We are all computer simulations - Not

[Edtharan]

This is a bit "tongue in cheek", but it does have some grounding as a serious question:

 

In the past few years there has been talk about if computers became so powerful that they could simulate reality, and that this computer power would keep growing, that these computers would produce so many simulations that just looking at it statistically, we are more likely to be in a computer simulation than reality.

 

But, computers are used more for entertainment than simulation. Think about the number of CPU cycles used on computers for entertainment (games, etc) than are used for reality simulations.

 

As more computer time is spent in entertainment, then statistically, we would be more likely to be in a Game than a Reality Simulation.

 

As we don't seem to be in a Game, then it seems that we must be in Reality and not a Simulation.

 

This seems to counter the "we are in a computer simulation as it is the statistically most likely" argument as we are statistically more likely to be in a Game.

 

What other arguments can we come up with, for and against the Reality Simulation proposition?

[alan2here]

'The Matrix'

FTW

[Ndi]

If we are in a computer simulation then our perception of statistics is generated by the computer thus unable to rely on it. You are also assuming that whatever computer runs us has such a concept as simulation/game/whatever.

 

Also, if we are in a computer game, computer reality of in the friggin MSPaint we wouldn't know. We define what we see as reality and it's what reality is to us. If we define it as a game, it's a game to us. It might as well be the screensaver someone left open and it evolved overnight.

 

This is the new look at the old question - what is yellow. We see colors and are taught it's called Yellow. Should we see through someone else's eyes, would it be Magenta? How can you tell what happens in someone's brain?

[Farsight]

I'm very cynical about the idea that we're all some figment of somebody's virtual reality. It's a cop-out that we could never prove. Interestingly, this was given prominence in the BBC Horizon "Time Trip" documentary, which I think is what got me thinking seriously about time. See transcript below.

 

http://www.bbc.co.uk/science/horizon/2003/timetriptrans.shtml

[Spyman]

As we don't seem to be in a Game

Why ? - How can you tell the difference ?

[Edtharan]

If we are in a computer simulation then our perception of statistics is generated by the computer thus unable to rely on it. You are also assuming that whatever computer runs us has such a concept as simulation/game/whatever.

I was mainly thinking Ancestor simulations, rather than Alien (whether extraterrestrial, machine or whatever) so the beings simulating this would be a future human (maybe a couple of hundred years into our future).

 

Interestingly, this was given prominence in the BBC Horizon "Time Trip" documentary

This is also what got me thinking about this.

 

They presented it as so matter of fact that if it was possible to create ancestor simulations, then the statistics say that we are most likely in an ancestor simulation.

 

But, being a computer game developer, I thought about the fact that more and more computers are being used for games. Looking at these statistics, the number of computers running games will far exceed any computers running ancestor simulations so statistically, we are more likely to be in a game than an ancestor simulation.

 

Also, if we are in a computer game, computer reality of in the friggin MSPaint we wouldn't know.

It doesn't matter weather we are in a game or an ancestor simulation. It doesn't matter if we can detect it or not. I was just making a point that using statistic to say that we are more likely to be in an "Ancestor Simulation" than reality does not take into account the other "Realities" we might actually be in.

 

As for being in mspaint:

 

I'm very cynical about the idea that we're all some figment of somebody's virtual reality.

I am too, but we can't be sure (although all these "paper clips" keep turning up all the time, so we might be in something like MSWord ...).

[Ndi]

Statistically we are most likely to be an unobserved life form.

 

If technology goes to a point where we have the power to run such complex simulations then we'll most likely be running other simulations for loads of time before. By the point when someone allocates such power to running a sim, the tech advance would have probably passed us by so we are in a side task. When we could simulate bacteria from the tech point of view we used that power to sim chemical reactions, design engines and other tasks way before bacteria growth become a game.

 

Besides we're really, really boring. To such a civilization we're below ants, much below apes as we can't sim apes. By the time they reached such a level we're insignificant. Who runs ant simulations for thousands of generations with complete physics?

 

I believe we're no fun, eat loads of memory and power so whatever is hard, smelly, sweaty, predictable and requires loads of resources is unlikely to be a game.

 

We could be a sim project of a student who stepped out to lunch. We'll get a TerminateProcess when he gets back and hits a key to return to his Parallel Universe Conglomerate Editor.

[Royston]

They presented it as so matter of fact that if it was possible to create ancestor simulations, then the statistics say that we are most likely in an ancestor simulation.

 

I vaguely remember this episode, but were they actually clear on what they were basing the statistics against i.e it's far more likely we are in a simulation than this all happening by chance ?

 

But, being a computer game developer, I thought about the fact that more and more computers are being used for games. Looking at these statistics, the number of computers running games will far exceed any computers running ancestor simulations so statistically, we are more likely to be in a game than an ancestor simulation.

 

The only problem I see here is you're comparing a 'current' market trend, i.e the popularity of computer games with the likelihood of a superior ancestral simulation. I'm not sure the two are comparable, games have a purpose to entertain, how would you ascertain what the purpose of this so-called simulation is, and why is the reality of the creators of the simulation the same as ours ?

 

 

It doesn't matter weather we are in a game or an ancestor simulation. It doesn't matter if we can detect it or not. I was just making a point that using statistic to say that we are more likely to be in an "Ancestor Simulation" than reality does not take into account the other "Realities" we might actually be in..

 

I'm not entirely sure what you mean by 'other realities we might actually be in' but if there are a multitude of simulations, how would we know anything about these other 'realities' to make them statistically viable i.e we'd need some numbers to go on.

 

I think there's too many unexplained attributes to what's outside this so-called simulation, to start comparing it's validity with attributes inside this so-called simulation. I know you were being tongue in cheek, but I think to tackle such a question with statistics is a bit of a no-go area, both to prove and disprove.

 

If this is a game however, then there's a severe lack of zombies, and as Ndi said, how incredibly boring...'your next quest, is to seek the post office, and settle the gas and electric bill.' I couldn't see this game gaining much popularity.

[Ndi]

I know. We're Sims ver 612!

 

They finally fixed that simspeak but apparently the free will subroutines are still buggy. Heck, judging by MY free time we might be in. Any of you recall stepping into oblivion when exiting the screen?

 

[sorry about the off-topic]

[Edtharan]

The only problem I see here is you're comparing a 'current' market trend, i.e the popularity of computer games with the likelihood of a superior ancestral simulation.

Humanity as nearly always played games as far as we can tell. So I see no reason not to assume that this will continue into the future.

 

I'm not entirely sure what you mean by 'other realities we might actually be in' but if there are a multitude of simulations, how would we know anything about these other 'realities' to make them statistically viable i.e we'd need some numbers to go on.

Well, I am just going off what was said in the "Time Trip" documentary. I have also heard this sentiment expressed in other places.

 

We can not know anything about "other simulations", but we don't need to know about them. We are discussing possibilities. If More's law continues on as it is (even if it slows a bit), or at a time where computers can be the size of a planet, then they will have the processing power to simulate a region of the universe (this doesn't have to be real time either - we would still perceive it as real time).

 

This region, if it as good enough simulation, would eventually be able to simulate its self, and so on. This is only considering 1 computer running an ancestor simulation, imagine if millions were doing it, or billions even.

 

There is no reason we could not create a"universe" simulation that had a type of physics that could produce a computer inside it. In fact this has been done with a very basic simulation called "Conway's Game of Life", they have managed to build a universal computer in it that is capable of running a Conway's Game of Life simulation which it's self would be capable of running a Conway's Game of Life simulation (and so on).

 

Imagine a simulation being created that would develop intelligent life that would create a simulation that would lead to intelligent life that would create a simulation...

 

Now imagine a civilization that spans multiple planets with billions of these simulations systems. There would be an awful lot of simulations going on.

 

But, the point of my post was that there would be much more Games being played than there would "Pure Simulations". There is a thrust in Games to produce better AI that more closely resembles Real Intelligence. So it would be more likely that a game would be running an AI that would be "intelligent" than one of these simulations (as there would be more people playing games).

 

Also games would benefit from a self constant "Physics" system (the rules of the game). We are even starting to do this now with games like "Second Life".

 

In games like second life you can create objects with various properties. Why not have an object that can simulate a computer that could simulate a computer...

 

If we are NPC/AIs complex enough to act with "Freewill" and can interact with the game components, then we would be able to make these objects too.

 

So, we have computer systems able to make simulations of computer systems that are capable of making simulations of computer systems (and so on). However, there would be more Games using this than there would be Pure simulations. Hence there would be more "Game" universe simulated than research simulations and statistically we would be more likely to be in a game simulation than a research simulation.

[Ndi]

Now imagine a civilization that spans multiple planets with billions of these simulations systems. There would be an awful lot of simulations going on.

 

[...]

 

In games like second life you can create objects with various properties. Why not have an object that can simulate a computer that could simulate a computer...

 

a) until some civilization doesn't simulate because they evolved differently.

 

b) because of overheads. In order to run a simulation at 1 GHz you need a 2 GHz host that computes all the physics for you. (numbers are exaggerated, you need billion times more to compute physics down to particles). So the host running it would be running at least 4 (if it's aware of the child) or at least 3+ if not. They don't span forever, and given the fact that we need massive computers for a simple particle - the computations exceed loads of limits.

 

Don't mistake simulation and emulation.

 

Remember - do not mistake the sim that walks around thump up its pants doing nothing. They are a bunch of pixels. We have a spinal chord, hundreds of muscles, interact and think, study particles and experiment with unset rules - someone can't just cheat and emulate the result, we check stuff along the way. You actually need to run every spec of dust. Optimized, yes, but still.

 

We don't currently have the processing power to simulate a single person's one sense, let alone billions with all their senses running. Billions of computers running simulations? Unlikely. There is such a thing as minimum data density.

[Edtharan]

We don't currently have the processing power to simulate a single person's one sense, let alone billions with all their senses running. Billions of computers running simulations? Unlikely. There is such a thing as minimum data density.

But, if More's law holds (every 18 months the computing power doubles) for 180 years, then that means that the computing power would be 2^12 or 4096 times the power we have today. So my 2.0GHz dual core would end up being a 8,192GHz dual core. What would the power be after 1,800 years...

 

As for minimum data density, I did talk about planet sized computers. Also if quantum computing is made workable (they have made the components, it is just scaling them up for large usage - and that is admittedly a significant problem) then the power of the computers will be enormous.

 

Also, even if the simulation was "Saved" and then the user exited it and resumed the simulation at a later date, the occupants of that simulation would not notice that jump in time. Neither would they "see" their universe running slower if the simulation was run slower than real time.

 

Intelligent algorithms would only simulate the important interactions. As each partial interacts with another partial the simulation would then apply the appropriate algorithms, while they are not interacting a simpler and faster algorithm can be run.

 

Finally, the occupants would not be able to know the "outside" physical laws. The laws of the simulation would be the laws of their universe, so this might mean that different forms of computing could operate in the outside universe which makes it easier to simulate a universe.

 

It is estimated that by 2030 (if more's law holds) we should have the capability to simulate an entire human nervous system. Even if we only come close by that time, it will not be long before that capacity is then available.

 

b) because of overheads. In order to run a simulation at 1 GHz you need a 2 GHz host that computes all the physics for you. (numbers are exaggerated, you need billion times more to compute physics down to particles). So the host running it would be running at least 4 (if it's aware of the child) or at least 3+ if not. They don't span forever, and given the fact that we need massive computers for a simple particle - the computations exceed loads of limits.

This is not true. You are assuming that the simulation program must be running at half the computer's capacity. You would not need to have separate processes for the inhabitants and the physics. Just give 100% of available CPU time to running the physics, our brains are made of atoms and they seem to do a good job of allowing the emergence of our brain function.

 

Of course, with programs, precomputed data is always faster to use than calculating the data on the fly. So much of the physics might be just down to a "lookup" table and the amount of processing needed is just for a simple search f this table. Vast memory banks and a simple physics system (simple doesn't mean it can't have a great depth - see Conway's Game of Life again) could allow you to simulate a universe even with a moderate level of computational power (it might even be possible with today's computing power, we would just need those lookup tables).

 

Also, it is not necessary for the simulation to have existed for 13 or so billion years. This could just be a data set loaded into the simulation and only the last second or so is actually being simulated. We could never know.

 

Remember - do not mistake the sim that walks around thump up its pants doing nothing. They are a bunch of pixels. We have a spinal chord, hundreds of muscles, interact and think, study particles and experiment with unset rules - someone can't just cheat and emulate the result, we check stuff along the way. You actually need to run every spec of dust. Optimized, yes, but still.

No, that sim is not just a bunch of pixels. It is in fact the result of a process. We just see the pixels as a result of part of that process. We are a process, we are the process of the physics and matter configurations that they are applied to.

 

Don't mistake what you see on a computer screen as to what is really going on inside the computer. Having a spinal chord, from a physics point of view, is no different than a lump of rock. The matter configuration is different, but the physics are not.

 

a) until some civilization doesn't simulate because they evolved differently.

But if only one civilization (the first I suppose) decides to attempt a simulation where the inhabitants attempt to set up a simulation, then you will get this recursion. 1 counter simulation (where they choose not to create further simulation) will not break the chain, but 1 civilization that does simulate will create the chain.

[Spyman]

I would say that the trend of games with better and better computers is to simulate more details in the games reality better. If we would be in such a supercomputer simulation, then it would most likely contain plenty of details to simulate the reality of the game very good.

 

The universe could very well just be a small detail, like the pattern on a wall in "Resident Evil", (someone mentioned zombies), that flashes by for a short duration of the hole game. The quest/purpose of the game could be very entertaining and popular for players even if the "lives" of the wallpaper on a wall is boring.

 

If we are SIMs then we don't know our meaning in this simulation or the purpose of it, hence we can't know wether we are in a reality simulation or a simulation in a game.

 

So the argument: "statistically we would be more likely to be in a game simulation than a research simulation" - Doesn't show: "that we must be in Reality and not a Simulation".

 

Unless if you somehow can prove that we not are in a game simulation...

[bascule]

Some transhumanists believe it's statistically likely we're all a computer simulation. I don't buy it.

 

That's not to say I don't believe the universe is really just a computer, just not one assembled and programmed by intelligent systems.

[Edtharan]

Unless if you somehow can prove that we not are in a game simulation...

Well games have a common theme to them. That is they are "games".

 

A very simple definition of a game is: A system where 2 or more players compete with each other within a set of rules to reach a goal (Note: one of the competitors can be the computer or game designer).

 

Now, for a competition to be fair (and although don't have to be fair, unfair games are not as fun as fair games) the rules should work out as a 0 sum game. That is, what one looses the other gains.

 

You can have non 0 sum games, and these tend to be games more like The Sims series (sim city, the sims, etc). However, even these games have a pay-off/penalty aspect to them.

 

These could be loosely interpreted into the conservation of Energy/Matter laws in our universe, but this is a very loose interpretation of them.

 

So, in answer to your question: It might be possible to determine that we are in a game as a game has certain aspects to it that mean that it must be different from a straight "Reality" simulation. These differences are the rules of the game, the goals of the game and the fact that there are Players.

[Ndi]

A very simple definition of a game is: A system where 2 or more players compete with each other within a set of rules to reach a goal (Note: one of the competitors can be the computer or game designer).

 

Human/nature. Me and you. That's not really an argument.

 

You say that a game can be with or without goal, with or without balance, thus we are in a game?

 

Games do nothing more than to replicate reality as we perceive it - as the familiarity does wonders. No game is based on nothing. This is a looping argument as everything we have is based on out observations and experience.

 

It's like saying that we create a simulation of life in a 2-d environment so the 2D humans have 2D games, thus they are in a game. They have no idea where they are, nor can they find out because they don't know where they are nor do they understand what is a game to us.

 

By my definition, a game is designed to be entertaining ("GAME"). This is why is don't like people who oppose single player cheats, but that's a different issue. If it was meant to be as realistic as possible, it's a sim. If it was supposed to be fun, it's a game. How we know what the intended purpose of this computer was is beyond me and I suspect beyond everyone as we have nothing to compare to.

[Spyman]

It might be possible to determine that we are in a game as a game has certain aspects to it that mean that it must be different from a straight "Reality" simulation. These differences are the rules of the game, the goals of the game and the fact that there are Players.

It might be possible but it might also be impossible...

 

To check wether the simulation is different from a real "Reality" or a game "Reality" one must know the real "Reality", which we are most likely not able to know from inside.

 

The Players doesn't have to be in the game and even if they are we could still be some small detail - they may not be Humans or even on Earth.

 

Either way, my main point was:

 

We have not yet determined wether we would be in a reality simulation or in a game simulation.

 

So it can not be used as an argument to prove wether we are in a simulation or in a real word.

[Edtharan]

So it can not be used as an argument to prove wether we are in a simulation or in a real word.

My main argument is not really about whether we are in the "Real World" or a simulation. It is about whether we are more likely to be in a Game Simulation or a Research/Pure Simulation.

 

My second argument (really a question), is: Is it possible, to determine if we are in a simulation?

 

There is only 1 assumption I am making: That if we are in a simulation, it is being run by humans (although what would be our future). The reasons for this is that if aliens were creating and running simulations, then the chances that the simulation would be of us is remote. So looking statistically, we are more likely to resemble the beings simulating us.

[Ndi]

We could be an unobserved side-effect of a simulation of Earth of Universe. It's much more interesting and useful to run a high-speed simulation of the Universe to predict changes and dynamics.

[Spyman]

My main argument is not really about whether we are in the "Real World" or a simulation. It is about whether we are more likely to be in a Game Simulation or a Research/Pure Simulation.

Guess I misinterpreted your first post then...

 

My second argument (really a question), is: Is it possible, to determine if we are in a simulation?

I think it's only possible if the system running the simulation allows itself to be detected.

 

There is only 1 assumption I am making: That if we are in a simulation, it is being run by humans (although what would be our future). The reasons for this is that if aliens were creating and running simulations, then the chances that the simulation would be of us is remote. So looking statistically, we are more likely to resemble the beings simulating us.

I don't agree with your assumption because:

 

1) It severely cut's down on possibilities and focus very tightly to only Earth and humans.

The Universe is very large, so to assume that we are the only ones able to make such computer simulations and that all those simulations only simulates Earth and humans, is in my opinion misleading.

 

2) I have plenty of games without anything even remotely resembling humans.

Not to mention all neutral games which any intelligent being could play or games which includes both humans and fantasy creatures/aliens. Any statistics on all our games to back up your reason ?

[Edtharan]

I think it's only possible if the system running the simulation allows itself to be detected.

Well, this would mean either a "perfect" simulation, which, unless you have massive (and by massive I mean being able to simulate a volume of space in less than that volume of space - which would be physically impossible - as far as we know) amounts of computing power, it would not be feasible (you could run it at less than real time, but then you wouldn't get the high number of simulations).

 

Other than that, you would need the simulating computer to have the ability to detect a self referential inquiry in the computer (ie something in the simulation tries to determine that it is in a simulation). This would be very difficult to do and would take up more computational time (much more), which would again reduce the total number of simulations. Also, the fact that we are discussing this and nothing has stopped us, indicates that, if we are in a simulation, then this is not the strategy that "they" are using to stop us finding out...

 

I don't agree with your assumption because:

 

1) It severely cut's down on possibilities and focus very tightly to only Earth and humans.

The Universe is very large, so to assume that we are the only ones able to make such computer simulations and that all those simulations only simulates Earth and humans, is in my opinion misleading.

 

2) I have plenty of games without anything even remotely resembling humans.

Not to mention all neutral games which any intelligent being could play or games which includes both humans and fantasy creatures/aliens. Any statistics on all our games to back up your reason ?

Ok, yes, it is an assumption and they are valid points. However, I will try to defend the assumption still:

 

1) If we open up the possibility that we could be simulated as part of an Alien Simulation (rather than an ancestor simulation), then this massively increases the potential that we might be in a simulated reality. Also it closes the door on whether or not we can determine that we are in a simulation.

 

Although, I still think that any "Intelligence" that could develop technology to a point where this kind of simulation is possible, and decide to make these simulations would have certain mental requirements. Namely that of "playing". Playing is not just the kind of "play" we associate with children, but our adult play stems from that period of our development.

 

Playing is essential to invention. Invention comes about through trial and error. This is a kind of play (and play is just a method of learning). It is through play that we have developed the science and technology we have today.

 

It might be possible that "Play" is not the only way to develop technology, but other forms would most likely be less efficient. Therefore, these Alien races would be far "older" than an equivalent race that used "play" and so these "non-Play" races would not have the time to develop the simulations before the "Playing" races.

 

2) This is the strongest point against my assumption.

 

However, over the years the capability of AI to imitate a living human has been increasing. We prefer to play against opponents that are comparable to us in ability (if only for the challenge). Also, because we are limited to our own psyches, we will tend for these opponents to resemble our own thought patterns. This makes the programmers more likely to use their own minds as templates for the AIs.

 

Admittedly, this is a weak argument for my assumption, but it does deserve some examination.

[Spyman]

Well, this would mean either a "perfect" simulation, which, unless you have massive (and by massive I mean being able to simulate a volume of space in less than that volume of space - which would be physically impossible - as far as we know) amounts of computing power, it would not be feasible (you could run it at less than real time, but then you wouldn't get the high number of simulations).

If we are in a simulation we don't know what a "perfect" simulation is, how massive the systems hardware is, how big space it's placed in, what real time is and so on...

 

Other than that, you would need the simulating computer to have the ability to detect a self referential inquiry in the computer

I did not mean that the system would hade any "agents" running to prevent us from detecting the system.

 

What I meant was that the system (both hardware & software) need to be made in such a way that detection is possible either by decision, no decision or by a flaw/bug/limit in the construction/program.

 

If we can detect the system then the system allows us to detect it, otherwise it don't, whether it's by the "creators" purpose or not.

 

Also, the fact that we are discussing this and nothing has stopped us, indicates that, if we are in a simulation, then this is not the strategy that "they" are using to stop us finding out...

We can discuss if we are in a simulation all we want but so far we can't "detect" the system, we don't even know how to.

 

If we open up the possibility that we could be simulated as part of an Alien Simulation (rather than an ancestor simulation), then this massively increases the potential that we might be in a simulated reality.

It also increases the potential of games simulations.

 

Also it closes the door on whether or not we can determine that we are in a simulation.

I don't think so, it depends on the system.

 

Although, I still think that any "Intelligence" that could develop technology to a point where this kind of simulation is possible, and decide to make these simulations would have certain mental requirements.

Speculations - we don't know which mental requirements is necessary and those creatures would likely evolve with such a technology to something we don't know either. Even future humans might be considered aliens to us.

 

Therefore, these Alien races would be far "older" than an equivalent race that used "play" and so these "non-Play" races would not have the time to develop the simulations before the "Playing" races.

(Counter speculation)

But these "playing" races got bored pretty fast and invented some new funnier technology and later on these "non-play" developed the simulations and since they don't play so much and develop slower they spend much much more of their resources on the simulations on a much much longer time scale.

 

We prefer to play against opponents that are comparable to us in ability (if only for the challenge).

But as I said before - we don't know the goal/rules/intention of the game. If the simulation is of the universe, or at least a very large part, then our solar system is a very small "brick" in a remote place, a very detailed piece in the backround that might be "played" but will likely be left alone for a long time, maybe for the whole duration of the game.

 

The players could play against opponents that are comparable to them in ability, but if we are the SIMs then we can't compare us to them and don't know their ability.

 

Also, because we are limited to our own psyches, we will tend for these opponents to resemble our own thought patterns. This makes the programmers more likely to use their own minds as templates for the AIs.

If the simulation is "only" of Earth then it would have needed as many programmers as we have individuals ? It would take a very complicated template with lots of parameters to be able to simulate different persons.

 

With such large computer power I think it's more likely that the programmers let the AIs evolve according to some rules in the simulation. If it's a science simulation it's probably necessary and if it's a game simulation it would be much more unpredictable and fun to play, from game to game.

[Edtharan]

If we are in a simulation we don't know what a "perfect" simulation is, how massive the systems hardware is, how big space it's placed in, what real time is and so on...

you misunderstood what I meant by "perfect". By "Perfect" I meant a simulation without any flaws that could be detected and exploited by the inhabitants of that simulation. Such flaws would show up as defects in the mathematics that underpin physics. That is, even though we had an exact match for the physics, there would still be situations where it would give an incorrect answer (essentially a logic bug in the simulation opposed to a syntax bug).

 

These logic bugs might stem from a "quick and Dirty" method of implementing the rules of the simulation. We use these kinds of fast approximations in simulations all the time.

 

If you want to simulate the behaviour of a bunch of atoms, then you either need to use at least as many atoms, make fast approximations, or take longer to simulate the atoms than it would to actually have those atoms do whatever it is you are interested in.

 

So either the size of the computer is larger than our universe, our universe is running slower than the creator's universe, or they are using fast approximations.

 

If they are using fast approximations, then there will be "artefacts" that can be detected. If either of the other systems are being used, then we can't detect it, but they would not lead to a massive amount of recursive simulations (which makes them less likely a place we will be in - as there will be far less of them).

 

I did not mean that the system would hade any "agents" running to prevent us from detecting the system.

 

What I meant was that the system (both hardware & software) need to be made in such a way that detection is possible either by decision, no decision or by a flaw/bug/limit in the construction/program.

 

If we can detect the system then the system allows us to detect it, otherwise it don't, whether it's by the "creators" purpose or not.

"Agents" (like in the matrix movies) are not needed. All the system needs to do is monitor any intelligence that arises due to the simulation (or if that intelligence is programmed in from the start it makes it much easier) and determine if they are attempting to uncover flaws (as I stated above) in the system. If you have a language translation program, you could eavesdrop in on any conversation (and if it is in a computer - like we are doing - it also makes it easier to detect) and determine if the inhabitants are attempting to determine that they are in a simulation.

 

It would then be a simple matter of "adding in" something that would stop that line of thought. A heart attack, lightning strike, just erasing them (including from all memory of the inhabitants), switching off the simulation, etc.

 

As this hasn't been done, and we are asking these questions, then either we are not in a simulation, or they don't care if we find out that we are in a simulation.

 

We can discuss if we are in a simulation all we want but so far we can't "detect" the system, we don't even know how to.

As I said, we are more likely to be in a simulation that used fast approximations than one that doesn't, therefore there will be flaws that can be detected (may be the whole Quantum Gravity issue is one of these flaws, or maybe the Dark Matter is the result of these fast approximations creating an error that we are detecting).

 

Yes, we can determine we are in a simulation if the simulation is using fast approximations, and we are more likely to be in a simulation that uses fast approximations.

 

It also increases the potential of games simulations.

True, and this supports my original proposition, that is we are in a simulation, it is more likely a game than not.

 

Speculations - we don't know which mental requirements is necessary and those creatures would likely evolve with such a technology to something we don't know either. Even future humans might be considered aliens to us.

Yes, this is speculation, but I think it is a fairly solid assumption.

 

Evolution will mentally equip a species for survival.

 

If ther eis a species that lives in a rapidly changing environment (on the time scale over a few generations - like the onset or end of an ice age, etc), then the ability to rapidly adapt to these changes will be a survival trait.

 

Learning and passing that learning onto the next generation (aka culture) will also be a big survival advantage. So in these situations, it would be expected that the ability to learn (and make inferences about related subjects - aka intelligence) would evolve. It wouldn't evolve every time, but it would be a likely path once the initial potential is there (some basic forms of learning can culture are there, the ability to manipulate their environment, etc ). Also, it might take several such highly variable environments to push a species far enough up the intelligence scale that they are to a point that they become technological. But once the species hits it, it is a massive survival advantage.

 

One of the fastest way to learn about the environment is to experiment, trial and error. This can only really exist in a safe environment. In an environment where a single mistake will kill the experimenter, then this kind of exploration can not survive long (and so will not evolve). However, if you have a safe environment, a "Nest" so to speak, then this kind of learning will flourish. "Nests" require the adults to protect the young, so this is now the seeds of culture.

 

Culture is also a fast way to learn, the individuals that have learnt from their experiences pass it on to the young. However, in a variable environment, this could lead to a stagnation of the knowledge and thus would not be a survival advantage.

 

In a constant environment, the overheads of having all this (large brains and groups) because a disadvantage. So evolution would not encourage it's development. Also, there would be no drive to learn new things, so they would not necessarily develop further.

 

The only way to avoid this stagnation is through experimentation, that is play. Play becomes a necessary aspect of survival.

 

What all this indicates is that Play is a necessary (or at least an extremely likely component) part of developing a species to a technological stage. Also, that a species might develop without play, but they would do so at a much slower rate than one that did employ play.

 

(Counter speculation)

But these "playing" races got bored pretty fast and invented some new funnier technology and later on these "non-play" developed the simulations and since they don't play so much and develop slower they spend much much more of their resources on the simulations on a much much longer time scale.

So the playing races would push their technology further and further at a faster rate than the non playing races. That means that the computing power of the playing races will be far in excess of the non playing races. So give a few thousands years of development time, the playing races would have much greater computational capacity and even if they only spent a small amount of time and resources on such simulations, they would still exceed the non playing races in the number of simulated worlds.

 

But as I said before - we don't know the goal/rules/intention of the game. If the simulation is of the universe, or at least a very large part, then our solar system is a very small "brick" in a remote place, a very detailed piece in the backround that might be "played" but will likely be left alone for a long time, maybe for the whole duration of the game.

If we are just a "Brick" in an out of the way place in a "Game" simulation, then why spend so much processing power on us? It would be more efficient to just have a very crude fast approximation algorithm handle us until it was "used" by a player.

 

So either we are not in a game simulation, or we have players all around us. As games seem to be the most common form of simulation (and therefore we are more likely to be in one if we are in a simulation), then we should have players all around us.

 

If the simulation is "only" of Earth then it would have needed as many programmers as we have individuals ? It would take a very complicated template with lots of parameters to be able to simulate different persons.

What is your reasoning here? Why must the number of people being simulated match the number of programmers.

 

The "template" might have been developed in another simulation altogether, it might have be developed over several generations of the creators. They might have created a program to develop such templates and the number of programmers would then be irrelevant as compared to the number of people being simulated.

[Spyman]

This will be a looong reply, so I will cut down the quotes and reorder them little.

 

you misunderstood what I meant by "perfect". By "Perfect" I meant a simulation without any flaws that could be detected and exploited by the inhabitants of that simulation.

What I meant was that the system (both hardware & software) need to be made in such a way that detection is possible either by decision, no decision or by a flaw/bug/limit in the construction/program.

No misunderstanding, and I think we agree on that part.

 

"Agents" (like in the matrix movies) are not needed.

I did not mean that the system would hade any "agents" running to prevent us from detecting the system.

I think we agree on this also.

 

As this hasn't been done, and we are asking these questions, then either we are not in a simulation, or they don't care if we find out that we are in a simulation.

Or we are allowed to discuss it but not to find out the truth.

 

Or they do care but have made a flaw that eventually will let us find out.

(Which will force them to terminate the simulation. )

 

Agreement on the options too.

 

Such flaws would show up as defects in the mathematics that underpin physics. That is, even though we had an exact match for the physics, there would still be situations where it would give an incorrect answer (essentially a logic bug in the simulation opposed to a syntax bug).

Yes, the flaws would show up...

 

But how do we know/prove that they are flaws of reality and not flaws of our theories ?

 

You meantion this yourself here:

(may be the whole Quantum Gravity issue is one of these flaws, or maybe the Dark Matter is the result of these fast approximations creating an error that we are detecting)

 

And another one, that "pops up" in my mind, is the Heisenberg uncertainty principle - the simulation don't calculate everything exactly until we force it to, with our measurements.

 

If you want to simulate the behaviour of a bunch of atoms, then you either need to use at least as many atoms, make fast approximations, or take longer to simulate the atoms than it would to actually have those atoms do whatever it is you are interested in.

Here you are assuming that we are in a reality simulation.

 

If we are in a simulation we don't know what a real reality is, physics as we know it might be different, especially if we are in a game.

 

So either the size of the computer is larger than our universe, our universe is running slower than the creator's universe, or they are using fast approximations.

If we are in a simulation we don't know what a "perfect" simulation is, how massive the systems hardware is, how big space it's placed in, what real time is and so on...

If we are in a simulation then the size of the computer is larger than our universe.

 

Then we are virtual and inside it, we don't know what a real reality is.

 

The Universe from start to now could be only an eyeblink for them.

 

If they are using fast approximations, then there will be "artefacts" that can be detected. If either of the other systems are being used, then we can't detect it, but they would not lead to a massive amount of recursive simulations (which makes them less likely a place we will be in - as there will be far less of them).

Better simulation = more computing power -> less computers.

 

Few "super" computers - propability of science simulation goes up, propability of us inside simulation goes down.

 

Plenty "super" computers - propability of pleasure simulation goes up, propability of us inside goes up too.

 

But what is the propability of massive amounts of "super" computers and their capacitivity ?

 

What we would call a scientific reality simulation today might be much less than what they use as fast approximations in the future.

 

If we are in a simulation our reality could be a fast approximation in the real world.

 

Yes, this is speculation, but I think it is a fairly solid assumption.

On Earth, up to today, yes. In other environments, it could be different.

 

And if we are SIMs in a "cheap" game, how solid would it be then ?

 

But I can agree with a fairly solid assumption.

 

So the playing races would push their technology further and further at a faster rate than the non playing races. That means that the computing power of the playing races will be far in excess of the non playing races. So give a few thousands years of development time, the playing races would have much greater computational capacity and even if they only spent a small amount of time and resources on such simulations, they would still exceed the non playing races in the number of simulated worlds.

So if our own development of computers would have stopped before windows, which world would spend the most CPU-cyckles on non-windows, (DOS), games ?

 

Why would they spend resources to upgrade old programs to fit their new hardware or repair their old computers if they think it's "boring" ?

 

In the far future computers might be a rarity for collectors, they might have something completely else rendering computers useless.

 

If we are just a "Brick" in an out of the way place in a "Game" simulation, then why spend so much processing power on us? It would be more efficient to just have a very crude fast approximation algorithm handle us until it was "used" by a player.

If we are inside a pleasure simulation in one of a massive amount of computers which all have plenty of processing power then probably the most detailed would be the most popular, wether the details are necessary or not.

 

And once more: from inside we don't know what "a very crude fast approximation algorithm" is on outside.

 

So either we are not in a game simulation, or we have players all around us. As games seem to be the most common form of simulation (and therefore we are more likely to be in one if we are in a simulation), then we should have players all around us.

As you may have noted I have changed the word "game" to "pleasure" which would also include movies, with good enough computers movies could also be simulated.

(No players)

 

And like mentioned before:

we don't know the goal/rules/intention of the game

The players might be in the game or not, even if they are in they might not be on Earth, and if on Earth we might not be able to tell.

 

What is your reasoning here?

Individuality:

It would take a very complicated template with lots of parameters to be able to simulate different persons.

 

The "template" might have been developed in another simulation altogether, it might have be developed over several generations of the creators. They might have created a program to develop such templates and the number of programmers would then be irrelevant as compared to the number of people being simulated.

More or less what I said:

With such large computer power I think it's more likely that the programmers let the AIs evolve according to some rules in the simulation. If it's a science simulation it's probably necessary and if it's a game simulation it would be much more unpredictable and fun to play, from game to game.

 

 

In my opinion, there is to many speculations here, which all greatly favors it to be a "game" with approximations of a real world.

 

That doesn't meant that the speculations or the conclusion is wrong, but I don't agree with the reasoning.

 

And you are still repeatedly pointing out that since there is no players this can not be a simulation, which you said was not your main argument.

 

If we are in a simulation and we are more likely to be in a game simulation where it's more probably to be players and still there are no players, then your reasoning is proved wrong.

 

To use the reasoning the other way around, to prove that we are not in a simulation, is not good enough. I think you would need a lot of very solid measurable physical proof and we are not able to measure anything real from inside of a simulation.

 

I don't think it's possible to prove that we are in a real world and not in an simulation. It might be possible to prove that we are in a simulation, depending on the system, but it will be really hard. To prove the purpose of the simulation will be much much harder.

[Edtharan]

Or we are allowed to discuss it but not to find out the truth.

 

Or they do care but have made a flaw that eventually will let us find out.

(Which will force them to terminate the simulation. )

Yes, these scenarios are quite probable too.

 

As far as we can tell, these "Simulation End" situations have equal probability of occurring. If the creators have an ethical code like ours, then they would see the ending of such a simulation as potentially being tantamount to mass murder on a Universe wide scale (they could see us as thinking, feeling beings).

 

So, if the creators have a similar moral and ethical code, then we can assume that they would not be completely arbitrary. Either they would intervene and prevent a situation that ended in the shut down of the simulation, or they would not shut down the simulation in case of the discovery of it.

 

If we now consider that the creator have other ethical positions about the shut-down of the simulation, then as far as we are concerned this decision will be completely arbitrary and the discovery or not of the universe being a simulation does not come into it (they could just choose to shut it down because they want to make lunch).

 

Yes, the flaws would show up...

 

But how do we know/prove that they are flaws of reality and not flaws of our theories ?

And another one, that "pops up" in my mind, is the Heisenberg uncertainty principle - the simulation don't calculate everything exactly until we force it to, with our measurements.

I didn't think of that, but it could be. Also the randomness associated with the Uncertainty Principle and Quantum Mechanics could also be a part of this fast approximation.

 

One way that might lead to the discovery of our reality being a simulation is to look at what might be the fast approximations. In a computer simulation (that we make) we sometimes use fast approximations. However, by looking at the way computation works we might be able to determine what the fast approximation is supposed to simulate.

 

What we need to do is look for ways that you could simulate a system, but using more computationally complex methods (and produce an equivalent result). This would be the reverse of algorithmic optimizations.

 

This is of course a much more difficult problem than the one of optimization, but it could be done (and would show up in the creator's system if they were looking for it).

 

Here you are assuming that we are in a reality simulation.

 

If we are in a simulation we don't know what a real reality is, physics as we know it might be different, especially if we are in a game.

Yes, but one artefact that would occur in a simulated universe is that they would be computable. That is, a simulated universe would contain the ability to perfectly simulate its self. If we can determine that our Universe is computable, then it strengthens the chance that our universe is a simulation.

 

This occurs in simulations that are not designed even to accurately simulate our universe (Conway's Game of Life for instance).

 

Our universe might be the top level of the recursive simulations, but statistically this would be unlikely.

 

If we are in a simulation then the size of the computer is larger than our universe.

What I meant is that if you are using a computer to simulate a number of particles, then the computer you are using must consist of more particles (also accounting for the interface and other necessary components) than you are simulating, or you will need to run it at less than real time. Simulations that are running at less than real time will not produce as many simulations as one that is running at real time (or faster). Unless they are using lots of fast approximations, which further increases our chance of finding flaws in the simulation.

 

If we are in a simulation we don't know what a real reality is, physics as we know it might be different, especially if we are in a game.

Yes, the "Laws" of physics may be different, but if they are using computation, then it will conform to the rules that govern computations (namely the Universal Turning Machine), which state that a Universal Turning Machine (a computer) can simulate any other Turing Machine even if it running a different "operating system).

 

If we are in a computer simulation, then our universe is run by a Turing Machine, there fore a Universal Turing Machine can simulate the universe.

 

We don't need to know what the physics of the creator's universe is (they may even be in a simulation themselves), all we need to know is that if we are in a simulation, we must be able to simulate the computer that is simulating us (although at a much slower speed than it is running at).

 

Better simulation = more computing power -> less computers.

I disagree with this. Take a look at what has occurred over the last decade. Computing power has massively increased and we have more computers. It seems as if More Computing power -> More computers.

 

Few "super" computers - propability of science simulation goes up, propability of us inside simulation goes down.

 

Plenty "super" computers - propability of pleasure simulation goes up, propability of us inside goes up too.

 

But what is the propability of massive amounts of "super" computers and their capacitivity ?

As computers can be used to design and manufacture computers and make improvements to them, this means that as the power of computers increases, the availability of computers will increase. So what would be a "super Computer" only a few years ago, can become the Typical computer several years later. So this means that as the power of computers increase, the more computers will exist.

 

A few "Super" computers will become Lots of super computers, and then a few UberComputer will be created which will lead to lots of Ubercomputers and so on...

 

So it might start off with a few science simulations, but the super computers will create a massive amount of non science computers and the total processing power of these non science computers will exceed the amount in the super computers. Also, there will be more computers running simulations (pleasure simulations) than the number in of science simulations.

 

So even if there is a few "Super" computers, this will lead to lots of non science computers and non science simulations (pleasure simulations).

 

So if our own development of computers would have stopped before windows, which world would spend the most CPU-cyckles on non-windows, (DOS), games ?

these universes would not have developed much in the way of simulated universes, so it si unlikely that we are on such branches.

 

Why would they spend resources to upgrade old programs to fit their new hardware or repair their old computers if they think it's "boring" ?

Again, this is an unlikely scenario (for a "Playing' race) and so would not contribute much to the probability that we are on such a branch.

 

In the far future computers might be a rarity for collectors, they might have something completely else rendering computers useless.

Yes, computers might become a rarity, but unlikely due to the fact that they are a tool and we are tool using (technological) species.

 

Also this situation would need to occur in all universe (if their is at least 2 simulated universes then it negates this argument). Any simulated universe that chose this path would not simulates many universe on computers. So the number of branches that would contain recursive simulations from this line would not be large and we have less probability of being on one of these branches.

 

If we are inside a pleasure simulation in one of a massive amount of computers which all have plenty of processing power then probably the most detailed would be the most popular, wether the details are necessary or not.

 

And once more: from inside we don't know what "a very crude fast approximation algorithm" is on outside.

But efficiency (needed if you are going to have a lot of users) would dictate that that you only spend processing time on detailed that are actually needed.

 

Fast approximations can be detected as I explained above, If we are in a computer simulation, then the simulation will be computable. Therefore there will be algorithms. If we look for algorithmic optimizations and the fast approximations that such algorithms would by necessity use, then it might be possible to detect that we are in a simulation.

 

As you may have noted I have changed the word "game" to "pleasure" which would also include movies, with good enough computers movies could also be simulated.

(No players)

Good point. The Players need not be visible to us at all. But, even if it was a simulated move, then that means that this region is being watched (even if the players are not visible, their presence can still be inferred - assuming we are in a simulation).

 

And I agree that Game does not cover all of these kinds of simulation, however I will put forward "Entertainment" simulations as opposed to "Pleasure" simulations as it is a broader category that includes Pleasure simulations.

 

Also, as it is more likely that a playing Race will develop the technology to create simulations, we are more likely to be in an Game form of entertainment simulation as there will be more instances of a game type simulation than an Movie type simulation as a movie type simulation can be broad cast to many watchers with just the one instance.

 

The players might be in the game or not, even if they are in they might not be on Earth, and if on Earth we might not be able to tell.

We might not be able to directly tell if there are observers, but we can infer their existence (if we are in a simulation) by the fact that a simulated reality would be computationally expensive and so optimizations would need to be made and Level Of Detail (LOD) is an easy and obvious optimization.

 

This means that any region of the simulation that has a high LOD will be one that is actively observed by the creators.

 

If the creators are not using optimizations, this then limits the number of simulations that can be run on their computers. This means there are less "daughter" simulations under them and make it less likely that we would be in that branch.

 

So we are mor likely to be in a simulation with optimizations than not.

 

Individuality:

I have had some experience with AI systems (basically as a hobby, not as a research position), individuality does not need complex templates. A learning system which is exposed to different stimulus during learning will show individuality.

 

In an experiment I did, I wrote a program that avoided object on the screen. The "Agent" would move across the screen and if it encountered an obstacle it would "look" in a random direction and see if it was open. If there was another obstacle there it would try another random direction.

 

If the space was open, then it would remember that and next time it encountered an obstacle it would first try that solution.

 

Even such a simple system as this demonstrated a form of individuality (and some people said they had a kind of personality) as different "agents" would develop different responses to objects. Some would be "aggressive" and move so they would just miss the obstacle, others would back up and appear to be fearful. It all depended on the history of the "Agent" and what it had found worked in the past.

 

So even simple systems can demonstrate individuality, complex templates are not needed.

 

In my opinion, there is to many speculations here, which all greatly favors it to be a "game" with approximations of a real world.

 

That doesn't meant that the speculations or the conclusion is wrong, but I don't agree with the reasoning.

As i said in my first post, this discussion is not to be taken too seriously. It is supposed to be just pure speculation.

 

I agree that different conclusions can be reached, and that it might be impossible to determine whether we are in a simulation or not. But it is interesting to speculate about it.

 

One idea I would like to put forward (in case we are in a simulation and anyone is listening in on this) is:

 

A technological race that is capable of simulating an entire universe with intelligent, self-aware beings in it, should have the capability of running a smaller simulation of just a single individual that would reside within that larger simulation.

 

They would also be able to connect various external components into that individual simulation and allow direct access to "The Real World" (through digital cameras, microphones and such).

 

There would be willing individuals within this simulation (me for one) who would like to know the truth of the matter and these individuals could be put into such a "machine" and direct communication could take place and this issue could be resolved.

 

And you are still repeatedly pointing out that since there is no players this can not be a simulation, which you said was not your main argument.

It is not my main argument, yes, but it does have relevance to it. If we are in a simulation and there are no players, then we are more likely to be in a research simulation that a an Entertainment simulation. As one of my points was to determine whether we are in an Entertainment simulation or a research simulation it does have relevance.

 

I don't think it's possible to prove that we are in a real world and not in an simulation. It might be possible to prove that we are in a simulation, depending on the system, but it will be really hard. To prove the purpose of the simulation will be much much harder.

Again, this is not supposed to be a serious discussion and come to a definitive answer (neither is it just about randomly spouting beliefs). It is supposed to be a discussion about something that is of interest to people (not all). Coming to a definitive conclusion would be good, but it is not part of my expectations.

 

Even though it might not be possible to determine whether we are in a simulation or not, or even determine what the purpose of that simulation is, it doesn't mean that any speculation and discussion on that topic is wasted time or effort. It can help us understand our universe better as we discuss topics of relevance.

 

Also the fact that it might be possible that we are in a simulation deserves to be examined. If we are in a simulation, then it could change the way that we think of ourselves and the world around us. Finding out the truth of the matter would be the biggest discover ever (except finding out that our creators universe is a simulation too).