4 ms·

You can't write a virtual machine with 20 GB of memory running on a real machine with only 10. Simulation isn't magic.

dwrowe 10y ago
I'm not understanding your point - which may be completely on me - but your example here - why would the "child" universe require more memory than the parent has available?
- JoeAltmaier 10y ago
  On my computer, I can write 2D simulations, or 1D. Or even 4D or 5D. The simulated universe doesn't have to have any relation to the host one.
  - fallingfrog 10y ago
    That's magical thinking. I was referring to the total simulated state space, which does not depend on the number of dimensions or any other parameter of the simulation.
    - JoeAltmaier 10y ago
      Not sure that sentence means anything. Of course the dimensionality has everything to do with the state space. It grows, by definition, geometrically with the number of dimensions.
      - fallingfrog 10y ago
        If your state space is 16 bits, then you can create a 4x4 2-dimensional space or a 2x2x2x2 4-dimensional space, but the total space is still 16 bits. You cannot simulate a 32-bit state space on a machine with only 16 bits of memory, regardless of how many dimensions the simulated space has.
        JoeAltmaier 10y ago
        Ok a precision argument I understand. But the number of states will grow geometrically. And also, we use multiple words to create larger numbers all the time in simulation. I'm not sure how important this is at all.
        fallingfrog 10y ago
        I'm not talking about 32 bits of precision; I'm talking about 32 bits of data. If you only have 16 bits then the number of unique states you can represent is 65536, and it does not matter how you divide that up into dimensions.
kyriakos 10y ago
You don't know the capacity of the host universe to start with though.
shawn-butler 10y ago
Sure you can. There is just noticeable delay and "glitches" in the simulation.
- fallingfrog 10y ago
  That doesn't even make sense, nor does it have anything to do with the argument I was making.
  - shawn-butler 10y ago
    I don't know what argument you were making and apparently neither does anyone else in this thread. Revisit your assumptions and use clarity. Also take your own advice and research memory compression and probability densities. It is very possible to write a VM that exposes 20 Gb of "memory" to an application in a host machine with only 10 Gb of physical memory by using cache which would introduce latency to the simulation. Here is an implementation for you to study: https://en.wikipedia.org/wiki/Zswap https://en.wikipedia.org/wiki/Zswap
nanny 10y ago
Maybe the host swaps to the disk? ;)
- fallingfrog 10y ago
  Right, but assuming we don't have some alternative supply of extra memory.. :)
gioele 10y ago
> You can't write a virtual machine with 20 GB of memory running on a real machine with only 10. Simulation isn't magic. You can, and it will work as long as the memory of the virtual machine can be compressed to less than 10 GB. Programs compiled to run on current amd64 CPUs have access to 2^48 bits of virtual memory, yet they run on real machines with much much less real memory. Hey, look, most of the space out there is a vacuum! Gravity may be the hashing/pre-compressor function the simulation we are living in. ;)
- fallingfrog 10y ago
  In general, no you can't. That's a fact and I encourage you to do some research if you don't understand why. It's for the same reason that there are strings of data which cannot be compressed by any form of data compression.
  - idm 10y ago
    Luckily, it does not appear that reality is uniformly distributed. Instead, we've found that quite a few phenomena can be modelled using one statistical distribution or another. In other words, reality is quite regular, which makes it unlike high-entropy strings that are difficult to compress. In fact, reality has been repeatedly shown to compress well: images, audio, and language all exhibit statistical regularities that support compression. Throughout nature, fractal structures manifest regular patterns. All of this compresses well.
oceanofsolaris 10y ago
As long as not all the memory is actually used, you can pretend that there is 20GB though. Once you actually need to map more than there is, you would need to stop the virtual machine. If you read e.g. the Bostrom 2003 paper, you will see that this is more or less what they claim: If your simulation technique is good enough, you can just fake most of the universe and only actually perform fine-grained simulations where it matters. This would e.g. assume that you have special simulation rules for brains, large scale chemical reactions etc., which allows you to just skip most of the details and make them up once you need them. You could even automatically detect whether your simulation is fine grained enough by running the same simulation at two different granularities at the same time, readjusting whenever the outcome between the two differs too much. I think the argument in its original form does not make too much sense (future humans simulating todays humans for research and fun). What you would probably simulate much more, are human-like entities in a much simpler universe (e.g. no weird quantum-mechanics that are really hard to simulate). So in my opinion, you could take away from the simulation argument that there is a non-zero chance that we are living in a simulation. But the computer that performs the simulation most probably lives in a universe that does not follow the same rules our universe does. It might be a universe with two or more time-like dimensions. It might be one where quantum mechanics are just super-easy to simulate for some reason (maybe in the same way it would be really easy to e.g. build Conways game of life in transistor hardware for us).
greendesk 10y ago
I have another counter-example. Why not simulate a smaller part of the universe, not the whole experience? Suppose you have a real machine with 10 GB. You simulate a world of 1 GB. Inside the world, the universe is contained in 1 GB (say, a smaller version of the universe, simplified physics, etc). Inside the virtual machine, in the simulation world, the perceived universe is 1 GB. They simulation agents don't perceive a short-coming because they have never been exposed to the out-the-simulation experiences.
- fallingfrog 10y ago
  Right. But if you were randomly spawned somewhere, you'd be 10 times more likely to be in the 10GB than the 1GB.
  - majewsky 10y ago
    Except we are anything but randomly spawned, on the scale of the universe.
- chriswarbo 10y ago
  > Inside the world, the universe is contained in 1 GB (say, a smaller version of the universe, simplified physics, etc). > They simulation agents don't perceive a short-coming because they have never been exposed to the out-the-simulation experiences. I agree, and this was the only part I found lacking in the article. For example, where it says: > The argument is quite simple: we already make simulations, and with better technology it should be possible to create the ultimate one, with conscious agents that experience it as totally lifelike. This is implying that the world we experience is somehow the "ultimate" one, and it would require super-intelligence to bring a mere simulation up to this pinnacle of existence. In fact, that's also an implicit assumption being made by those searching for 'errors' (not to dismiss such efforts; chasing empirical evidence is the best contribution to make in such arguments). Yet if our universe is a simulation, there's nothing to say it's not actually a crappy, low-res, toy-physics sandbox; the simulators may consider it to be as "ultimate" and "lifelike" as we find 80s home computer games.
jeremiep 10y ago
Why is it so hard to imagine universes bigger than our own? That's like saying "I only own a computer with 10Gb of memory, so certainly a program using 20Gb is impossible!"
- fallingfrog 10y ago
  Go back and read what I wrote again.
kamaal 10y ago
What makes you think beings simulating us are in a universe with resources lesser than us?

I'm not understanding your point - which may be completely on me - but your example here - why would the "child" universe require more memory than the parent has available?

JoeAltmaier 10y ago
On my computer, I can write 2D simulations, or 1D. Or even 4D or 5D. The simulated universe doesn't have to have any relation to the host one.
- fallingfrog 10y ago
  That's magical thinking. I was referring to the total simulated state space, which does not depend on the number of dimensions or any other parameter of the simulation.
  - JoeAltmaier 10y ago
    Not sure that sentence means anything. Of course the dimensionality has everything to do with the state space. It grows, by definition, geometrically with the number of dimensions.
    - fallingfrog 10y ago
      If your state space is 16 bits, then you can create a 4x4 2-dimensional space or a 2x2x2x2 4-dimensional space, but the total space is still 16 bits. You cannot simulate a 32-bit state space on a machine with only 16 bits of memory, regardless of how many dimensions the simulated space has.
      - JoeAltmaier 10y ago
        Ok a precision argument I understand. But the number of states will grow geometrically. And also, we use multiple words to create larger numbers all the time in simulation. I'm not sure how important this is at all.
        fallingfrog 10y ago
        I'm not talking about 32 bits of precision; I'm talking about 32 bits of data. If you only have 16 bits then the number of unique states you can represent is 65536, and it does not matter how you divide that up into dimensions.

kyriakos 10y ago

You don't know the capacity of the host universe to start with though.

shawn-butler 10y ago

Sure you can. There is just noticeable delay and "glitches" in the simulation.

fallingfrog 10y ago
That doesn't even make sense, nor does it have anything to do with the argument I was making.
- shawn-butler 10y ago
  I don't know what argument you were making and apparently neither does anyone else in this thread. Revisit your assumptions and use clarity. Also take your own advice and research memory compression and probability densities. It is very possible to write a VM that exposes 20 Gb of "memory" to an application in a host machine with only 10 Gb of physical memory by using cache which would introduce latency to the simulation. Here is an implementation for you to study: https://en.wikipedia.org/wiki/Zswap https://en.wikipedia.org/wiki/Zswap

nanny 10y ago

Maybe the host swaps to the disk? ;)

fallingfrog 10y ago
Right, but assuming we don't have some alternative supply of extra memory.. :)

gioele 10y ago

> You can't write a virtual machine with 20 GB of memory running on a real machine with only 10. Simulation isn't magic. You can, and it will work as long as the memory of the virtual machine can be compressed to less than 10 GB. Programs compiled to run on current amd64 CPUs have access to 2^48 bits of virtual memory, yet they run on real machines with much much less real memory. Hey, look, most of the space out there is a vacuum! Gravity may be the hashing/pre-compressor function the simulation we are living in. ;)

fallingfrog 10y ago
In general, no you can't. That's a fact and I encourage you to do some research if you don't understand why. It's for the same reason that there are strings of data which cannot be compressed by any form of data compression.
- idm 10y ago
  Luckily, it does not appear that reality is uniformly distributed. Instead, we've found that quite a few phenomena can be modelled using one statistical distribution or another. In other words, reality is quite regular, which makes it unlike high-entropy strings that are difficult to compress. In fact, reality has been repeatedly shown to compress well: images, audio, and language all exhibit statistical regularities that support compression. Throughout nature, fractal structures manifest regular patterns. All of this compresses well.

oceanofsolaris 10y ago

As long as not all the memory is actually used, you can pretend that there is 20GB though. Once you actually need to map more than there is, you would need to stop the virtual machine. If you read e.g. the Bostrom 2003 paper, you will see that this is more or less what they claim: If your simulation technique is good enough, you can just fake most of the universe and only actually perform fine-grained simulations where it matters. This would e.g. assume that you have special simulation rules for brains, large scale chemical reactions etc., which allows you to just skip most of the details and make them up once you need them. You could even automatically detect whether your simulation is fine grained enough by running the same simulation at two different granularities at the same time, readjusting whenever the outcome between the two differs too much. I think the argument in its original form does not make too much sense (future humans simulating todays humans for research and fun). What you would probably simulate much more, are human-like entities in a much simpler universe (e.g. no weird quantum-mechanics that are really hard to simulate). So in my opinion, you could take away from the simulation argument that there is a non-zero chance that we are living in a simulation. But the computer that performs the simulation most probably lives in a universe that does not follow the same rules our universe does. It might be a universe with two or more time-like dimensions. It might be one where quantum mechanics are just super-easy to simulate for some reason (maybe in the same way it would be really easy to e.g. build Conways game of life in transistor hardware for us).

greendesk 10y ago

I have another counter-example. Why not simulate a smaller part of the universe, not the whole experience? Suppose you have a real machine with 10 GB. You simulate a world of 1 GB. Inside the world, the universe is contained in 1 GB (say, a smaller version of the universe, simplified physics, etc). Inside the virtual machine, in the simulation world, the perceived universe is 1 GB. They simulation agents don't perceive a short-coming because they have never been exposed to the out-the-simulation experiences.

fallingfrog 10y ago
Right. But if you were randomly spawned somewhere, you'd be 10 times more likely to be in the 10GB than the 1GB.
- majewsky 10y ago
  Except we are anything but randomly spawned, on the scale of the universe.
chriswarbo 10y ago
> Inside the world, the universe is contained in 1 GB (say, a smaller version of the universe, simplified physics, etc). > They simulation agents don't perceive a short-coming because they have never been exposed to the out-the-simulation experiences. I agree, and this was the only part I found lacking in the article. For example, where it says: > The argument is quite simple: we already make simulations, and with better technology it should be possible to create the ultimate one, with conscious agents that experience it as totally lifelike. This is implying that the world we experience is somehow the "ultimate" one, and it would require super-intelligence to bring a mere simulation up to this pinnacle of existence. In fact, that's also an implicit assumption being made by those searching for 'errors' (not to dismiss such efforts; chasing empirical evidence is the best contribution to make in such arguments). Yet if our universe is a simulation, there's nothing to say it's not actually a crappy, low-res, toy-physics sandbox; the simulators may consider it to be as "ultimate" and "lifelike" as we find 80s home computer games.

jeremiep 10y ago

Why is it so hard to imagine universes bigger than our own? That's like saying "I only own a computer with 10Gb of memory, so certainly a program using 20Gb is impossible!"

fallingfrog 10y ago
Go back and read what I wrote again.

kamaal 10y ago

What makes you think beings simulating us are in a universe with resources lesser than us?