In an absolute win for the tech sector and classical culture enthusiasts everywhere, a bunch of private software developers and Silicon Valley donors have used AI to virtually unwrap a 2,000-year-old papyrus scroll that was burnt to a crisp by Mount Vesuvius in AD79. The scroll, designated PHerc 1667, was recovered from an ultra-wealthy Roman villa in Herculaneum. Instead of waiting around for tax-funded academic committees to draft another useless proposal, a global competition called the Vesuvius Challenge bypassed the bureaucracy, threw a massive bag of cash at the problem, and cracked the code using machine learning. This is what happens when you let developers cook instead of leaving things to slow-moving institutions.

For centuries, historical preservation was dominated by clumsy academics who did more damage to these ancient artifacts than the volcanic eruption itself. In their infinite wisdom, past researchers trying to unroll the carbonized papyrus physically managed to break the scroll in half and cause the outer layers to flake off and disintegrate. What remains of PHerc 1667 is a pathetic 8 centimeters tall and 2 centimeters wide—half of its original size. It took modern computer science to fix the mess left behind by generations of physical 'preservationists' who couldn't handle the delicate material without turning it into a pile of ash.

The scroll itself was recovered from the library of a luxury Roman villa in Herculaneum, near Naples, that was blasted by intense heat and buried under a mountain of ash in AD79. This volcanic eruption completely destroyed nearby Pompeii, but ironically preserved this massive library of carbonized scrolls. However, because they were burnt to a crisp, they were essentially unreadable time capsules. For decades, the academic consensus was basically that we would never know what was written inside them without destroying them in the process.

Enter Professor Brent Seales from the University of Kentucky, who realized that instead of physically tearing the scroll apart, they could just scan it. Seales trained machine-learning algorithms to spot microscopic differences in the papyrus fibers on high-resolution X-ray images, detecting where the ancient ink was applied without ever touching the delicate carbonized lump. This virtual unwrapping technique is a literal game-changer, proving that code can solve the physical limitations of archaeological science.

Armed with this technology, the Vesuvius Challenge—backed by private Silicon Valley money—dished out hundreds of thousands of dollars in prizes to anyone who could write code to digitally unwrap and read the scrolls. This global contest, launched in 2023, completely decentralized the research. Instead of gatekeeping the scrolls in elite university basements, they put high-resolution X-ray images online and let global coding teams go to town on them. The strategy worked flawlessly, producing rapid breakthroughs that traditional academic models could only dream of.