Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[wierd]

There's a problem with that premise.

The simulation could be run on an "Analog computer".
https://en.wikipedia.org/wiki/Analog_computer

The analog computer is able to process a modelled universe "inexpensively", because it forces the parent universe to perform the computations, via raw physics.  This is especially true of a universe that is not quantum. (EG, all forces and measures are infinitely sub divisible, but are still fully consistent.)

By necessity, the simulated universe will be less complex, and will have the total spectrum "smashed" by the constraints of the computer environment.  The resulting degree of quantum limitation would be imposed by the selection of components of the analog computer.  (EG, the thresholds over which the analog computer performs true-false computation)

The use of digital computers is NOT a requirement for the simulation theory.

[PTTG??]

Ok so here's a thought.

If there is a simulation tree/stack, then eventually each simulation will be identical to the one that preceded it -- that's the only stable state. Because the identical states are literally infinite, then all of the previous versions, being finite, are impossibly rare. Even if it takes a googleplex iterations to reach the point where each simulated universe is identical to its superior universe, then the infinite identical ones outnumber them infinitely.

In such a universe, there is only an infintesmal chance that we exist in a unique universe. We are certainly somewhere in that infinite stack of identical universes.

We can prove that false, however, if we ourselves cannot make an identical universe to this one. By definition, if we can't simulate this universe, the identical universe "above" us couldn't have created ours.

So, can we simulate our own universe? It doesn't look very likely.

[Reelya]

If there is a simulation tree/stack, then eventually each simulation will be identical to the one that preceded it -- that's the only stable state.

How? By what mechanism must this be true? If we run a simulation in our own universe on hardware there is no requirement that eventually the simulation will turn into a copy of our universe.

What if we model different rules of physics than our real universe, for a start? Any race running a simulation will make alterations to the simulation vs reality, because if they didn't intend to make any alterations, then they could just more easily do a real-world experiment instead of making a simulation.

EDIT: also, consider something like a very large DF dwarfputer capable of running a universe-simulation because you've allocated so much memory to it, then you simulate a complex universe inside the Dwarfputer. This shows that containing universes don't need to be more complex, the complexity of modeling can go up and down as layers rise and fall. The Dwarfputer just runs very slowly.

This suggests other things you can do: you can simulate a bigger universe by trading off speed for memory. You compress the data, so that more actually fits in. In fact, that's one end-olf-the-universe scenario I've heard. If things get to the point where energy density is too low to keep people alive, but you can run computing devices - very slowly - but on residual energy such as virtual particles / vacuum energy, then you could create a whole new universe as a simulation. The outer-universe computing would be horribly slow, but the subjective internal rate of time would appear normal. If the outer computing devices keep slowing down for infinity, then this is still sufficient to model infinite more useful living time in the inner simulation.

[PTTG??]

Any universe that can be simulated has rules that can be used to emulate a second-iteration simulation -- and a simulation that doesn't include sufficient rules to simulate a sub-universe is so simple that it's hardly a "universe."

So nesting is possible, and infinitely possible, since each universe can simulate at least a similar universe.

The reason we know it eventually becomes an infinite series of identical universes is because no other stable state is possible, with the trivial exception of a cyclical set where Universe A simulates universe B simulates universe A and so on. It might have billions of iterations but it's still an infinite identical number of universes being simulated.

Note that each subsequent iteration might be slower than the last; there's no reason the believe that the universes are simultaneous.

But again, our rules seem to rule out simulation because there is behavior that is both rule-based and too complicated to simulate efficiently, as above.

[Egan_BW]

61st from the bottom!

[wierd]

61st from the bottom!

Already posted.

snip

snip

http://www.smbc-comics.com/comic/2012-02-29

[KittyTac]

Here's the problem with generating an universe exactly like ours, but without the little parts: the little parts are required in the early stages.

[Reelya]

We're also missing the fact that any simulated universe is not required to be "less complex" by any degree, just smaller. For example, you could simulate the Earth in full detail with enough processing power. You can simulate physics as complex or even more complex than the current universe, but the resultant universe will be smaller*. So, it's meaningless to ask whether our universe could "fit" inside another universe's simulation without knowing how big the outer universe even is.

* at least, smaller in total information content, not necessarily perceived spatial size. if you had completely continuous analog matter to work with, you could in theory construct a vector-processor capable of doing infinity operations per second (since each atom effectively has an infinite number of bits of precision in it's position, momemtum and other values), thus an outer-outer universe could model infinitely large universes in a finite computing space.

[wierd]

It has to be less complex in some capacity, otherwise you run smack into complexity theory.


Either you simulate less material with a smaller total volume, or you simulate a universe with simplified physics, and fudge it so that the same higher-order outcomes happen.


Either way, you are making a less complex universe.  The only way to run a 100% parity copy of a universe, is to devote 100% of the parent universe to run it natively. (EG, you re-arrange an identical universe so that it perfectly matches, then let it go.)

This is because the native state of the universe is already its most efficient simulation of its state.  (well, assuming false vacuum and pals are not true anyway.)

[Reelya]

Not if you have matter or values within matter that are divisible in some countably-infinite way. 50% of infinity is still infinity.

EDIT: for example if you take a real-number value you can encode countably-infinite other real-number values into the same space that the one number takes. A single real-number variable contains no more data than an infinite collection of real-number variables. That's why the outcome is different if you assume that the ultimate outer universe is purely analog.

[wierd]

I brought that up when I mentioned that an analog computer could be used to perform the simulation.  However, it still has to be less complex, because then you run into the caveats and gotchas of the materials present in the parent universe.  (what you can theoretically do, and what you can actually do reliably, are not the same thing.)

[Sirus]

Hate to break up the discussion, but has anyone in this thread mentioned or noticed that the Japanese have landed some probes on an asteroid and have gotten back some pictures? Because that's pretty cool.

[Madman198237]

Hey, the Bay has been down since that happened, so, shush, you

[Sirus]

I thought that went down before the forum went poof. My bad.

[Madman198237]

Not long before, I think.