Note that all models are physical because all modelers are finite and all modeling is process limited to available resources. I think true/false 0/1 is asymptotic. 0 and 1 is as 'reachable' as the speed of light and the absolute zero or the planck or hubble scales. The relation thing-model is always of 'resemblance'. Euclidean geometry is 'like' the world, and it's 'laws' are neat form of compression, outputting good enough and cheap results to be used in practice. To 'model everything' ofc the modeler must BE at least as big (in Shannon terms) as the modeled. As big in complexity terms indeed, which implies also resources in the light of stronger algo = doing more with less. This gives some hope a part of the universe to grok it, as long as the algo which outputs universe is no longer than its full bitstring.
You are viewing a single comment's thread from: