In April 2000, shortly after MicroProse’s flight simulation studios were shuttered by Hasbro Interactive , an anonymous developer (later identified as Kevin Klemmick) leaked the source code—specifically a version between 1.07 and 1.08—onto a public FTP site.

The package was addressed to the company's lead programmer, John. Curiosity piqued, he opened the box to find a single, sleek CD-ROM with a label that read: "Falcon 4.0 Source Code - Confidential".

, a custom-built, high-quality dataset derived from web crawling that was extensively filtered and deduplicated.