Post
by Speaker to Animals » Tue Feb 28, 2017 6:34 pm
There exist some very good reasons to claim it's not at all a simulation. For one, if it were a simulation, you should be able to detect that fact. You have to look at it from a theoretical CS point of view.
For instance, we can't really simulate our own universe. A simulating universe has to be quite a lot larger than the simulated universe in order to host the computer that runs the simulation, for example. Inside the simulation, there will be tell-tale signs that everything is computational. Granted, we simply might be unable to detect this yet, but then again, some of the nondeterminism inherent in our universe indicates that it's not the case.
For instance, some of the problems that would have to be solved in order to run a universe simulation are among the hardest classes of problems, such as NP-Complete and harder. Even the problem of computing orbital mechanics of every object in a star system is PSPACE complexity. But consider the n-body problem has to be solved for every star system and every galaxy and every galactic cluster and every group and so on. All of these things would be computed simultaneously. The actual universe in which we would live, the one containing our simulation, would have to be orders of magnitude larger than this one, and most of it converted to a computer.
Evidence in support of it would be the little details like how subatomic particles can only be pinned down to an orbit after they are observed. Spin of electrons can only be determined after observation. That might belie a kind of design pattern we use in our own software whereby we delay calculation until the calculation is actually needed. If I were to design a very large simulation, I would use a principle like this.
__________ 
