### Wednesday, November 03, 2004

## Life goes on...

Well, I have been quite negligent of my poor blog for the last month and a half - should try not to let that happen again. All sorts of thoughts should be recorded... Start with a small one - riding back from the movie theatre on Halloween night with Katie (saw the spooky "The Grudge" appropriately enough), and early on during the drive a big grasshoper hopped onto the windshield down by the wipers. We figured it'd be blown off as we accelerated to higher speeds, but it hung on, first sheltering behind the wipers, and then just holding onto the windshield even as we went past 50 mph. It made it all the way home, where upon - of course - we poked at the intrepid insectoid castaway, which in turn naturally prompted a startling POP-jump and away it flew.

New thoughts on statistical metaphysics, which I'm thinking of changing the name to the Observer Class Hypothesis. I think it is key that we are not abstract turing machines out floating unmoored in the ensemble (so to speak), but rather physically realized computers embedded in a reality. This is because, being so embedded, we are constrained in the information we can process - we only have so much memory M, and can thus only consider ideas Z that can be encoded in M or less bits: I(Z) ≤ M (all of this will be later upgraded to be functions of time: M(t)...). Now, when first considering the ensemble it may seem like a hopeless mess - any group of structures of a certain type would be offset by another group of the opposite type, or for any 'function' ƒ(...) you will also have it's inverse, so you can't prove ƒ. But this does not hold if we order the information by size. There, if we, say, only consider all programs Z such that I(Z) ≤ N for some bit length N, then you only get some program behaviors and not others. Constraining the system brings about structure: Constraints ⇒ Structure. I.E. that's why we are physically realized computers, embedded in an immediate exterior reality from which we draw most of our information, and which we must successfully explore and manipulate so that M(t) ∼ e^(αt) ...

More soon.

New thoughts on statistical metaphysics, which I'm thinking of changing the name to the Observer Class Hypothesis. I think it is key that we are not abstract turing machines out floating unmoored in the ensemble (so to speak), but rather physically realized computers embedded in a reality. This is because, being so embedded, we are constrained in the information we can process - we only have so much memory M, and can thus only consider ideas Z that can be encoded in M or less bits: I(Z) ≤ M (all of this will be later upgraded to be functions of time: M(t)...). Now, when first considering the ensemble it may seem like a hopeless mess - any group of structures of a certain type would be offset by another group of the opposite type, or for any 'function' ƒ(...) you will also have it's inverse, so you can't prove ƒ. But this does not hold if we order the information by size. There, if we, say, only consider all programs Z such that I(Z) ≤ N for some bit length N, then you only get some program behaviors and not others. Constraining the system brings about structure: Constraints ⇒ Structure. I.E. that's why we are physically realized computers, embedded in an immediate exterior reality from which we draw most of our information, and which we must successfully explore and manipulate so that M(t) ∼ e^(αt) ...

More soon.