The scientific paper deals with "the hypothesis that the observed universe is a numerical simulation performed on a cubic space-time lattice or grid".
Just because the universe may operate with computer-like regularity and grid-like data structures does not mean "the universe is a computer simulation." At best it means the fundamental laws of physics are "computer-like".
Whether "resembles a computer simulation" equals "is a computer simulation" is a metaphysical/philosophical question.
Finding uncanny resemblance between our own reasoning concepts and the universe is not new - and this uncanny mirroring between abstract mathematical concepts and the laws of physics has been known for quite some time. It's a mystery, not an answer. See: "The Unreasonable Effectiveness of Mathematics in the Natural Sciences" by Eugene Wigner:
Indeed. And conversely, proving that the observed universe does not operate on a cubic space-time lattice would not prove that it is not a simulation of some sort--it would merely prove that it is not a simulation of that sort.
...can't Wigner's "unreasonable effectiveness" be explained in some way via the anthropic principle? (there should be a way to reason that intelligent life like us could only have evolved in an universe where predictions about physical systems can be made considering only a small subset of the initial condition variables, and also that "abstract mathematical ideas" could only appear in someone's mind from a sort of intuition of the universe's laws, so that the mathematicians' intuition can't come "out of the void" but is actually subconsciously "inspired" by the laws of nature at some level and this is why it explains them, and that what "feels" "simple and elegant" to a mathematician actually feels this way because it's in accordance with the universal laws of nature by witch his mind also works...)
If we're simulation, we're almost certainly emergent phenomena (laws of physics as we discovered them doesn't say anything about humans, they operate on much lower level of abstraction). Those, that run our simulation may even don't know we're there.
Imagine intelligent life evolved in huge game of life grid (it's possible assuming inteligence is turing-complete). We may look at the strange patterns and never know they are intelligent, never mind communicating with them, or decoding what they're up to from the patterns.
This work attempts to draw out the presence of a simulation by looking for tiny violations of known/expected symmetries, which is always worth doing. We should note, however, that an excellent simulation could exactly cover its tracks, lest the underlying architecture cause unintended artifacts.
From a CS perspective, the simulation's developer might create unit tests that guarantee the preservation of symmetries.
A theorist who develops an experimental test that could prove we do not live in a simulation should be lauded. There are experimentalists (like me) just waiting for such guidance. John Bell's theorem [1] was/is a shining guidestar for physics.
I think they're assuming that the simulators don't have infinite resources, so eventually we'll find a 'wall' in the simulation.
"With finite resources at their disposal, our descendants will likely make use of effective theory methods,
as is done today, to simulate every-increasing complexity, by, for instance, using meshes
that adapt to the relevant physical length scales, or by using fluid dynamics to determine
the behavior of fluids, which are constrained to rigorously reproduce the fundamental laws
of nature."
It's possible that if we were simulated, the simulators aren't actually trying to hide the nature of our universe from us. It's also possible even though the simulation started with a 'simple' framework that they could have hidden better if desired, that by becoming a self aware information system we can no longer be switched off / restarted without violating their ethical constraints. So it's possible that the limits a simulation actually are detectable.
Of course as someone said, if the simulators were actively monitoring your thoughts with the intent of stopping you from finding the boundaries then there truly would be nothing you could do.
Our decedents might have a certain level of technological advancement that Moore's law works a lot faster for them. If it is above a certain level, we would not fast enough to detect that limit.
That's a pretty silly statement. Surely you can prove that 1+1!=3, which is the negative of 1+1=3. You can also prove that pi is not a rational number, 1/2 is not an integer, etc. There's a post on the front page about attempting to prove P!=NP.
Is there a distinction between proving and demonstrating by proof? I did a quick search (I'm hardly a mathematician), to quote Wikipedia:
> a proof must demonstrate that a statement is always true
It would seem that, if you can demonstrate a statement is true for a given logic, that is proof. Contradiction and exhaustion (which is a bit more controversial, I think), while less tidy than one might expect, are valid techniques for showing that a statement is always true, and therefore are valid ways to construct a proof.
Note: I encourage any and all maths professors to come and tell me off if I'm way off base.
This strikes me as equivalent to researchers hundreds of years ago saying "well if our universe is a clockwork simulation then we should be able to detect the gears turning by this method".
Certainly interesting ... I'm not sure that we could currently understand a single principle of a system capable of simulating our universe.
A more apt analogy would have been testing for a specific number of gears with a specific number of teeth in a specific configuration. Choosing very specific hypotheses purely based on your ability to falsify it and no other supporting theory or evidence seems like a low reward strategy.
I'm certainly not trying to argue against pure research in any way.
No it can't. The simulation hypothesis will never be testable. All our assumptions about the hypothetical computers our universe runs on are based on things as they are in this universe. We have no basis at all for applying these assumptions to the host universe. We don't know if the simulation runs on a discrete computer, a finite computer, or whether it resembles a Turing machine in any way. We don't know what kind of physical limits there are on computation, or to what extent those are even recognizable concepts, and that's just a few highlights. We have no idea what it might be like up there, or what can be done.
And don't tell me about "likely", you have no data worth mentioning, in particular about "likely" physics or "likely" motivations of simulators.
This idea is way older. Various mystics already envisioned the metaphysical structure of the world in such way (i.e. creation of reality through symbolic transformation - computational idea similar to Markov's algorithms).
See for example Taam Eytzo by R. Pinchas Eliyahu Horowitz of Vilna (18th century philosopher and mystic): http://www.hebrewbooks.org/21931
If the universe is a simulation, what's to stop the simulation from recognizing that it's being tested and dedicating more resources to making the test result indicate that there is no simulation going on.
The assumptions in this paper are quite weak. It's like testing whether or not your code is running in a virtual machine. It can be done, but only if the virtual machine authors have made some glaring mistake.
Yes, computation is finite, but that doesn't mean that the authors of the simulation can't use more resources than usual simulating your test of the simulation. Not all parts of the simulation need to run at the same level of detail.
Imagine, in the far future, you were able to to create a simulation that was so advanced that it could accurately recreate everything that has happened in the universe so far, so that it was an exact replica of history. Also, that there is a way which you can determine whether you are in a simulation or not. Assuming you are in the real universe, if you make this measurement, it will show false, however when your simulation reaches that point in history, the result would show up true, and reality and simulation would diverge. The only way to stop this would be to make a more advanced simulation that could not be detected with any method, or to have some part of your simulation check for when this test is made, and spoof the result...
This assumes that reality is simulated using a grid, and particles are either in that cube or that one.
But that's not the only way to run a simulation - you can make the fundamental particles the building pieces and simulate them - they can move anywhere they want, in any direction, they don't live in a grid.
The only thing that matters is their relationship to the particles near them.
Which follows very nicely from relativity which shows that there is no absolute coordinate system, or absolute energy, only relative distances and relative energy differences.
i.e. like the difference between polar coordinates and rectilinear ones.
Also the simulation doesn't really even have to involve particles. Stephen Wolfram did some interesting writing on concept of our reality / laws of physics emerging from simple computational systems like cellular automaton [0]. Also Eliezer had some interesting points about how a continous interpretation of causal networks gives rise to some concepts we recognize from physics, like space, time and the speed of light binding them together [1].
food for though (mostly offtopic): what if most simulations are actually nurseries, run by our descendants as an alternative to mostly random procreation?
...I imagine that in a very distant and complex hypothetical future, the only way that super-intelligent beings could reproduce in a way that the offspring would truly be similar enough to their parents, so that they could feel empathy one towards another would be either by: (1) clone and mutate an existing individual (think mind-cloning, like software copying) and mutate, and (2) create "children from scratch" and raise them, for at least a few "eons" in successive simulations of their parents or ancestors pasts (any other alternative, I'd imagine, in a post-singularity world evolving incredibly fast, would result in offspring so different from their parents that it would make empathy and the existence of a shared culture with shared artistic and aesthetic values impossible...)
...maybe we're just "babies" in the first level of a simulated nursery, waiting to complete the stages to "adulthood" in world of unbound, "software like", intelligent beings :)
EDIT+: ...just realized that this could be made up into pseudo-scientific pseudo-theory of reincarnation (which I don't believe in, but I find it interesting) ...if any buddhist or hinduist mystic reads this and takes my idea and does something intersting with it, please give me credit as original author :)
Does it matter? Think about a simulation in a game on your PC. Nothing in that game actually exists outside the simulation. i.e. as far as anyone/anything in the simulation is concerned, the simulation is real. You cannot leave the simulation. You cannot end the simulation. There's nothing unreal about the simulation for anyone in it.
this happens when you die, universe segfaults, or it continues running in hell, limbus or heaven. depends on where it lands. hell has 9 rings. goal is to end up in ring 0 . heaven
At any level of simulation we know that when we take time, it takes time in the reality that hosts us too. It's nice to know we'll always have that in common.
Me layman's hunch is that we would only be able to test this if either (a) the simulation is flawed, or (b) the simulation designer chooses to give us hints (perhaps more subtle than 'miracles' or revelatory visions).
The mystery between DMT and brain chemistry is also a valid area to understand such an idea. DMT is a naturally occurring substance that's in the human body, animals and various plants. The mystery is that it is almost identical to serotonin and tryptophan.
But it is not like any other drug. It's almost unfair to call it a drug. Under its effects, reality looks radically different. You can see the math and inherent aspects within physical reality. With your eyes closed, the experience is even more profound.
Of course, it's likely just an artificial experience, a chemical delusion. But it an amazingly mystical experience, that can be duplicated over and over. It's unlike any other psychedelic and is the only scientific way I'm aware of that induces a "real" mystical/spiritual experience.
How is "fake" pain different from real pain? From my understanding, our neural correlate of consciousness is in fact a transparent self model simulation our brain is running, so in that sense all conscious pain experience is "fake". Or do you mean something else?
While I don't know if it's actually deserving of the title "implausible nonsense", I think its important to remember that stupid is as stupid does, and same goes for smart. Someone isn't smart because they have the trappings of intelligence, its because they did something smart.
"its because they did something smart"
So intelligence == 'achievement'? That makes no sense.
As for nonsense, ok I'll retract that, it is an interesting but highly implausible topic that gets a large amount of discussion relative to it's likelihood. It's actually just another way of phrasing "Does God exist?" If there was a computer (God) simulating this universe and controlling it, then it'd require a host universe for itself to exist, and it's complexity has to be greater than what it is simulating. Therefore it is less likely, and not by an insignificant amount.
I guess its true that you can't just equate intelligence with achievement. In my own life, I think I've increasingly used achievement where I used to use intelligence to gauge how 'smart' someone is. That comment was also trying to get at the annoyingly pervasive belief people seem to have that you must be smart because you are a mathematician, say. As an example, novices might think someone is good at basketball based on how they look on the court- but you wouldn't say someone is good unless they actually had a big positive effect on how often their team won. In general, I guess when I say smart, I mean "good" in some abstract sense.
What would be the purpose of a simulation with so much pain and suffering. Where children get raped, flayed, abused, etc... And don't give me any of that matrix crap, that was only to make a flawed concept make just enough sense that you didn't fully think it through until the next bullet time scene. I'm certainly not on the outside paying for this. I have a degenerative bone disease that's caused me to have 6 major surgeries in 6 years. If this is a simulation, when I get out, I'm going to kill the son of a bitch that put me in here.
That question is answered by theodicy- the philosophical field of justifying God's dealings with the universe. I recommend C.S. Lewis's _The Problem of Pain_. The short version is that if you give people free will, then there is a risk that they will try to hurt each other; if you always intervene to ensure that they can't hurt each other, then their free will is worthless; and if they don't have free will, then what's the point of them living?
Now as I said, that's the short version, and in such drastically abbreviated form the argument has holes you could drive a humvee through. An alternate explanation is that the creators of the simulation simply don't care about us at all, and we and our problems have nothing to do with its purpose. One might also ask, if the creators do care about us after all, why didn't they just make everyone morally perfect so they wouldn't freely hurt each other? In which case me-with-my-theologian-hat would say "maybe a life of pain is the very process by which morally perfect people are created". But theodicy is a very large field which I cannot adequately summarize in this post, so if you want real answers, just start with that book.
I sometimes envisage the simulation as the random by-product of some other computation. That we exist in the noise, the margins of some extremely large (maybe "game-of-life"-like) computation. The runners of the computer computation don't know what is happening at our level, they just see the underlying binary representation and the eventual result. There can be a large stack of emergent patterns, from the bottom-most jumbling of zeros and ones to quarks and eventually atoms and molecules. Even if they are looking at molecular level they would have no idea that anything alive and conscious and suffering is in there. Just funny moving patterns! How could they recognize a completely alien ecology running at arbitrary speed?
The problem with applying that (very reasonable, in my opinion) argument to this discussion is that it invokes assumptions of benevolence, or even planning and/or intent.
Those things are largely irrelevant to the question of simulation. There should be no implicit assumption that a simulation must have any sort of "deity" behind it. Nothing with intent, purpose, planning, ethics, etc. It seems much more reasonable to me that such a simulation, if it is in fact the case, 1) arose "naturally", whatever that may mean, and 2) does not care about us. This universe, simulated or otherwise, does not care about humans. Why should we assume the proposed simulation does?
Well, the idea that the simulators care is the basis for the standard argument that we're probably in a simulation. After all, it's supposed to be posthumans running simulations of their evolutionary ancestors. I don't know why they would do that, but given that premise, why run simulations of your ancestors if you don't care about your ancestors?
Of course, that provides another answer to the problem of pain- that this is how it was the first time 'round, and they want the simulation to be accurate.
Is the simulation just of us though? Or is the simulation of a much larger portion of the observable universe? We naturally want to be important, but I don't see any particularly compelling reason to think that we are. I don't see any reason at all to assume that if we are in a simulation, it is being run by anything remotely resembling us (or used to resemble us), or indeed by anything at all.
Is that such a strange concept? The prevailing school of thought, if the universe is not a simulation, is that there is no mastermind behind it. Seems natural to me that we would assume the same about any universe simulator.
That made me thing of a tangent - could the incredible distances between stars and the relatively slow speed of light be intentional, a way of preventing the simulation from grinding to a halt if any of the intelligences ever learn how to travel in space?
Maybe there's not enough resources to completely model everything we see in the sky, so there needs to be some mechanism to throttle exploration.
It seems to me that's kind of the defining feature of a simulation. If there's no one behind it, if it just exists entirely on its own, then it is the natural universe, not a simulation. Even if it looks computationally-based, that's just The Unreasonable Effectiveness of Mathematics at work.
A simulation means "someone made it". They may not be masterminding details of our lives, or our civilization, or even our solar system if they only happen to care about something completely different from that and incomprehensible to us, but nevertheless calling it a simulation implies that a creator at least exists.
You are just pulling that concept from the traditional definition of "simulation", but there is no particular justification for doing so.
If we detect evidence of what the author of the paper is calling a "simulation" (that is, our universe is being simulated in a 'meta-universe'), then we know only one thing about the meta-universe: It can simulate ours. We don't know that it supports some sort of "life" of it's own, contains anything with "intent", or contains anything like that at all. We do not know that there are meta-men in the meta-universe.
Perhaps there better terminology for such implications than saying a 'meta universe is simulating our universe', but the fact that we are using that terminology in the meantime does not mean it is reasonable to make the assumptions about it that you are making. Just because we call our universe a "simulation" does not mean that there must therefore be meta-men.
(Also I do not understand the point you are trying to make about it being the "natural universe". If some sort of meta-man made a simulation that we live in, instead of the simulation arising through non-deliberate means in the meta-universe, is his meta-universe any less "natural"? Either way the meta-universe is "natural"; whether meta-men exist or not)
In a nutshell what I am saying is that we should not assume intent unless there is evidence for intent. Evidence for 'simulation' itself should not be misconstrued as evidence for intent; it would be evidence for nothing but 'simulation' itself.
Gods and simulators are very equal in my book. The key isn't that I don't believe in either (though it's true that I do not), it that both constructs (god and simulators) being completely absent from any interactions with the human life in any recognizable way, simply are irrelevant. Both are not falsifiable, nor are either provable. It would be akin to me wondering if dragons exist on a planet that orbits V12 of NCG 4203.
Though to be honest, I would give better odds that there are dragons on a planet orbiting V12 of NCG 4203 than either of the above existing.
Just because the universe may operate with computer-like regularity and grid-like data structures does not mean "the universe is a computer simulation." At best it means the fundamental laws of physics are "computer-like".
Whether "resembles a computer simulation" equals "is a computer simulation" is a metaphysical/philosophical question.
Finding uncanny resemblance between our own reasoning concepts and the universe is not new - and this uncanny mirroring between abstract mathematical concepts and the laws of physics has been known for quite some time. It's a mystery, not an answer. See: "The Unreasonable Effectiveness of Mathematics in the Natural Sciences" by Eugene Wigner:
http://www.dartmouth.edu/~matc/MathDrama/reading/Wigner.html