Get the latest tech news
Primitive Kolmogorov complexity is computable
Defining a computable analogue of Kolmogorov complexity and Solomonoff induction over primitive recursive functions that is more practically relevant to intelligence.
Solomonoff induction uses this principle to assign a prior probability to a hypothesis based on its Kolmogorov complexity, creating a universal predictor. For most practical applications in science and engineering, the functions required to model systems and make predictions are primitive recursive. The physical world does not appear to demand that intelligent agents compute non-primitive recursive functions to survive and operate effectively.
Or read this on Hacker News