Vesuvius Challenge Grand Prize: AI Reads Ancient Scroll That Was Unopenable for 2,000 Years

Remember those times you tried to read a crumpled receipt from the dry cleaner? Now imagine it's 2,000 years old, carbonized, and can't be unrolled. That's exactly what the Vesuvius Challenge participants tackled.

What happened

The Vesuvius Challenge, launched in 2023 with a $1 million prize pool, aimed to read text on papyrus scrolls from Herculaneum that were carbonized during the eruption of Mount Vesuvius in 79 AD. These scrolls are so fragile that any attempt to unroll them would destroy the artifact.

The winners — students Luke Farritor, Youssef Nader, and Julian Schilliger — applied computer vision and machine learning to "see" the ink on the rolled scroll. They used high-resolution CT scans and trained a neural network to recognize the texture of the ink, invisible to the human eye.

Result: over 2,000 Greek characters read — fragments of a philosophical text about pleasure, likely by an Epicurean author. The organizers confirmed the method works and can be scaled to other scrolls.

Why it matters for IT

Technically, this is a breakthrough in computer vision and image processing. The model, trained on synthetic data, generalized features to real objects — a classic industrial problem (e.g., defect detection).

Moreover, it's a great example of scientific crowdsourcing: the problem was broken into sub-tasks, and open data allowed hundreds of teams to compete. The winners published their code — welcome to the era of collaborative AI.

Studio METABYTE comment: Pattern recognition technologies are becoming so accurate that they can read what's hidden — be it an ancient scroll or a bug in logs. If you need to automate data analysis or recover lost information, we know how to apply these approaches to your project.