These ideas may be fun to think about, but ultimately they're meaningless. It's certainly not a new idea that some being on another plane of existence brought our universe into being on a whim, and can modify the parameters of its existence at will.<p>If that's so, it's highly unlikely we'll be able to poke any holes in it. Nor is there any reason to believe our memories or any of history is real. Maybe scientists have poked holes in the simulation a million times, and each time, the supreme Code Monkey suspends the simulation, fixes the bug, and rolls back the state to before the hole was poked.<p>For that matter, it's silly to believe that our puny existence has even been noticed by whoever spun up the entire Universe.<p>Most of the questions that this premise supposedly solves are mostly based on fallacious reasoning. Why is the universe fine-tuned for us? What are the chances of this specific universe? How likely is it that intelligent life exists on our planet? You may as well ask what are the chances of your particular DNA sequence. Just because your specific genetic code is extremely unlikely doesn't mean it's impossible that you exist.<p>Ultimately, such an idea only raises more questions. So, say we are in a simulation. Now, how does the "real" world work? Where did it come from? How do we know <i>it</i> isn't a simulation?<p>If these are the answers that make the most sense to you, you've stumbled down the wrong path, and it's time to take a fresh look at the world.
This idea of a computer simulation is what made me become agnostic. At first I was an atheist, I wasn't believing that any powerful entity would create us like this, a world so complex and coherent, just to test our belief.<p>If I had the capacity to build this kind of computer simulation, I would personally do it in an heartbeat. In fact I'm feel like I'm doing this on some games. I would also consider myself to be the god of that simulation, I'm omniscient, omnipotent, omnipresent, etc...<p>That's why I believe there could be a god, it's possible that he isn't even aware of our existence and he is just currently playing a big game of Universe Sandbox. However I also came to the same conclusion of that article (at least, based on the title, I still haven't read it, it's on my to read list), does it really matter?
> Should any error [in the program] occur, the director could easily edit the states of any brains that have become aware of an anomaly before it spoils the simulation.<p>This seems really problematic to me. Why would you assume that a full-on universe simulator has a concept of a brain? If you're in the business of simulating the interactions of nanoparticles, having a "brain" object doesn't seem to make much sense. I suppose that you could implement some heuristic to identify minds with some degree of accuracy, but then you're getting into some weird territory. Someone wants to model the <i>entire universe</i> just to observe life here on earth? Why assume the author of this system is even aware of us at all?<p>> It could be the case that one planetary civilisation is all that can be simulated, without running into computational capacity issues.<p>Again, why would a system like this be optimized toward the civilisation level?<p>The idea is cool, maybe probable, but this write up seems pretty fanciful.
Obviously this is a theme that has been explored in SF a fair amount (recent examples include Ken Macleod's The Restoration Game and Iain Banks' Surface Detail). One obvious problem is that it appears to be highly immoral to generate a simulation in which millions or billions of sentient beings suffer, which presumably would weigh heavier on a more advanced species (human or otherwise).
There is a very general argument that we can't simulate anything remotely resembling our universe inside our universe: <a href="http://www.tjradcliffe.com/?p=1220" rel="nofollow">http://www.tjradcliffe.com/?p=1220</a> Short version: at the precision required for the degree of agreement we see between theory and experiment, we could simulate at most 0.5% of our universe, which simulation would not be capable of star formation because there wouldn't be enough matter in it, and a universe "above" ours that was simulating us would collapse immediately after the Big Bang because it would contain at least 200 times as much matter as ours.<p>The responses to that argument are two-fold:<p>1) start madly making up increasingly implausible auxiliary assumptions to save the claim that we might be living in a simulation<p>2) admit that the odds we are living in simulation are rather low, and in particular the argument that we almost certainly are thanks to recursion is extremely poor.<p>I favour the latter response, as the former seems to involve either claiming that the universe that is simulating us has fundamentally different physics, or that the simulation isn't actually a simulation but rather a game-like approximation that somehow gets updated with sufficient local detail that no matter what kind of experiments we do or observations we happen to make, there is special-case code for ensuring the results look like a consistent underlying physics.<p>Claiming the universe simulating us has fundamentally different physics requires that we drop any claims about it based on the physics of our universe, in which case it becomes untestable because anything we see may or may not be due to a simulation being run in a universe whose laws are not like our own.<p>Claiming the approximation is fixed up by special-case code whenever anyone thinks of doing an experiment or making an observation that might reveal the lack of a full physics engine is likewise putting the hypothesis beyond testability.<p>So either the simulation hypothesis is wrong, or untestable. Neither of these is very interesting.
The speed of light is just a clever hack for rendering performance much like binary space partitioning made the rendering of our own early 3D worlds performance good enough to enjoy.
I think there could be some interesting conclusions or capabilities from being in a simulation.<p>First, we would expect that the simulation may use heuristics; that some things not "observed" may be simulated in a quicker or more crude way.<p>Second, there may be a way to access the computational substrate. If the universe exists within a simulation, is there a way to run software directly against the universe "computer"? Are there security vulnerabilities in the simulation that could allow access to the higher level substrate, either to discover more about it or enact otherwise impossible changes in our universe?<p>Finally, there is the potential for external threat. In a simulation, one time step may take a variable amount of "real" time to complete. If we do find a way to tap into the computational substrate, or otherwise find some actions that burden the simulation excessively, the amount of time each of our universe's timesteps takes to complete in the external host "universe" will increase, possibly to the point where the author of the simulation interventions to correct the problem.
1) Probably.<p>2) No.<p>Hell, even if it isn't, is time real? Is anything? Or do we create the reality we perceive through our perception of it?<p>Mused the other day on Descartes, and "je pense, donc je suis" is tautological. Everything is, as any definition of anything by us is inevitably from our reference frame, and is based on precepts we aren't even aware of.
If it's the case that we are currently a part of a simulation then we can assume those above use are part of a simulation too and eventually we will create a simulation below us and it's turtles all the way down....and up.
It seems to me that a consequence of the second principle of thermodynamics is that your simulations are always going to be (much) smaller than your resources (I'm using the word loosely).<p>So you can't go very far down the road of simulations in simulations in simulations before the "deepest" simulation is to small to have conscientious beings. So the argument that we are in all probability in a simulation seems to break down.
People always saw the universe based on their current knowledge, in the digital era we believe it's a computer simulation some time ago we thought there were turtles all the way down.<p>We are still at the philosophical stage, science is too young for these kind of answers so the question is when we will reach the necessary knowledge level to comprehend it and if this is possible because we are also part of it.
The notion of simplification, which I hadn't considered, might explain - if testable - certain things such as Heisenberg's Uncertainty Principle, i.e. limiting the precision with which one can measure a particle's position or momentum would make it more difficult to discover the underlying true reality.<p>I'm tempted to use Occam's Razor but probably can't because the simulation hypothesis assumes an intelligence created it. And intelligent beings have demonstrated more than once the capability for subterfuge.<p>I think to test the simulation hypothesis one might have to assume a couple of things - 1) the sim has resource constraints, otherwise the creator might as well build a real universe (thanks to Pete Bonani of Falcon 4.0 for sharing a similar insight); 2) every object in the sim has a state and a function which acts upon that state.<p>I'm not sure how someone inside a sim might test these assumptions though I suppose we could try making babies and hopes that overloads the sim before we run out of food.
An aspect that the blog post did not get into as much is what does it mean for something to be a computation or a simulation. Formalization of computation is well trodden ground.<p>Lets take the state transition view of it. Our PCs, phones or any other physical thing that we agree is a system 'that computes' is just following the laws of physics, or state evolution. We engineer the initial condition so that it converges to something that is of interest to us. So does it mean it is necessary for the existence of a conscious observer / intervener / interpreter of state for something to be deemed a 'computation' ?<p>Even if no one is there to observe a particular state or have interest in it, if the system happens to be initialized at some state it would 'compute' the result (end state) no matter what. So what would be that fundamental difference between a simulation and a universe following physical laws ? I think the issue is not whether this is a simulation but whether someone is consciously simulating it. The thing is that it need not be an external entity, embedded entities themselves may reside (perhaps voluntarily) in a simulated experience (hallucinogenic drugs), insanity (socially imposed conformance / compliance), schizophrenia.<p>The notion of reality is messy business that ties one up in knots. What I find interesting is how two ancient cultures : (i) native American and (ii) Indian thought about it, how they answered when I am dreaming is that real or is that fake. It is really hard to argue that what we call real is a more privileged position than the dream world or a hallucination when we are in it. Some native American philosophies decided that both are equally real. If some one flies in a hallucinogenic trip the person is really flying in that world. There are ways to get in and out of those worlds. The ancient Indians or the vedic philosophers took another route, they chose that both these worlds are just equally unreal.<p>The question of whether this real or not is pretty much as old as thought.
There seems to be this background idea here that "If you can't disprove it, that means it's plausible," which is totally mistaken. Not every idea that we can't disprove is worth seriously considering.<p>Let me put it this way. If a fundamentalist Christian made the same kind of arguments, and said "therefore God probably exists", would you believe that it really was probable? Most likely you'd say "I'm pretty sure that God doesn't exist, even if I can't prove it, and I'm not going to change my view of the probability just because this person says something that sounds good." But if you think that's a reasonable response, why is not the same approach reasonable in response to this "the universe is probably a simulation" stuff?
Maybe I am ignorant but trying to distinguish between simulation and reality seems to me a lost case right from the beginning. You have to look for a difference but we know neither the rules of the simulation nor of reality. If we look at the universe and find a difference between what general relativity predicts and what the universe does, maybe it even looks exactly like floating point rounding errors - this tells us nothing. It might be an imprecise simulation but it may as well be that general relativity is just an approximation and the real behavior of the universe is floating point general relativity.
Energy, though.<p>I mean, the host universe may have completely different laws of nature, but assuming that the 2nd law exists in it...<p>Organising information (storing it in RAM, carving it on a rock...) decreases entropy and thus takes work (entropy overall increases, because you had to get the energy to do your work by increasing entropy somewhere else by more.)<p>The cost of accurately simulating an entire universe, down to atoms, would be <i>extreme</i>, like you'd need galaxys worth of stars' to even begin.<p>At which point, the question becomes, why not just observe the real universe?
"However, Bostrom recognises that a complete emulation of reality on every level is likely to be impractical, even for powerful computing systems."<p>Doesn't this assume that the simulation is running in a universe where our laws of physics apply? Couldn't what we think of as our laws of physics be a programmed part of the simulation, running on computers in a completely different universe where it's impossible for us to do more than speculate on matters of practicality?
> Simulating an entire Universe with sufficient detail to include conscious minds will be complex, even if the fundamental rules underlying the program are simple. It seems needlessly baroque to programme something as complicated as that, when you can learn just as much from something simpler.<p>Scientists and engineers use many-body simulations to investigate emergent properties of systems all the time. Not all understanding comes from first principles only.
I really like this short story that illustrates the concept:<p><a href="http://qntm.org/responsibility" rel="nofollow">http://qntm.org/responsibility</a>
Are you kidding! Of course it matters!! If it's a VM, we can start hacking it, until we find the universal equivalent of a buffer overflow or some kind of edge case branch mispredictoin, and can start injecting our own payload. After a little bootstrapping, and experimentation that hopefully doesn't crash the VM, we are now running whatever physical laws we want. Matrix time!
I have been giving this a lot of thought lately, especially after reading Superintelligience. I actually think that it is somewhat likely, or even highly probable given how we would model out our own behaviour and hardware. Further, to use a video game analogy, we may actually be non-primary characters, that are being modeled by a different, more sophisticated, group of actors.
If you're interested in this idea, here's a fun short story: <a href="http://qntm.org/responsibility" rel="nofollow">http://qntm.org/responsibility</a><p>Also, Permutation City by Greg Egan is a good read.