Compressing Data to Create Intelligence

A cash prize rewards people for compressing a massive text file, based on the theory that perfecting data compression is identical to achieving human-level artificial intelligence.

This challenge assumes that to compress a file perfectly, a computer must fully understand the logic, patterns, and meaning of the language within it. By shrinking text as much as possible, machines are forced to learn how to predict information just like a human brain does. The contest uses a one-gigabyte snapshot of Wikipedia to test these boundaries. It turns out that mastering the art of file size might be the secret path to building a truly sentient AI.