User Controls
The Simulation Hypothesis
-
2017-10-04 at 3:03 AM UTC
-
2017-10-04 at 5:22 AM UTC
Originally posted by Falco Let me start with the simplest one:
I'm sure you are familiar with the n-body problem in mathematics (and physics). Even if you assume that interactions between n>2 bodies can be regularized and apply Sundman's globalised solution to generate every "frame" of reality as a simulation and are not doing it in real time, by the evolution of even a small system (less than 5 physical elements), eventually the simulated elements in the system will quickly grow to larger than the number of atoms in the known universe, and there is no way to represent them. It's basically mathematically impossible to simulate.
Take the example of a quadruple pendulum for example
Simulations such as this one are purposely designed to cull many, many elements of the actual system as it evolves. In the case of the quadruple pendulum, without saying "hey, we just aren't going to include any forces except these", the system would evolve at the rate of n^x^z if n is the initial number of forces being simulated, X is the number of frames into the simulation, and z=(x+1)^((x+1)^x) until it would quickly approaching an infinite amount of time to generate the next frame of the simulation.
It seems like computational issues with solving the n-body problem only applies to relativistic models (we posited that future societies would live in a universe physically similar to our own, not that they would have the exact same physical models as we do). Just as n-bodies was analytically solvable under a Newtonian model, I can't see any reason to deny an analytic solution under any possible unified field theory which we know relativity is not.
Also we have relatively efficient approximation techniques of n-bodies problems under relativistic assumptions, again, there's no requirement that if our reality is simulated it be a perfect simulation total physical simulation.
And just to throw in the cartesian doubt angle, how realistic do you need to make a physical simulation to simulate your consciousness? When was the last time you empirically verified physical models hold? Have you done this enough and with enough scrutiny that simulating all the local cases necessary to give your consciousness the impression of a consistent reality is computationally infeasible? I don't know about you but I can't even see the pixels in my screen anymore, the level a physical simulation needs to be realistic at to convince me is maybe a hundredth of an inch at most. It's not that uncommon to put one's shirt on backwards, it's not like the contents of your experience need to be that detailed.The following users say it would be alright if the author of this post didn't die in a fire! -
2017-10-04 at 1:28 PM UTC
Originally posted by Lanny It seems like computational issues with solving the n-body problem only applies to relativistic models (we posited that future societies would live in a universe physically similar to our own, not that they would have the exact same physical models as we do). Just as n-bodies was analytically solvable under a Newtonian model, I can't see any reason to deny an analytic solution under any possible unified field theory which we know relativity is not.
Yes, because the Newtonian models were close, but wrong. In practice, an n body simulation using a Newtonian model would actually be wildly incorrect experimentally. It's not solvable in Newtonian models vs unsolvable in relativistic models due to any fundamental problem with relativity as a theory. Newtonian models simply treat bodies as if they exist in a static space, no tricksy spacetime etc. Each element in a relativistic model, however, is an additional element that fundamentally changes the nature of the playing field every step in a simulation, and exponentially adds complexity to it.
A GUT that somehow removes the exponentiating nature of the problem is improbable enough to the point where I don't think it's really worth discussing.Also we have relatively efficient approximation techniques of n-bodies problems under relativistic assumptions, again, there's no requirement that if our reality is simulated it be a perfect simulation total physical simulation.
And just to throw in the cartesian doubt angle, how realistic do you need to make a physical simulation to simulate your consciousness? When was the last time you empirically verified physical models hold? Have you done this enough and with enough scrutiny that simulating all the local cases necessary to give your consciousness the impression of a consistent reality is computationally infeasible? I don't know about you but I can't even see the pixels in my screen anymore, the level a physical simulation needs to be realistic at to convince me is maybe a hundredth of an inch at most. It's not that uncommon to put one's shirt on backwards, it's not like the contents of your experience need to be that detailed.
Let me put it another way; if you pick out 189 stars close to one another, and simulate their a gravitational interactions with 3 forces attributed to each body (hideously inaccurate) in a relativistic model (or any other model that doesn't simplify the special cases of any model down to something like newtonian mechanics in special cases), and each force interaction is somehow represented on one bit, and each bit is one atom large, you would run out of atoms in the universe to represent their interaction...
...
After one frame of your simulation.
So you look up at the sky, if shit isn't majorly fucking up and wildly deviating from our existing relativistic models, you can be pretty sure that it's not running on a computer, but somehow the "thing in itself" rather than a siulation of the "thing in itself", to stab at the philosophical side. Although that's a difficult distinction to make.
I mean we can't 100% rule out a simulation theory but there is no reason to believe it is possible.
Post last edited by Captain Falcon at 2017-10-04T16:09:17.381645+00:00 -
2017-10-04 at 2:34 PM UTCMost of the theories I have been looking at aren't suggesting that reality is literally running on a computer but more that it is simulation-like, or dream-like, that these are only metaphors for what is actually going on.
-
2017-10-04 at 4:07 PM UTC
-
2017-10-04 at 4:17 PM UTCHere's my theory. Life is a game of who can acquire the most. Therefore, he who dies with the most toys wins.
-
2017-10-04 at 4:18 PM UTC
-
2017-10-04 at 5:55 PM UTC
Originally posted by Open Your Mind 1. a figure of speech in which a word or phrase is applied to an object or action to which it is not literally applicable.
2. a thing regarded as representative or symbolic of something else, especially something abstract.
No I was asking what the fuck "that" means.
Also, what idea of higher reality do you have, which contains this "dream like" or "simulation like" world of ours? The noumenal world? -
2017-10-05 at 12:57 AM UTC
Originally posted by Captain Falcon Yes, because the Newtonian models were close, but wrong. In practice, an n body simulation using a Newtonian model would actually be wildly incorrect experimentally. It's not solvable in Newtonian models vs unsolvable in relativistic models due to any fundamental problem with relativity as a theory. Newtonian models simply treat bodies as if they exist in a static space, no tricksy spacetime etc. Each element in a relativistic model, however, is an additional element that fundamentally changes the nature of the playing field every step in a simulation, and exponentially adds complexity to it.
I wasn't saying that Newtonian models can be used to produce convincing simulation, just as an example of a case where simulation is infeasible in one model but feasible in another.A GUT that somehow removes the exponentiating nature of the problem is improbable enough to the point where I don't think it's really worth discussing.
How have you determined how probable that is? How can I estimate the probability of future theories having certain properties?Let me put it another way; if you pick out 189 stars close to one another, and simulate their a gravitational interactions with 3 forces attributed to each body (hideously inaccurate) in a relativistic model (or any other model that doesn't simplify the special cases of any model down to something like newtonian mechanics in special cases), and each force interaction is somehow represented on one bit, and each bit is one atom large, you would run out of atoms in the universe to represent their interaction…
Again, I'm not denying that simulating n-bodies is infeasible to simulate on a relativistic model. I'm saying we can use approximation strategies that are or we may find it is feasible under some heretofore undiscovered model. Maybe you can make an argument that any model that predicts the motion of bodies will have this issue, similar how we can demonstrate incompleteness theorems in all systems with certain basic primitives, but I've yet to see this argument advanced.The following users say it would be alright if the author of this post didn't die in a fire! -
2017-10-05 at 10:33 AM UTC
Originally posted by Lanny I wasn't saying that Newtonian models can be used to produce convincing simulation, just as an example of a case where simulation is infeasible in one model but feasible in another.
I didn't say you did, I'm saying that the reason it's sokvabble in a Newtonian model is because there is fundamental omission in it that makes it much simpler, to where each body does not increase the complexity of the problem because it does not affect the other bodies in anywhere near the same way; there is *one* force atteibuted to each body in Newtonian models.How have you determined how probable that is? How can I estimate the probability of future theories having certain properties?
I would counter by asking how you can find it reasonable to believe that future, more accurate models, will be simpler for special cases than older, inaccurate models, specially when thus far there are no phenomena that contradict relativity. A newer model can probably give a better explanation of what exactly is going on for what specific reasons (that's the whole point of shit like string theory) but until there is a reason to doubt the nature of the phenomena described by relativity, how is it reasonable to believe that the problem will go down in complexity?Again, I'm not denying that simulating n-bodies is infeasible to simulate on a relativistic model. I'm saying we can use approximation strategies that are or we may find it is feasible under some heretofore undiscovered model. Maybe you can make an argument that any model that predicts the motion of bodies will have this issue, similar how we can demonstrate incompleteness theorems in all systems with certain basic primitives, but I've yet to see this argument advanced.
I've already given you the practical problems with any theoretical solution; to suggest that it's possible for us to solve them would require evidence on your part. I'm not saying the simulation hypothesis is impossible, I'm saying that there's no good reason to believe it. -
2017-10-05 at 5:06 PM UTC
-
2017-10-05 at 5:16 PM UTC
-
2017-10-05 at 5:18 PM UTC
-
2017-10-06 at 9:08 AM UTC
Originally posted by Captain Falcon I didn't say you did, I'm saying that the reason it's sokvabble in a Newtonian model is because there is fundamental omission in it that makes it much simpler, to where each body does not increase the complexity of the problem because it does not affect the other bodies in anywhere near the same way; there is *one* force atteibuted to each body in Newtonian models.
I would counter by asking how you can find it reasonable to believe that future, more accurate models, will be simpler for special cases than older, inaccurate models, specially when thus far there are no phenomena that contradict relativity. A newer model can probably give a better explanation of what exactly is going on for what specific reasons (that's the whole point of shit like string theory) but until there is a reason to doubt the nature of the phenomena described by relativity, how is it reasonable to believe that the problem will go down in complexity?
It's reasonable to at least float the idea that complex models don't give way to simple ones, indeed much of the sciences progress by taking simplistic models and creating more complex but accurate incremental replacements. But simpler models outmoding more complex ones is not without precedent. Phlogiston models of combustion are frequently considered more complex than their modern replacements, and a great example in the motion of heavenly bodies is the heliocentric model's simplicity relative to the geocentric model it replaced. Anticipating the motion of planets was a lot more complex under the geocentric model. At at very least we can rule out any general principle that more accurate models need be as or more complex than the ones they replace.I've already given you the practical problems with any theoretical solution; to suggest that it's possible for us to solve them would require evidence on your part. I'm not saying the simulation hypothesis is impossible, I'm saying that there's no good reason to believe it.
You seem to be ignoring the approximation algorithms I've mentioned a couple of times now though. Would you mind speaking to that point?The following users say it would be alright if the author of this post didn't die in a fire! -
2017-10-06 at 9:11 AM UTC
-
2017-10-06 at 9:13 AM UTC
-
2017-10-06 at 9:26 AM UTCFML, chrome shut down in the middle of me writing a giant ass reply to Lanny
-
2017-10-06 at 9:56 AM UTCI'll just keep this short
Originally posted by Lanny It's reasonable to at least float the idea that complex models don't give way to simple ones, indeed much of the sciences progress by taking simplistic models and creating more complex but accurate incremental replacements. But simpler models outmoding more complex ones is not without precedent. Phlogiston models of combustion are frequently considered more complex than their modern replacements, and a great example in the motion of heavenly bodies is the heliocentric model's simplicity relative to the geocentric model it replaced. Anticipating the motion of planets was a lot more complex under the geocentric model. At at very least we can rule out any general principle that more accurate models need be as or more complex than the ones they replace.
Phlogiston theory, for all it's mistakes, could still be used to make a valid model with practical predictive power, even if that model would fail in other aspects.
For example, I could make a model of phlogistination that could tell you how long a flame from a certain object of a certain mass would last in a certain volume of air.
Similarly, I could make the claim right now that the earth is flat and instead of gravity, the reason why things go "down" is because the earth is accelerating "up" at 9.8 m/s, therefore everything falls "down" at 9.8 m/s. That wouldn't be correct but the experimental predictions in a certain context would hold up, and I doubt any model will disprove that.
In much the same way, I don't know how one can justify the idea that something with so much physical evidence as the idea that each body with a mass has a buttload of individual, interactive elements that can happen with other masses and is thus of an exponentiating nature, can be disputed. This almost seems like a self evident truth at this point, although it doesn't need to be.You seem to be ignoring the approximation algorithms I've mentioned a couple of times now though. Would you mind speaking to that point?
I'm sorry, I tried to address it. Let me give it another shot:
You'd agree that an approximation model is close, but incorrect (that's basically a tautological truth) right?
Would you say that with a relatively complex system, with many of these approximated errors interacting, these approximations would cumulatively deviate noticeably from real observations over time? Genuine question, I don't know what kind of error correction goes on in current star simulations etc.
If not, and we can keep it looking 99% correct over long timescales with most elements of our simulations looking good, I concede. -
2017-10-06 at 10:32 AM UTC
-
2017-10-06 at 10:36 AM UTC99 posts of shit in this thread, 99 posts of shiiiiit. Take one down, pass it around, 99 posts of shit in this thread.
