Fwd: Is the universe computable?

jbone at place.org jbone at place.org
Mon Nov 3 22:27:27 PST 2003


Hal chimes in w/ the perfect response on the thread mentioned 
previously.

--

Begin forwarded message:

Resent-From: everything-list at eskimo.com
From: "Hal Finney" <hal at finney.org>
Date: Tue Nov 4, 2003  00:23:36 US/Central
To: everything-list at eskimo.com
Subject: RE: Is the universe computable?

IMO the best idea we have discussed for why the universe is and remains
lawful is that the set of descriptions (equivalently, programs) for
the universes are governed by the Universal Distribution.  This is the
description where a string whose shortest description has length n bits
is given measure 1 / 2^n.

An heuristic argument for this distribution is that if programs are
self delimiting, then there are 2^x more programs of length n+x than
of length n, created by appending the 2^x x-bit strings to each n-bit
program.  Since the appended x bits are never executed, all 2^(n+x)
of these programs are the same as the basic 2^n programs.

A program which says "obey these simple laws" is shorter than a program
which says "obey these simple laws for a zillion steps, then start
obeying these other laws", or a program that says "obey these simple 
laws
everywhere except where this incredibly complicated configuration 
occurs,
and then do this complicated other thing."

Hal Finney



More information about the FoRK mailing list