Schmidhuber's "New AI" seeks to base AI in prediction. It benefits by being based purely on the theory of computation. It need thus only be generally computable and not necessarily regular:
The Hutter Prize reflects that we cannot compress natural language by as much as we would expect:
"...in 1950, Claude Shannon estimated the entropy (compression limit)
of written English to be about 1 bit per character [3]. To date, no
compression program has achieved this level."
(http://cs.fit.edu/~mmahoney/compression/rationale.html)
It is inspired by the work of Marcus Hutter to show that compression can be used as a functional definition for intelligence.
The idea of the prize is the old one that we (can't predict, and thus