4 ms·

We don't live in a computer simulation. Here's why: No matter how efficient your computer is, the total state space of all the simulations in the universe is l

We don't live in a computer simulation. Here's why: No matter how efficient your computer is, the total state space of all the simulations in the universe is less than the total state space of the containing universe. In fact even if you dedicate the total energy output of the universe to running simulations, the total simulated state space is much much less than the state space of the containing universe. And that applies recursively- all simulations contained in the simulations can't possibly add up to more than a tiny portion of the volume of the containing universe. So, you're incredibly more likely to be in a real universe.

CuriouslyC 10y ago
Your argument only knocks down the idea that the universe is simulations all the way down, assuming a finite prior probability distribution on the size of the universe (which I would be inclined to agree with). Even if each simulation was three orders of magnitude less efficient than its parent universe, that would still leave a lot of room for simulation.
- fallingfrog 10y ago
  Right, but what I'm saying is that the total "room for simulation" is at most as large as the room in the parent universe, and probably much smaller. So if you were randomly spawned somewhere in the total space made up of all the real and simulated universes, the chances are very good you'd be in a real universe, not a simulated one.
  - CuriouslyC 10y ago
    That would be true, assuming a uniform prior probability of complex life in the real universe and a simulation. In actuality the simulation would almost certainly have rules that encourage the aggregation and stability of information, such that the probability of complex life would be much, much higher.
    - fallingfrog 10y ago
      If this universe is designed to be efficient in terms of the ratio of living matter to inert matter, I'd hate to see what the parent universe looks like. Also, the matter in the parent universe running the simulation has got to be active matter using energy somehow, so why not construct native life forms? The simulator itself is a device for maintaining information, if you want to create intelligent life, (or any other life), it's always more efficient to just create an AI than a virtual universe with an AI in it.
      - CuriouslyC 10y ago
        Why not create life forms? Why not create an AI? Why create a simulation at all? In my opinion, the best argument for why is entertainment. The originator of a simulation of the universe this nuanced probably has existence on full autopilot, what is left but recreation? You could think of the universe like minecraft (the best selling PC game of all time, btw) raised to some ridiculous power.
dwrowe 10y ago
I've never understood this argument in the context of this discussion. Why would the constraints of the simulation apply to that which is running the simulation?
- michaelfeathers 10y ago
  Exactly.
  - fallingfrog 10y ago
    OK, there are lots of people here insisting that you can simulate a larger machine inside of a smaller machine, in terms of the total state space. Let me assure you that in general, you cannot. That's absolutely a fact. Windows can fool you into thinking it can simulate a 20GB system inside a 10GB system, because it's using the hard drive as the extra 10GB. Which means you really still have a 20GB+ simulating system! I'm talking about a system in which you don't have some other form of memory to fall back on. Data cannot be infinitely compressed, and state space is finite no matter how it's partitioned into dimensions and so forth.
    - JoeAltmaier 10y ago
      Still simulation means you may not have to represent every particle/state but can instead estimate. Or even not represent them at all until they are observed. Like a graphical system renders only what's in front of the camera. So the effective universe size could be enormous and use fewer resources than is apparent.
      - fallingfrog 10y ago
        Actually, the opposite is true. Read this great explanation by Richard Feynman: https://people.eecs.berkeley.edu/~christos/classics/Feynman.pdf https://people.eecs.berkeley.edu/~christos/classics/Feynman....
- fallingfrog 10y ago
dwrowe 10y ago
I've never understood this argument in the context of this discussion. Why would the constraints of the simulation apply to that which is running the simulation?
xyzzy4 10y ago
If the laws of physics are different in the universe that's running the simulation, then all bets are off.
- fallingfrog 10y ago
  Wrong.
Phemist 10y ago
This is the German Tank problem applied to simulated universes. What is the maximum likelihood state space of the simulating universe, given that the simulated universe has a state space of X?
- fallingfrog 10y ago
  Yes! And the state space of the simulating universe must be at least as large as the state space of all the simulated universes in it.
- fallingfrog 10y ago
  So far you're the only person who really got my argument at all. What was I unclear about? Maybe I assumed too much knowledge of math and physics.
  - Phemist 10y ago
    What I can distill is that a simulating universe may be computationally stronger than the simulated universe, and not computationally equivalent. However, given Church-Turing theorem (ie, the simulatung universe does not need to be in a higher computational class to simulate equivalent universes), you can make a similar argument. What is the likely computational class of the simulating universe, given a class X of the simulated universe
zielmicha 10y ago
Simulated universe might have more live-density per state space bit then host universe. I.e. most power of host universe might be dedicated to running "stupid" processes (using, for example, 10^200 bits per sentient being), while simulated universe might be better optimized (using only e.g. 10^40 bits per sentient being). So even if simulated universe is much smaller, it still will be more likely to "be born" in it. (This argument is similar to other arguments against your claim, but rephrased)

CuriouslyC 10y ago

Your argument only knocks down the idea that the universe is simulations all the way down, assuming a finite prior probability distribution on the size of the universe (which I would be inclined to agree with). Even if each simulation was three orders of magnitude less efficient than its parent universe, that would still leave a lot of room for simulation.

fallingfrog 10y ago
Right, but what I'm saying is that the total "room for simulation" is at most as large as the room in the parent universe, and probably much smaller. So if you were randomly spawned somewhere in the total space made up of all the real and simulated universes, the chances are very good you'd be in a real universe, not a simulated one.
- CuriouslyC 10y ago
  That would be true, assuming a uniform prior probability of complex life in the real universe and a simulation. In actuality the simulation would almost certainly have rules that encourage the aggregation and stability of information, such that the probability of complex life would be much, much higher.
  - fallingfrog 10y ago
    If this universe is designed to be efficient in terms of the ratio of living matter to inert matter, I'd hate to see what the parent universe looks like. Also, the matter in the parent universe running the simulation has got to be active matter using energy somehow, so why not construct native life forms? The simulator itself is a device for maintaining information, if you want to create intelligent life, (or any other life), it's always more efficient to just create an AI than a virtual universe with an AI in it.
    - CuriouslyC 10y ago
      Why not create life forms? Why not create an AI? Why create a simulation at all? In my opinion, the best argument for why is entertainment. The originator of a simulation of the universe this nuanced probably has existence on full autopilot, what is left but recreation? You could think of the universe like minecraft (the best selling PC game of all time, btw) raised to some ridiculous power.

dwrowe 10y ago

I've never understood this argument in the context of this discussion. Why would the constraints of the simulation apply to that which is running the simulation?

michaelfeathers 10y ago
Exactly.
- fallingfrog 10y ago
  OK, there are lots of people here insisting that you can simulate a larger machine inside of a smaller machine, in terms of the total state space. Let me assure you that in general, you cannot. That's absolutely a fact. Windows can fool you into thinking it can simulate a 20GB system inside a 10GB system, because it's using the hard drive as the extra 10GB. Which means you really still have a 20GB+ simulating system! I'm talking about a system in which you don't have some other form of memory to fall back on. Data cannot be infinitely compressed, and state space is finite no matter how it's partitioned into dimensions and so forth.
  - JoeAltmaier 10y ago
    Still simulation means you may not have to represent every particle/state but can instead estimate. Or even not represent them at all until they are observed. Like a graphical system renders only what's in front of the camera. So the effective universe size could be enormous and use fewer resources than is apparent.
    - fallingfrog 10y ago
      Actually, the opposite is true. Read this great explanation by Richard Feynman: https://people.eecs.berkeley.edu/~christos/classics/Feynman.pdf https://people.eecs.berkeley.edu/~christos/classics/Feynman....
fallingfrog 10y ago

dwrowe 10y ago

I've never understood this argument in the context of this discussion. Why would the constraints of the simulation apply to that which is running the simulation?

xyzzy4 10y ago

If the laws of physics are different in the universe that's running the simulation, then all bets are off.

fallingfrog 10y ago
Wrong.

Phemist 10y ago

This is the German Tank problem applied to simulated universes. What is the maximum likelihood state space of the simulating universe, given that the simulated universe has a state space of X?

fallingfrog 10y ago
Yes! And the state space of the simulating universe must be at least as large as the state space of all the simulated universes in it.
fallingfrog 10y ago
So far you're the only person who really got my argument at all. What was I unclear about? Maybe I assumed too much knowledge of math and physics.
- Phemist 10y ago
  What I can distill is that a simulating universe may be computationally stronger than the simulated universe, and not computationally equivalent. However, given Church-Turing theorem (ie, the simulatung universe does not need to be in a higher computational class to simulate equivalent universes), you can make a similar argument. What is the likely computational class of the simulating universe, given a class X of the simulated universe

zielmicha 10y ago

Simulated universe might have more live-density per state space bit then host universe. I.e. most power of host universe might be dedicated to running "stupid" processes (using, for example, 10^200 bits per sentient being), while simulated universe might be better optimized (using only e.g. 10^40 bits per sentient being). So even if simulated universe is much smaller, it still will be more likely to "be born" in it. (This argument is similar to other arguments against your claim, but rephrased)