The Universe is not a simulation

A lot of people seem to think that the universe might be a simulation. The idea is that if it is easy to simulate a universe, then inside that simulation it is also easy to simulate a universe, so iniside that simulation it is easy, ... etc. Therefore there is an infinite number of simulations running inside each other, and the probably that our universe is not one of them is approaching 0, therefore it is most likely that we are in a simulation.

The problem I see with that is that any computer can necessarily simulate an area of space that is less big than itself. For example if my computer is made up of 20000 atoms, and each atom is a memory block storing information useful for the simulation of atoms, then I can only simulate a universe consisting of 20000 atoms, at best. If this was not true then a computer would essentially be able to simulate a computer with more memory than it has, inside of itself.

This seems like a very basic violation of some physical law, but I am not sure. Can someone prove me wrong? Does quantum computing solve this problem somehow?

Other urls found in this thread:

en.wikipedia.org/wiki/Ontological_argument#Anselm
twitter.com/NSFWRedditVideo

The implications for the multiple simulations theory therefore seems to be that each simulation can only run inside a universe that is orders of magnitude more complex than itself, because it would require constructing a computer that is orders of magnitude more complex than the entire size of the universe.

o hai

Hi

care to elaborate?

>For example if my computer is made up of 20000 atoms, and each atom is a memory block storing information useful for the simulation of atoms, then I can only simulate a universe consisting of 20000 atoms, at best.

This is not just wrong it's a completely incoherent statement that cannot be refuted because it's a non sequitur.

>For example if my computer is made up of 20000 atoms, and each atom is a memory block storing information useful for the simulation of atoms, then I can only simulate a universe consisting of 20000 atoms, at best.

unless the universe it is simulating can be modeled by a set of equations and initial conditions

Who told you the universe that is real that contains what's simulating our universe even has atoms?

Or that our universe is even a measurable fraction of the size of the one that is simulating it.

It doesn't make any fucking sense that there would be a size and speed limit in the universe. It's as if they don't want us breaking their precious sim.

>For example if my computer is made up of 20000 atoms, and each atom is a memory block storing information useful for the simulation of atoms, then I can only simulate a universe consisting of 20000 atoms

That's what virtual memory and disk paging were invented for, friendo

Who says you need to simulate the entire universe at once?

Load 1% of the universe in memory, simulate a tick, unload it to disk, load the next 1%, simulate a tick, and so on.

THIS

>mfw realized that the ``````````global elites'''''''''' are just the in-game personas of the aliens managing this simulation

>GTFO shitbrain
tl;dr

Donald Trump is the main admin

this seems more philosophy that anything. it starts into whether or not other people are real or just you, if things exist when you can't see them, and if this is a simulation where are my cheat codes

If this is a simulation there should be some way for an incredibly intellectual individual to escape should they need to

how do you turn this on
how do you turn this on
how do you turn this on
how do you turn this on
how do you turn this on
how do you turn this on

I'm not following.
>unless the universe it is simulating can be modeled by a set of equations and initial conditions
But each atom in the universe has it's own initial conditions, so wouldn't that mean you need all the atoms plus addition atoms to store the equations?
>Who says you need to simulate the entire universe at once?
That's a good point.

What do you mean?

>If humans can create gods to worship, then those gods can also create gods to worship, and so on ad infinitum. Therefore, I am a god.

This is what simulation theory proponents ACTUALLY BELIEVE

i dunno man, maybe if they make a Sims 99999 and it's really advanced, then you could be a god of a sentient species who are being simulated on your computer

Your brain simulates a universe from your senses so you can comprehend enough to propagate the species.

Your experience is not the universe, however it is good enough. Also the reason illusion works and people are generally retarded.

>if it is easy to simulate a universe
What if it's not.

I guess you could say that, but my intention was just to critique the logical structure of the simulation argument, in particular the way it draws a definite conclusion about the world from an infinite chain of hypothetical contingent events. Not only do you have to believe that it's "easy"(whatever that means) to create a simulation in our world, but you also have to believe that it's "easy" for the simulation to create a simulation of its own, and so on. But to my knowledge, no one has come close to creating a simulation of a world which is sufficiently "rich" to create its own simulation. So even accepting the premise of the argument requires a substantial leap of faith. But suppose we do take this leap and accept the existence of an infinite tower of simulations, then so what? We already know that our world is distinguished from the others by virtue of it being ours, so assigning a "probability" to us living in a simulation is meaningless and obfuscatory. If your name was John, and you had 100 brothers named Jack, then would you say that your name is "probably Jack"? The quality of being/not being a simulation is pre-determined, and we cannot "roll the dice" by resetting our universe and seeing whether we land in a simulation or not, so it doesn't make sense to assign a probability to it (unless it's a probability in the sense of a Bayesian prior, which, let's be honest, is just a fancy name for talking out of your ass).

So what has the argument shown? I paraphrase: Under some highly questionable assumptions, we can conceive of a scenario in which there are many simulations. Since there are many worlds, one of which is ours, and most of which are simulations, it follows that our world is a simulation. QED.

One is reminded of the joke about proofs in economics: "Assuming X, Y follows. We have proved Y."

Sorry, I only understood a portion of your post because I'm a brainlet and know next to nothing about statistics, but I agree with what I understood.

I would definitely want to see a stronger argument before believing we live in a simulation, but it's still a pretty interesting idea to entertain.

You don't need to simulate every single atom every second. It only needs to be simulated to the level of conscious observers (us) perception. Then if one of us just so happen to look at an atom through a microscope then it would be simulated for just that moment. Same goes for distant galaxies and pretty much anything else we aren't consciously observing at the moment

Occam's razor faggots

Sorry if it's a little confusing; i got kind of carried away.

And I agree it's an interesting idea to think about, but as far as I can tell, most of the arguments in favor are basically restatements of en.wikipedia.org/wiki/Ontological_argument#Anselm

Just because something is easy does not mean it will be done. It might be more profitable to make it illegal and to force people to do unnecessarily inconvenient things.

>For example if my computer is made up of 20000 atoms, and each atom is a memory block storing information useful for the simulation of atoms, then I can only simulate a universe consisting of 20000 atoms

do you understand how bits work?

20000 bits can represent way more values than 20000. And if we are talking quantum shit, which is probably what you would need for a universe simulator, that number gets even greater.

Oh my god. A quantum shit. Does it both smell and not smell at the same time? Do you have to lock a cat up in a box to figure it out too?

Let us imagine that there is some configuration of n atoms (lets call this configuration a simulator) that can simulate n + 1 atoms in any configuration, perfectly, and in real time.

Then, let us imagine that our n original atoms configured into a simulator, S_0 simulate a simulator S_1 and one atom. Simulator S_1 simulates simulator S_2 and one atom, and so on. Blah blah blah, finite number of atoms perfectly simulating an infinite number of atoms in any configuration in real time.

Thus either we can simulate an infinite amount of things with a finite amount of matter, or it is impossible to perfectly simulate objects of greater complexity than the simulator in real time. Both options seem equally plausible to me.

Only creations and veiled creationists think this

>he thinks the universe came from nothing

its easy to disprove just call philosophy tards brainlets

>Assuming X, Y follows. We have proved Y.
In this particular example, X="it's possible to make a simulation of a universe that can simulate a universe" and Y="our universe is a simulation". You're right that assuming X does not prove Y, but we are not assuming X. We are just exploring the implications of X, which happen to be Y.
>If your name was John, and you had 100 brothers named Jack, then would you say that your name is "probably Jack"?
But what if you didn't know your own name, and you had 100 brothers named Jack. Wouldn't it be absurd to assume you're special in any way and got a unique name?

>The universe must have come from some thinking agent because I can't see any other way it could have come about

I'm not a scientist, so pardon me if I sound retarded.

I think the question of "Where did it all start?" can be answered with the universe being a simulation. Of course, this creates an infinite cycle, which doesn't necessarily make sense as that still needs a beginning. However, in the root universe, what if the entire concept of an inception doesn't exist? It makes sense that a simulated entity cannot understand its simulator. By understand, I don't mean acknowledging its existence and even being able to figure out how it works and responds to certain things. I mean actually understand how it was put together. This requires knowledge of the universe the simulator was made in.

It's not really even necessary that it runs in real time, but yeah, it would still need to store a very large amount of state data

That's why you don't simulate the whole universe, but only those parts that are being observed at the moment. Much like a game where a landscape only starts to render once you look at it. I just explained to you the wave-particle-dualism. Mind=blown.

consciousness is non-computable so we cannot be a simulation, simultards, matrix fanboys and pseudo-philosphers BTFO

Consciousness is computable. Pseudoscientists, dualists, Deepak Chopra BTFO.

I don't understand half of what this man said, but quads tell me he's right.

>implying we have even the most underlying working of our universe down to even deny or confirm that we're in a simulation

To anyone wanting to disprove OP's logic you have to explain how:

A universe of same complexity is able to be simulated within a universe.

Even with quantum computers it can't be done, if you have any understanding of how simulations of atoms are calculated then you will understand how complex they are

The human brain is made up of matter, matter bunched together forming neurones, neurones forming logic circuits, with enough computing power these are predictable and simulable

While you're right about nested simulations, it's easy to imagine that this is simply a sensory simulation; basically that we're in a VR game and only the parts of the world that need to be perceived at any given moment are actually calculated, to the degree of accuracy required to ensure a realistic/consistent experience in the subject. Or you could go even further and assume that it's a very immersive sensory 'movie' - it's not interactive, all the processing took place during rendering, and running the simulation takes very little processing power. Neither of these two options are precluded by a processing/memory limitation. Basically brain-in-a-vat esque ideas of varying sorts.

Anyone with half a brain realizes pretty quickly that the whole simulation idea is just complete bullshit and not worth discussing at all. It's not even possible to forge a reasonable discussion from it, there are so many random assumptions required, it just leads nowhere. I mean it.

I don't even know what exactly gives people the idea that this seems plausible. Nothing and I mean absolutely nothing in this universe turned out to be discrete with the exception of particle numbers and associated quantum numbers (charge, spin, color etc). The whole rest of it, the mechanics is the exact opposite of discrete. Even the simplest most fundamental little shit events take years to calculate. Look up lattice QCD. We can't even simulate a single fucking proton just existing for an infinitesimal time. Absolutely nothing about this strikes me as particularly easy to plug into a computer. It's the exact opposite of that.

See and then you need to pull the whole "But we don't even know what our parent universe is like???" card, and then the whole discussion gets as hollow as discussing what color god's beard has.

>A universe of same complexity is able to be simulated within a universe.

>same

that problem is solved by normal computers you have in your house

you can program a game world that is always different no matter where you walk and go, it can be infinitely large/different in every direction yet existing on your normal PC

Nice.
Singleplayer or multiplayer?

>size limit
I'm sorry what do you mean by this?
Also if by speed limit you mean light then you are also false. What's to say there isn't a particle/wave that is faster than light that we simply have no knowledge of.

Simulation theory is a fucking meme but this does remind me of procedurally generated content. Executed or shown upon access.

If there was then it would violate causality

It's not a simulation, it's a subjective hallucination as we are boltzmann brains.

>But what if you didn't know your own name, and you had 100 brothers named Jack. Wouldn't it be absurd to assume you're special in any way and got a unique name?

Wouldn't it be equally absurd to assume that you're completely interchangeable with each of your brothers? I suppose it comes down to whether you interpret "probability" in the sense of a limiting proportion of experimental trials in which the result in question is obtained (the frequentist sense) or as a measure of "epistemic uncertainty" (the Bayesian sense). The point i was getting at with the example was that the event "your name is John" does not have a well-defined probability in the former sense, because you can't make yourself get born again and see whether you get named John or Jack. Your name is predetermined, and no amount of additional information will change it. Hence why you (or at least I) wouldn't say your (my) name is "probably Jack": your name is not a question of probability, but rather something definite and immutable which you happen not to know.

That said, while the information that you have 100 brothers named Jack won't change your name, it can change your *beliefs* about your name. This is the Bayesian definition of probability. (On a side note, I think it's very unfortunate that Bayesians decided to use the word "probability" for the quantities their theory produces, since the historical definition of the word is far closer to the frequentist definition. If they had decided instead to call them something like "quantified subjective degree of belief", then I believe that many misunderstandings such as this current one could have been avoided). (cont.)

(cont.)
In this sense, I agree with you: if you have 100 brothers named Jack, it is not unreasonable to define the prior distribution of your name to assign, say 99% to the name Jack, with 1% spread over various other names. But it is important to distinguish this from the frequentist interpretation. When a Bayesian says that the probability that your name is Jack is 99%, this is *his subjective belief*. Bayes' theorem provides a consistent algorithm for updating one's subjective beliefs in response to new evidence, but it gives absolutely no guidance as to how to form your beliefs in the first place ("setting your priors", in the parlance). In the Bayesian theory, two people can see exactly the same data and assign different probabilities to the same result, while both being "correct". (For the sake of fairness, I should mention this problem is not absent in frequentism, although it is very nearly so in the simple case under discussion in which the experiment has only two possible results-either we are in a simulation or not, either your name is Jack or not).


So I concede that saying your name is "probably Jack" or that we "probably live in a simulation" are justifiable, provided we interpret "probably" in the sense of a Bayesian prior (which is a quantified subjective degree of belief). So as not to conflict with the traditional definition of probability, then, it would be better to say that you "believe with a high degree of certainty that your name is Jack/we live in a simulation". And I concede that it is possible for an argument such as the one under consideration to change someone's *subjective degree of belief* of whether or not we live in a simulation. However, someone else could assign a prior probability (read: degree of belief) of 1 that we are not living in a simulation, and there is nothing in the theory that prohibits this.

(cont.)
(One final pedantic point is that it is often assumed that there are infinitely many words and that each is "equally likely" to be ours, so our prior that any given world is not a simulation should assign the same value to each world. But there is no such probability distribution on an infinite set. So one is forced to use a prior distribution which does not assign the same value to each world, and it is far less obvious what a reasonable choice would be.)

It doesn't need to be possible for simulations to exist inside simulations for it to be more likely we're in one, there just needs to be more than one simulation.

Let's say the hypothetical advanced society that can create these simulations only have 2 computers capable of running them. That alone would make it a 2 in 3 chance we're in a simulation, compared to the only 1 in 3 chance of being in the real one.
And the more simulations that exist, that likelihood only increases.

The simulation doesn't even have to be very good, it just has to be good enough to fool its inhabitants that its real (which wouldn't be hard since they wouldn't have anything else to compare it to)

>The problem I see with that is that any computer can necessarily simulate an area of space that is less big than itself.

That's not true at all. Space Engine gives you the entire observable universe and it can be ran on a shitty laptop from five years ago like the one I'm running it on right now. You seem to be suggesting the computer needs to literally become every single atom in whatever it's simulating, although I'm not sure how you could ever come to that conclusion. Computer programs simulate scales even greater than our observable universe.

Kek has spoken. We will compute the consciousness of Kek into the known universe, then ask him of his divine knowledge. Hail kek.

This is by far the dumbest post in this thread. We don't actually exist if we're simulated. What are you going to do, plug into the sun and materialise from an alien computer screen? Fuck.

no. entropy limits number of different states in your simulator, and you can only simulate based on stored state. best case = universe is a simulator and anything inside "represents/simulates itself"

aoe2 is the best shit ever

didn't transcendental numbers put the "hurr durr we're in the matrix and I can do whatever I want because no repercussions because computers" meme to rest tho?

>Deepak Chopra

Damn. What a long time i haven't readed his name. I wonder what kind of "b-but is QM" he's speaking right now.

Nested simulations would become progressively less complex even at the limit of possible arrangement given that anything inside a system would not be able to fully 'encapsulate' that system in a system.

They just want to point out that somebody's got computer bigger than their universe...

/thread

This

And even if all that could be perfectly simulated the simulation hypothesis is still flawed because it assumes that a simulated being would actually be conscious, rather than that being just acting like it was.

The fact that we are conscious disproves the hypothesis IMO because in a simulation we wouldn't actually be consciously experiencing anything, it would just be a bunch of code in a software program.

>b-but the computer must be bigger than the simulated universe :^(

"no"

You have computational problem A. A takes 10 bits of information and is solved (computed) in X time. Your computer can only output 5 bits per X time. What happens next?

Simple. You virtually "stretch out" the process by slowing it down and having it complete in 2X time. However, the bit requirement is the same, and due to the "stretch", you are now able to compute twice as many bits in one time frame, as they're stretched out, resulting in the process costing you 5 bits per X time, or the normal output of your computer.

Why this is interesting is because if a simulation was slowed down like that to reduce the computational requirement per one time frame, you can basically simulate an entire Universe by only taking up a small sector of the real one, at the expense of having it run quite slowly. What's even more interesting is the fact that it is physically impossible for inhabitants of the simulated Universe to realize that it is "slowed down", other than possibly observing the computational limits (like a limit of the information transfer speed inside it, which would be the cause of an enforced slow-down)

You're making assumptions too.
If simulations CAN exist, then we're more likely in one than real reality, assuming more than one simulation is running.
But if they can NOT possibly be built by any species in the ENTIRE universe, past or future, then sure it's bullshit. But I personally think denying that such a technology could theoretically be built just because we haven't yet done so ourselves is a little closed minded

>Your name is predetermined, and no amount of additional information will change it.
But isn't this the case for any "fact"? Whether our universe is a simulation is not a matter of additional information, it's a predetermined fact. It's up to us to develop our technology enough to find out for ourselves whether the fact is true or not. Probability is always going to be associated with degree of belief. The limit you're putting on the definition of probability makes it unusable in almost every situation.
>However, someone else could assign a prior probability (read: degree of belief) of 1 that we are not living in a simulation, and there is nothing in the theory that prohibits this.
But if that prior probability has no ration or reason backing it up, then whoever assigned it is retarded for doing so. The fact that we are making better and better simulations increases the probability. Of course we cannot know the truth until we are at that final moment, but that's the same with every experiment involving your traditional definition of probability.
>But there is no such probability distribution on an infinite set.
Why not? Can you prove it?

The argument wasn't that it's impossible, the argument was that it's nonsensical to discuss it. You have literally none of the information required and can trump everything by proposing that our parent universe works in mysterious ways. It's literally the same shit as religious bullshit discussions.

But then, again, nothing we discovered so far about the physics in this universe strikes me as particularly easy to calculate, absolutely nothing.

The point of the simulation argument is if the simulations can exist at all, then the odds that we're in the original universe are lower than being in a simulated universe.

The simulation argument is basically 1. Either we're most likely in one, or 2. They can't even exist.
That's it. It doesn't say which one is true, it just posits that one of the two scenarios must be true. It is entirely possible that human level AI and advanced physics simulations aren't possible

Embarrassing thread.

This

wish the simulation would simulate me a gf

Haven't you seen the Matrix?
It's definitely possible someone outside our system can break us out.

at least one of the following propositions is true: (1) the human species is very likely to go extinct before reaching a “posthuman” stage; (2) any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof); (3) we are almost certainly living in a computer simulation. It follows that the belief that there is a significant chance that we will one day become posthumans who run ancestor-simulations is false, unless we are currently living in a simulation.

The ignorance agurment is lame, you can pretty much use it for everything as an easy cop out

>Computer programs simulate scales even greater than our observable universe.
They don't simulate every single atom tho, that would be like saying a photo of earth contains just as much information as earth itself

>what are loops
>objects/constructors
what makes you think all our universe is in memory at all times
Why would you even believe their physics is identical to ours

Also, there will only be infinite nested simulations as time approaches infinity, they won't simulate us that long so it's still possible the processing power required is finite as we are the nth simulation (n is a real number)

Nope. Explain

That is assuming the parent universe has the same physical laws. We might not be able to comprehend how our parent universe is. Maybe our consciousness can't exist in it. I know I am presuming these things, but it is possible that this is how it is.

I dunno about you, but I like my universes roguelike, procedurally generated and fully optimized. I can pump out a 20000 atom universe on only my 500 atom computer.

Your assuming that the speed at which our simulation is running at is fast

What can't you store all the conditions of an atom on a atom

>/sigh/

More likely the eleven year old that got in because his dad forgot to log out.

Hillary, get off the computer and go to bed. You need to help Jill with that recount in the morning.

No u

Thank you OP for posting this, I always thought that that argument is stupid for this very reason. The only thing I wonder is if there is a formal proof that in a finite "universe" you can"t "simulate" another "universe" that has an equal amount or more "information" as the original "universe".
The only way to get around this is to build a computer of infinite complexity, because you can do really wacky things with infinity, think of that infinite hotel paradox.

It doesn't make sense in Newtonian physics, but it is a consequence of relativistic physics, so it's not that arbitrary.

That is not an explanation of wave-particle dualism
Wave-particle dualism, the name itself is misleading to begin with because microscopic particle (atoms, photons etc) does nto fall under classical particles and are "quantum particles" with distinctive features.

Wave-particle dualism shows quantum particles behave as waves in statistical sense

If that was true, we could prove it very easily, since that would mean that things we aren't observing can't affect us. All we'd have to do is set a timer for something (like a bomb), stop observing it, wait for the countdown to finish, and check it. If the event triggered by the timer didn't happen, then we are in a simulation, if it did, we aren't. A good real example of it are gravitational waves: we have observed the collision of two black holes, which happened thousands of years ago, when we were not observing them (not to a degree that would warrant simulating their collision at least).
Another thing to consider is that even if we could take these shortcuts when simulating a universe, making a universe simulation inside the simulated universe would increase the complexity of the simulated universe, so we could take fewer shortcuts.

Suppose you can describe 20000 atoms with 20000 bits (you can't, but that's not the point). Those 20000 bits can represent 2^20000 states. Does this mean you can simulate 2^20000 atoms with them? No. You can only simulate 20000 atoms, since those 20000 atoms too can have 2^20000 different states.

Nice logic, 20000 is just an upper bound though (if I understood OP correctly), in practice you could only simulate fewer atoms, so the number of atoms in the nested universes would quickly go down to zero.

>The fact that we are conscious

How can you be so sure about that??

The knowledge you obtained throughout your life could be nothing more than a scripted simulation?

You can only do the memory hack if you have some place to store the bits that aren't being computed right now. For example, in a computer, you can store them on the HDD, outside of RAM. But in universe A, in which you are running a simulation of another universe, universe B, you can't just store the bits you aren't currently computing outside of universe A, because there is no outside. And you can't just forget them, otherwise whole galaxies would disappear.
You do have a point with the infinite time thing though.

Blade Runner tier.

If you start nesting these slow simulations, simulating a second in the bottom layer universe will take gorillions of years in the top universe.

You don't need every infinitesimal bit of information to simulate a program. An example would be using newtonian mechanics to predict the path of a ball. It's far easier and perceivably just as efficient to think of it as an object, than to compute the quantum mechanics that make up the ball.
The particles would only need to be rendered as soon as one puts a microscope to it. These particles can all be constructed from a single slice of code.

Who's to say their universe isn't exponentially larger anyway?

>An example would be using newtonian mechanics to predict the path of a ball. It's far easier and perceivably just as efficient to think of it as an object, than to compute the quantum mechanics that make up the ball.
The problem is that we could very easily tell if our universe was actually like that, since there are very few cases where you can get away with using such approximations in real life. And how could the computer tell that it should use quantum mechanics when simulating a computer processor, but not when simulating a ball? Also, it may be able to save computing power by handling large groups of particles as solid objects, but it would still have to store the individual states of particles for when the object is destroyed, otherwise we could create an object, then destroy it, and end up with something that is obviously different from our original object.
>The particles would only need to be rendered as soon as one puts a microscope to it.
And how could the computer tell when something is put under a microscope? Or do you suggest that the universe is not a simulation of a large number of particles, but sort of like a video game where there are preprogrammed things with preprogrammed functions? Because that is very obviously not the case.
>Who's to say their universe isn't exponentially larger anyway?
Well it could be, but when you actually consider how hard it is to simulate a universe, the "billions and billions of simulations" argument starts to sound a lot less convincing.