Thursday, March 02, 2006

I probably shouldn't have done that...

Just talked to a local journalist, got diverted into chatting and mentioned Tegmark's stuff on "anything goes".

Journo of course picks up on this and managed to fish a quote out of me:

"The universe is like a video game, and all possible ways to play it are played".

I may come to regret having said that. Even if it is true.


Blogger Travis Garrett said...

Hi Steinn

Why would you regret it? I think the case for the many worlds interpretation is quite compelling, eternal inflation + string landscape is also quite intriguing, and I think that once one (reasonably) concludes that the universe is a mathematical structure (an explicitly realized one as opposed to an implicit one - think about the concise algorithm that generates the Mandelbrot fractal and the behavior of all the points in the complex plane under it) then one is inexorably led to the conclusion that all mathematical structures exist (i.e. there is no distinction between possible and realized abstract structures). Actually I may well disagree with Tegmark in the details of the last point - currently he is suggesting that the universe is a Godel complete structure in order to explain its apparent simplicity, while I suspect that this is a temporary state of affairs. Note also that the existence of the ensemble does not mean that *anything* goes - for instance {4 is a prime number} won't be an element of the ensemble, in fact the vast majority of random statements you could form do not correspond to any structure in the ensemble. There is natural structure to the ensemble...

Is it because some people think that these ideas "aren't scientifically testable"? I'm not actually sure that's the case (if we're a little patient...). As quantum computing improves it will become ever more obvious that arbitrarily large collections of atoms follow unitary time evolution, and as neuroscience and artificial intelligence advance it will become (even more) abundantly clear that conscious experiences are equivalent to the patterns formed by clusters of firing neurons in the brain. I also wouldn't bet against the testability of string theory landscape/inflation ideas - at least in the long term. I went to a talk earlier today that included teasing out information about inflation from primordial gravity waves like we might see with LISA (if we're lucky...) I also propose that if the mathematical ensemble exists it will have (rather dramatic) testable consequences. Namely, that any mathematical structure can be considered as a limiting case of a more complex structure, and thus for statistical reasons if the ensemble exists we will forever be finding new phenomenon that require us to expand our previous theories. Sort of a meta-prediction really, but the longer that we do continue to discover new things then the more confidence we can have in it (just like any other scientific theory...). Note also that we will likely need to significantly upgrade our mental hardware with nanotechnology in order to manipulate theories orders of magnitude more complex than string theory :-). It will be very interesting to follow computer hardware through the 2010-2020 - if the performance/watt/price ratio continues its exponential march (via increasing parallelism to a large degree) then we should have cheap pcs that match the brain's power by the beginning of the 2020s (assuming an effective information processing rate of about a kiloflop per neuron). True human level AI, and all that entails, could then reasonably be expected by 2030...

And that's that! :-)

12:28 AM  
Blogger Steinn said...

I actually think this is very interesting stuff, and it gets people going.
If you accept something like holography, or if there is a finite "bitness" to the universe, then necessarily a "large enough" universe, and it doesn't even have to be infinite, will explore all allowed states.
It is not obvious to me that this will either repeat all past states, or repeat all states with all allowed variations (the "alternate Earth reality" that got the journalist going off on the tangent) - in addition to "4 is not a prime" there may be many excluded states for simpler reasons of symmetry, or they are just not accessible given the initial conditions, even if allowed.
Clearly this is untestable now, it is not clear that it is untestable in principle, and may be testable in practise. It will be interesting to see.

BUT, the reason I think I may come to regret this is twofold: first I was commenting on a colleagues discovery having to do with space based observations of astrophysical objects. Letting the discussion diverge from supermassive black holes, to how many black holes, to can they be wormholes/what happens when you fall in, to parallel dimensions, to quantum information and universe as finite state system is RUDE and I should have pulled back. It could distract from the point of the final article. Problem was it is fun, and the journalist knew enough to ask dangerous questions.
Secondly, he's now likely to bug me again, if his editor will let him.

We'll see what the outcome is when the article comes out in ~ 10 days I expect...

6:00 PM  
Blogger Steinn said...

Well, I survived that one without any collegial damage.
only innocuous quotes made into the final version

5:06 PM  

Post a Comment

<< Home