Most people do not think about file formats. We focus on the words, the contracts, the medical records, and the transactions. The format is invisible, like plumbing or electricity, until it fails. And when it fails in healthcare, legal, or financial systems, the consequences are severe.
Right now, we are doing something highly illogical: we take structured database rows, serialize them into static visual coordinates (a PDF), and then pay neural networks and large language models billions of dollars to scrape those visual coordinates back into database rows.
We are literally paying artificial intelligence to undo the failures of our own file formats [1].
There is a better way.
It is free, open source, and already working.
