Is Our Reality a Simulation? Exploring the Simulated Reality Hypothesis

[SubTachyon]

I've been reading book called 'The Big Questions: Physics' by Michael Brooks during my trip to Warsaw. It's an enjoyable collection of essays despite the fact that they were perhaps slightly too basic at times for a physics student like me. What grabbed my attention though was the simulated reality hypothesis (http://en.wikipedia.org/wiki/Simulated_reality).
Up until now I treated the idea of us all being stuck in Matrix spin-off as an entertaining thought but unscientific, untestable and solipsistic. However what sparked my interest is the argument from "conservative computing":
- You never want to waste computing power when programming a new software.
- This means that any simulation will not be infinitely smooth.
- When we get down to the quantum level things get - as we all know - very wierd.
Can it be that the quantum wierdness is simply us getting closer to the very basics of the running simulation? Is the cat both dead and alive until someone measures it a way of saving computing power when nobody and nothing interacts with a particle hence why waste the rendering power?

I won't go very deep into it right away. Surely somebody in here must have heard of this before. How valid is this hypothesis? Is it scientifically sound or just philosophical and untestable hogwash?

Cheers.

 

[mgb_phys]

If this was the matrix - wouldn't the food be better?

 

[harroxelas]

If this was the matrix - wouldn't the food be better?

LOL!

Seriously now, I read an article on this a few days ago and it mentioned a paper by a guy named Nick Bostrom. Googled it. Check it out http://www.simulation-argument.com/"

 

[Pythagorean]

So if this is a "simulation", what's the "real world"? And if this simulation is as complex as it is, how much more rich must be the "real world". The "real world" would then have completely different physics (for example, following from your post, energy wouldn't be quantized, so the "real world" would have infinite energy).

Since the physics are completely different, and we're here experiencing this "simulation" is it really a "simulation" of anything? Or is it it's own "world"?

 

[GeorgCantor]

I've been reading book called 'The Big Questions: Physics' by Michael Brooks during my trip to Warsaw. It's an enjoyable collection of essays despite the fact that they were perhaps slightly too basic at times for a physics student like me. What grabbed my attention though was the simulated reality hypothesis (http://en.wikipedia.org/wiki/Simulated_reality).

While a simulation is consistent with all the evidence, there is something that can wipe the floor with the idea in the end - "I exist".

A character in a simulation doesn't exist, unless he has a mind of his own, which would simply negate the idea of a simulation. If you are not self-aware you cannot be said to exist, so how would a desktop icon be made to exist?

By extension, does a stone exist? It's not self-aware, doesn't have memories, doesn't seem to have a Mind mandatory for personal experience, so except for being part of your personal experience, can a stone really be said to exist?

For the simulated argument to hold, I myself have to be part of the team that developed the software and took a dive into it for a short ride. But seeing the mess that is here, i should have been out by now fixing the injustice.

 

[Ferris_bg]

A character in a simulation doesn't exist, unless he has a mind of his own, which would simply negate the idea of a simulation.

For him - yes, for the creator of the simulation - no. We don't know his purpose, the simulation could be some sort of perpetual motion machine, or some sort of a game.

How can we distinguish a perfect simulated reality from a non-simulated one? We cannot in any way, because we are part of the simulation, only an observer could, so we could try to leave the simulation, but because it's a perfect one, we won't be able to.

 

[SubTachyon]

@GeorgCantor: Why can't you be just very well developed NPC? If you have enough computing power to simulate some sort of universe surely simulating human brain shouldn't be such a problem.

 

[GeorgCantor]

For him - yes, for the creator of the simulation - no.

Well, the thing is that i do know is that I exist. Whether you believe it or not, that is the ONLY thing one can be certain of in this universe. Even the idea of physical death as a final termination of our personal experience isn't 100% certain.

Any theory that has at its root the contention that i don't exist is 100% wrong.

How can we distinguish a perfect simulated reality from a non-simulated one? We cannot in any way, because we are part of the simulation, only an observer could, so we could try to leave the simulation, but because it's a perfect one, we won't be able to.

A perfect simulation cannot simulate me as a self-aware entity. How would you put a self-aware "I" into a programming language? And how would a piece of software code become self-aware? If i am not self-aware than i don't exist, and if i don't exist i have no logical basis to contend that some kind of software might exist in some other plane of existence.

 

[GeorgCantor]

@GeorgCantor: Why can't you be just very well developed NPC? If you have enough computing power to simulate some sort of universe surely simulating human brain shouldn't be such a problem.

Why would someone go all the way to claim that there is no "SubTachyon"? Have you lost your sense that there is a SubTachyon that exists in some form?

A software code is exactly NOTHING without a MIND somewhere along the chain of simulations. If that is what you were implying, why don't you tweak your argument to become that the "simulation" is actually existence in the Mind of God, which would not only be harder to counter but impossible?

 

[heusdens]

Your idea makes absolutely no sense, and you are mixing things up. The "simulator" idea (all reality as perceived is just an imagination or virtualization of some all-encompassing super-reality in which the perceived reality is just a virtual reality) and the Idealistic interpretation of Quantum Mechanics (quantum events only exist when they are perceived through a conscious observer).

The first idea is just an idea, not based on reality, and as far sought as solipsism.
There is no way you can ever test that the idea is false, yet that does not make the idea any sound either.

Science dismisses this idea at the virtue that it makes needless extra assumptions. A "genuine" reality of some sorts exist anyway, so why not assume that the one we currentely perceive is that reality, which with the help of science and instruments can be approaximated and understood better and better, but never with unlimited knowledge.

Moreover the idea does not in any way contribute to a better understanding of the reality, and has the same fallacy as the idea that "God did it" since it places a limit on what we can know and understand about the world, and places all things we currently do not understand into a created being (wether that be the deity or the super reality) of immensely more complexity of which we can not understand anything at all.

Second, the idea that the perceived reality only exists as far as a conscious observer observes it, is just an interpretation of QM and one which is Idealistic in nature.

Moreover, your intermixing of these two things, in which you suggest that the rendering of this virtual reality would be energy/time efficient if non-perceived objects don't consume calculation time/energy, makes no sense, since the same calculations (how these objects which are temporarily not perceived must behave) must be made anyway upon perception (they change the same way, wether they were perceived or not), so there is no gain in calculation speed or efficiency.

 

[Meatbot]

Seems like for something to be a simulation, it must be an approximation of the original such that it lacks the same degree of detail. If it's exactly as detailed then it's a copy or another "original".

Does that not imply that if we are in a simulation, then there must exist a realm with more detail than we can perceive in our approximation? How might that manifest itself? Could we ever see physical evidence of this? Could we get to a point where it's obvious more detail is required, but we just can't see it? Are we there already? Dark matter, perhaps? Can we guess at what the detail might be based on the approximation?

Also, software frequently has glitches. I will concede that a civilization advanced enough to simulate us might have conquered bugs, but maybe not. Would we be able to detect a software bug? Compromises sometimes need to be made and evidence of that could be available. For one thing, I have heard of the idea that the speed of light might be the computing speed of the universe (Wolfram?). It can't calculate any faster than that and that makes it what it is. Perhaps it's really the computing speed of the hardware we're being simulated on.

 

[brainstorm]

I think it is important to note that the only reason the issue of whether "reality" is a simulation or not can even emerge in the first place is when humans are in an aesthetic-evaluative cognitive mode. When you are interacting with aspects of reality directly, the status of what you are interacting with is irrelevant to the task of what you are achieving by interacting with it. E.g. If you're cleaning a floor, what difference does it make if the floor's status is "simulated" or "real," or whether the act of cleaning it is a practical activity or a simulated one? All that matters is the instrumentalism of getting the floor from dirty to clean. To re-phrase my point at a general level: "real" is an aesthetic status that only matters in non-practical/non-instrumental evaluative-reflection. Within the context of any simulation, activities are as real as their effects.

Still, as I say this, I am thinking about the difference between a direct mechanical system and an informatic command-control simulation. E.g. when you clean a floor, the floor becomes clean because you remove the dirt from it physically. In a simulation, it might only become clean because you fulfill the conditions the programmer established for fulfilling the requirements of "cleaning." This could be 10 minutes of mopping with a minimum amount of effort. Once you fulfilled the criteria, bling, the floor is magically clean - only because the simulation was triggered to replace the dirty floor simulation with one of a clean floor. So there would be no direct causality between actions and consequences; i.e. only command-control logics. The simulation is ultimately controlled by the parameters of the program instead of interactions between elements directly affecting each other by virtue of their specific mechanics.

Still, one can imagine that there are states of consciousness in which reality is indistinguishable from simulation. A child who believes in the tooth fairy, for example, cannot distinguish between the causal mechanics of putting a lost tooth under a pillow for money and planting a seed to grow a plant. To the child, both processes are equally natural. Only a critical thinker can reason that there's no physical mechanics that can cause a tooth to be replaced with money under a pillow and that photosynthesis, irrigation, and nutrition causes a seed to grow into a plant and how. A simulation is magic and reality is ultimately mechanically explicable.

 

[Boy@n]

I find it impossible to program / simulate pain and pleasure, even harder to simulate self-awareness and sens of "I am".

How would you make a robot really feel pain, or pleasure? Say someone hits a robot; you could program it to try to evade being hit (to protect own existence), you could even program it to "pretend" it's in pain if hit, but it wouldn't actually feel that pain, it might even "think" it's pleasurable if being hit or even destroyed, depending on code a programmer gave it.

Pain/pleasure are, IMO, very effective tools for self preservation and evolution. Even the simplest of living organisms posses these tools, but really, no idea how anyone could put this into a computer language, ever.

(I am a computer programmer, well, wouldn't say I am genius, but know how computers work.)

 

[Borek]

You don't have to program everything. Set up a world and let it evolve. Accidentally I have just mentioned simulated world of bitozoa in another thread - do we know what they feel? No, but it doesn't matter, what matters is that they get effective at what they do.

 

[brainstorm]

I find it impossible to program / simulate pain and pleasure, even harder to simulate self-awareness and sens of "I am".

How would you make a robot really feel pain, or pleasure? Say someone hits a robot; you could program it to try to evade being hit (to protect own existence), you could even program it to "pretend" it's in pain if hit, but it wouldn't actually feel that pain, it might even "think" it's pleasurable if being hit or even destroyed, depending on code a programmer gave it.

Pain/pleasure are, IMO, very effective tools for self preservation and evolution. Even the simplest of living organisms posses these tools, but really, no idea how anyone could put this into a computer language, ever.

(I am a computer programmer, well, wouldn't say I am genius, but know how computers work.)

The issue isn't whether you can program a robot to feel pain or pleasure. The issue is whether you can program a computer to deliver pain and pleasure as behavioral stimuli to create associative reasoning that has nothing to do with actual mechanics of functional (non-simulated) human life.

For example, when humans used to farm, producing sufficient crops resulted in full bellies while insufficient production resulted in hunger. Pain and pleasure was directly related to the functioning of the farm. Now, if you separate the people from the farm and control the amount of food they get according to, for example, how often they say "please" and "thank you" in a meaningful way each day, then their farm labor has effectively become simulated by social politeness labor.

I guess a more direct simulation would be if they had to play some kind of computer game where they grow food and they would receive food according to how much "simulated" food they produced in the game. Then you could change the parameters of the game to create a bumper crop or a drought-famine and distribute pleasure/pain according to the logic of the simulation.

 

[mugaliens]

Pain/pleasure are, IMO, very effective tools for self preservation and evolution.

Provided they are tied to both self-pleasure and evolution, yes. But both have been subverted for other reasons, and not to very good effect, I'm afraid.

Even the simplest of living organisms posses these tools, but really, no idea how anyone could put this into a computer language, ever.

It would have to include strong cooperative/collaborative issues among the organisms in your simulation, along with serious detrimental effects for lack thereof.

 

[Boy@n]

Got a "bad" idea, instead of using "dead" matter, of which computers are build today, we could perhaps make "organic computers", perhaps build out of smallest living organisms which already have ability to feel pain and pleasure, and use that abilitiy combined with computerized logic to achieve the "need" for self-preservation and evolving? Organic simulation? Sounds familiar? Our very existence? ;)

 

[Meatbot]

...instead of using "dead" matter, of which computers are build today, we could perhaps make "organic computers"...

What's the difference? They're both made of quarks and electrons. I think the term "alive" is not really very useful.

 

[brainstorm]

What makes a simulation a simulation is not what things are made out of physically. It is the "operating system" or organizing logic of interactions and events. The Truman Show movie is a good example of a simulated living environment that involves real people acting according to planned scripts and command-control logics of the director and writers. The defining factor of a social situation being simulated is that the logic of interactions is not governed by functional logic but instead rules and other artificial types of interactions and exchange where participants are prohibited from making choices outside the parameters of the simulation. In the Truman Show, for example, whenever Truman tries to escape or confront the simulated nature of his existence, there are elaborate obfuscation mechanisms and punitive measure designed to discourage him.

 

[Boy@n]

How many fps can our best computers make in games, 200 fps or so max?

How many fps are there in reality, 10^32 fps? Perhaps more?

So, is it even possible to run a computer simulation with so many fps?

Would this be a proof enough, that we are not "living" in a simulation?

( Realized this after posting https://www.physicsforums.com/showthread.php?p=2865413&posted=1#post2865413 )

 

[Blenton]

Do you know how far computers have come in the last decade Boy@n? Moores law will probably remain for a LONG time as chip makers come up with new cleaver ways of cramming more transistors onto the given area of chip.

How many fps can our best computers make in games, 200 fps or so max?

You're thinking of a nice home computer. They are nowhere near the power of supercomputers used to run nuclear simulations for the government and probably in the next decade our home computers will be as powerful as the supercomputers we have today.

That in point since the power of computers seems to be increasing almost exponentially, the point that we're never going to reach the computing power required for real life simulation is not very strong.

 

[Newai]

You're going to need something a wee bit different than a home PC to simulate a set of life experiences. Or even a computer alone, for that matter.

 

[Boy@n]

The power of computers seems to be increasing almost exponentially, the point that we're never going to reach the computing power required for real life simulation is not very strong.

That's because you didn't consider how much computer power would be needed to compute all the frames for all the motion.

For covering all the steps of a single point moving 1cm in 1s (as happens in our reality) computer would have to calculate 6*10^32 FPS.

That calculation alone I'm sure cannot be done with ALL computers on our planet (super+home computers).

Now consider that eveything in our Universe in constantly in motion...

That, for me, is enough of proof that our reality cannot be simulated by computers, even if there would be computers incredibly faster than the ones we got today.

 

[GeorgCantor]

That's because you didn't consider how much computer power would be needed to compute all the frames for all the motion.

For covering all the steps of a single point moving 1cm in 1s (as happens in our reality) computer would have to calculate 6*10^32 FPS.

That calculation alone I'm sure cannot be done with ALL computers on our planet (super+home computers).

Now consider that eveything in our Universe in constantly in motion...

That, for me, is enough of proof that our reality cannot be simulated by computers, even if there would be computers incredibly faster than the ones we got today.

That's a rather strong statement for the idea implied in it.


How many people would have to be simulated? Me, you, 1000 or 6.7 billion? Who decides? There could be a simulation of just MY viewpoint, which wouldn't take up all that many resources. Or of 200 people's viewpoints or any arbitrary number for that matter. Nobody requires that a simulation has to necessarily include 6.7 billion viewpoints of the world or the 20 billion people that ever lived on the planet. The simulation could have started with my birth, or with the birth of anybody. In fact, when you play a primitive sim game on your computer, just 1 viewpoint is simulated, the other viewpoints are merely implied, they don't really exist within the virtual world you are playing in. You imagine they do for even more realism, but they are deterministic, non-aware digital codes that lack a first person perspective.

A simulation of just my or just your personal experience of a world, isn't going to be exceptionally difficult as to render it impossible. Michio Kaku made a similar point that the universe couldn't be simulated in his book "Paralell Universes", but his argument is flawed.

 

[Meatbot]

For covering all the steps of a single point moving 1cm in 1s (as happens in our reality) computer would have to calculate 6*10^32 FPS.

That's one reason you might be able to tell if you were in one. What if it's running slower than you suggest, but fast enough that you can't tell the difference normally? If that's the case and you can manage to construct an experiment sensitive enough, you might see that the particle was skipping over space repeatedly instead of moving "continuously". If you zoom in far enough you might see the stuttering.