Re: Kolmogorov complexity and prediction

From: Nick Hay (
Date: Thu Oct 18 2007 - 19:00:30 MDT

On 10/18/07, Bryan Bishop <> wrote:
> On Thursday 18 October 2007 11:30, Nick Hay wrote:
> > So systems of fixed complexity can both understand, and build
> > understandable, systems of greater complexity than themselves.
> Interesting implications. How would this be applied to say, the
> universe? The universe is a complex system, and I always thought that
> no system could completely understand itself without violating some
> gödeleans (incompleteness or inconsistency).
> When you first said "bridge," in a previous email, I thought you meant a
> sort of theoretics-bridge to take a program of some complexity and then
> have it breach past some finite complexity-production limit. Apparently
> not. :)

Yes, I meant it literally :)

> It is also important to point out that no system can generate infinitely
> more complex systems, implying that even if the universe is infinite
> (do we argue this here? I hope not), that the accessibility of such
> vastness is limited to our finite programs/designs/specifications.

Infinitely more complex would mean uncomputable, in this context. It
may be that a computable system cannot create an uncomputable system.
Or, rather, if IS true if you define create one way, and it's false if
you don't.

The universe may be simple as a whole. See if you haven't.

-- Nick

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:58 MDT