logoalt Hacker News

gnramiresyesterday at 5:29 PM3 repliesview on HN

> to surmise that our universe is computable by a deterministic computer program

Please an expert chime in to complement, but I believe simulation in quantum mechanics is a quite interesting and not so simple topic.

I believe the more correct thing to say is that quantum mechanics is approximable than simply computable. Don't forget, QM and current physical theories use real numbers, which are not digitally representable; although it seems also the case that information is locally bounded. The finite distribution of states and probabilistic dynamics is I believe is fundamentally what makes information bounded, (which forbids for example naive hypercomputation and I think several paradoxes that might arise from unbounded local information) but don't forget in general the dynamics is still based on continuous quantities as far as we know, not discrete quantities like in a computer. QM is about discrete states within a continuum set interacting with continuum dynamics, I think is accurate to say. Quantum field theory probably complicates things more but similar conclusions might be valid. This continuum dynamics can't be computed exactly, at most you could achieve a pretty good approximation; you'd also need to provision probabilities, either from pseudorandomness (which technically would invalidate the probabilistic model, although in practice this invalidation is not feasibly distinguishable by an observer of current human-scale computing power) or gathering environmental noise.

It's really interesting to me how nature is kind of surprising and defiant of simple hypothesis we come up with. But I think each of those characteristics plays a role, is an important part of existence somehow, and we should keep trying to understand why each aspect of nature is the particular way it is for their numerous insights into our condition and existence.


Replies

wongarsuyesterday at 9:43 PM

While we model our world as continuous quantities, doesn't QM give the necessary leeway here by imposing limits on measurement? If you computed fields as grids with grid points that are half a plank length apart, then added some "quantum noise" (just good pseudorandom noise in the right distribution), could we tell?

For things like light frequencies I believe we are not even sure if they are truly unquantized, so they might be discretely computable for all we know

Time is the tricky one. But if you get away from computing the whole universe and just try to compute the viewpoint of one observer that might be quite solvable

layer8yesterday at 7:54 PM

I agree about the continuum only allowing an approximated simulation by Turing computation. But that’s orthogonal to QM, you already get that with classical fields. In Everettian QM, the evolution of the quantum state of the universe is still deterministic, following the Schrödinger equation. There is no probabilistic aspect in simulating it as a whole.

Regarding the continuum, the final word hasn’t been spoken with regard to the physical world, and probably also with regard to mathematics, since there is some level of controversy about the continuum even in the foundations of mathematics. We know for example that the continuum hypothesis is independent of ZFC, so what’s up with that. It might very well turn out that there is no continuum in physical reality. However, it seems highly unlikely that it’ll be anything like the naive conception of a cellular automaton.

legulereyesterday at 7:34 PM

Any text/program/formula describing a real number can be represented as an integer number, therefore there's only a countable amount of real numbers that can be worked with.