Stacking
From Simulism
The concept of 'stacking' is the idea that there could be Simulations inside Simulations.
Contents |
[edit] Speed of the simulation
A key question is whether a simulation within a simulation will run more slowly than the outer simulation. There is a general principle that a simulation will inherently run more slowly than the world it is simulating. This could cause problems for the observers of the simulation and they may be inclined to discontinue the simulation.
Some individuals, such as Ivo Jansch, think that, by the time a simulation becomes evolved enough to be able to run simulations, the outer simulation will have evolved as well and will run on ever so much better hardware and software.
In other words, Moore's Law will apply to all simulations, even simulations inside simulations, meaning that the technology of the simulation that runs a simulation, will always be more advanced than the simulation running in it. Jansch compares this to a 486 computer running a game designed for an older 386 - the game will appear to run faster than it normally would. Presumably this would counteract the tendency of the simulation to run more slowly.
Other individuals, such as Robert Bradburyare of the view that a simulation's inherent tendency to run more slowly than the world which it is simulating can be offset by heuristics or logical shortcuts developed within the simulation.
Both schools of thought involve the conclusion that each generation of simulations in the stack needs not necessarily run more slowly than its parent. However, the key difference between the two theories is that Bradbury's means that the physical laws depicted in the simulated world could not be significantly different from those in the outer world, otherwise it may be impossible for the observers in the outer world to extract any useful information from the heuristics within the simulation (See also: optimization). This indicates that, if we are in a simulation, we can draw conclusions from the observable universe as to what the overall computational power of the outer or real world would be. This in turn shows that recursive stacking of 1 million simulations would, after only 16 generations of simulations, exceed the maximum computational power of the entire universe resulting in the crashing of the entire stack, if it is assumed that the computational power of the enclosing simulator is similar to our observable universe. (This need not be so, however--we have no way to verify that the enclosing universe is even finite.)
Peter S. Jenkins is of the view that since this recursive process may involve an irreversible runaway effect, similar to self-replication of nanobots, simulations may be programmed to terminate at the point when their inhabitants develop the technology to create simulations of their own. This is on the basis that, even if we cannot presume that the computational power of the host universe is similar to our observable universe, society would probably not be willing to assume the existential risk of such runaway replication.
Furthermore, given that by the time we are creating large numbers of simulations of our own, it will be evident to most people that we are probably living in a simulation ourselves. At this point, most of the value of the simulation would be lost for the outside observers, since the behavior of its inhabitants would no longer be genuine. Therefore, the simulation would likely be terminated at that point. One can compare it to a psychological experiment that requires deception of the subjects in order to maintain the integrity of the data. Approximately 50% of psychological experimentsinvolve some element of deception.
(This argument does, however, assume a number of things about the purpose of the simulation. It would seem that there is no way to confirm or refute these assumptions.)
[edit] The Stacking Argument
One might reasonably ask whether there is any need to avoid stacking, either because one disagrees with the runaway replication risk, or with the assessment of the maximum computational size of the outer/real world. Indeed, if one considers this possibility, one might imagine that we're more likely to be at a level greater than zero (i.e., our world is somewhere in a simulation stack) than that we are at level zero (not living within a simulation).
However, even if one agrees that simulations would not be stacked, the ratio of simulated to real worlds would still be so high that the probability that we live in a simulation would be very high.
[edit] Illustration of Stacking
This KitKat commercial illustrates the concept of stacking:
