double-entry, the Venetian method: every entry has a twin, so the book checks itself. And from that day to this, as reliably as the sun sets in the west, people have tried to get around it.
Fun fact: they didn't catch Capone for what he did. They caught his books.
A Negroni, please
I like building things that challenge the status quo — preferably in a field I have absolutely no clue about, because that forces me to read, to innovate, and it hands me an angle. I like design; there's a graveyard of a few hundred Figma projects that will never see the sun. And I like writing copy. A Negroni is my favourite drink: one part gin, one part vermouth, one part Campari, equal measures, nothing hidden. Burry is the same recipe. One part curiosity, one part design, one part writing.
The curiosity part: this work is a lot of rows, it's new to almost everyone, and it is impossible to get anyone excited about, because it's boring. That's exactly why I wanted it. I spent weeks learning the trade — how a certificate is built, what an examiner looks for, where the losses came from — and then taught a machine to read a book the way an examiner does, one row at a time, every row, and tie each one to the bank statement.
Most rows tie, and the report says so. Some don't, and they get flagged, with the row and the source document beside them. Some can't be checked either way, and the report says that too. Three answers, never a verdict.
And when you read every row instead of a sample, you find things that weren't meant to be found. Not because anyone is a criminal — because a sample is the part of the book that was prepared for a sample, and the rest wasn't prepared for anyone.
The goal
Become the new standard for how a lender's collateral is examined. Put plainly: make it damn hard to cook the books.
I ran the method on First Brands' public record first, because the docket is open and you shouldn't have to take my word for it: getburry.com/first-brands.
I'll write one of these most Fridays: the book, the cases, what it takes to read every row. If your shop has a borrower you'd like read this way, write me. Or just write.
/E