[ExI] Simulation argument again (was Re: Computronium planet.)

[Anders Sandberg]

My problem with the QCD paper is that it assumes the Simulators use a 
particular cubical grid (they admit this is a weakness, the paper is not 
stupid by any means). The QCD guys look at systems were you calculate 
every point in the volume, rather than the very anisotropic distribution 
of stuff we have on larger scales. A galaxy simulation has most galaxies 
and matter in clumps in a small subset of the full space, and molecular 
simulations obviously has discrete molecule conglomerates.

One interesting thing I would love to see is what an octree or multipole 
code would produce in terms of anisotropies as seen from the "inside". 
Any numerical analysis hotshot who knows?

This of course relates to Moravec's caution way back in "Mind Children" 
that the universe could be running like HashLife: the mapping from our 
spacetime to the Real system implementing it could be very nontrivial.




-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University