5 of 8
5
Simulation Argument
Posted: 06 October 2011 07:13 AM   [ Ignore ]   [ # 61 ]
Jr. Member
RankRankRankRank
Total Posts:  89
Joined  2011-09-20

From wiki, Bekenstein article:
“the maximum number of bits to perfectly recreate the average-sized U.S. human adult male down to the quantum level on a computer is 2.0057742×10^45 bits”
That sounds like an awful lot of bits.  I know were talking about ‘advanced civilisations’ and some pretty mean computing hardware, but I think we’re losing contact with reality here.

Profile
 
 
Posted: 06 October 2011 07:24 AM   [ Ignore ]   [ # 62 ]
Administrator
Avatar
RankRankRankRankRankRankRankRankRankRank
Total Posts:  15370
Joined  2006-02-14
keithprosser2 - 06 October 2011 07:13 AM

From wiki, Bekenstein article:
“the maximum number of bits to perfectly recreate the average-sized U.S. human adult male down to the quantum level on a computer is 2.0057742×10^45 bits”
That sounds like an awful lot of bits.  I know were talking about ‘advanced civilisations’ and some pretty mean computing hardware, but I think we’re losing contact with reality here.

Maybe it can be compressed by some algorithm without being too ‘lossy’ ... who knows?

 Signature 

Doug

-:- -:—:- -:—:- -:—:- -:—:- -:—:-

El sueño de la razón produce monstruos

Profile
 
 
Posted: 06 October 2011 08:00 AM   [ Ignore ]   [ # 63 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02
domokato - 06 October 2011 12:32 AM
Write4U - 05 October 2011 05:59 PM

Uncertainty pertains to our ability to predict, but that does not mean the process is random.

But we must be able to predict it in order to simulate it. Is this prediction possible? Maybe Otis knows.

I’m not really sure precisely what it is that you want to predict. If you mean every interaction that makes up our universe then no we can’t predict those, the measurement process in quantum mechanics is not deterministic.

domokato - 06 October 2011 12:32 AM

This discussion seems very quantum woo all of a sudden

I’m pretty sure we’re all laymen here, unless you have some expertise?  grin  Thanks for the paper, I’ll read it.

The only expertise I can claim is background in undergraduate/graduate physics.

Alexander80 - 06 October 2011 06:19 AM

“Regarding the Bekenstein bound, it is of interest and is discussed briefly in Seth Lloyd’s paper on the limits of computation which is referenced as supporting material by Bostrom in the argument.”

Supporting?
Doesnt it show the opposite?

Bostrom believes it supports his argument due to his naive estimate of the number of operations necessary to simulate human civilization, 10^36 operations if memory serves me, and thus he treats Seth Lloyd’s paper showing that a ‘computer’ capable of running such a simulation is not impossible by the laws of physics as supporting material, which is again a naive supposition.

Profile
 
 
Posted: 06 October 2011 09:22 AM   [ Ignore ]   [ # 64 ]
Sr. Member
RankRankRankRankRankRankRankRankRank
Total Posts:  207
Joined  2011-09-23

10^36?

No way!
This would maybe work for us alone but what is with the Rest of the universe?
I think it is nonsense to say we (our world) are the only thing which is simulated, that would be to minimalistic for most things I can imagine.
To do serious work it would be necessary to simulate a big portion of the universe, if it is not for pure entertainment, but in this case i would say that our world is too boring.
Also it does not look like the other Stars, Galaxys, Planets etc. are some sort of Wallpaper.

Another Question: Would it be possible for simulated beings, if they are not the center of the simulation but only a byproduct of a simulation of Galaxy movement, would it be possible for them to figure out that they live on a “chip”, if they are not decieved in any way by the simulator?

Profile
 
 
Posted: 06 October 2011 10:01 AM   [ Ignore ]   [ # 65 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02
Alexander80 - 06 October 2011 09:22 AM

10^36?

No way!
This would maybe work for us alone but what is with the Rest of the universe?
I think it is nonsense to say we (our world) are the only thing which is simulated, that would be to minimalistic for most things I can imagine.
To do serious work it would be necessary to simulate a big portion of the universe, if it is not for pure entertainment, but in this case i would say that our world is too boring.
Also it does not look like the other Stars, Galaxys, Planets etc. are some sort of Wallpaper.

If you look at footnote 10 in the paper, Bostrom’s estimate for the number of operations required to run an ancestor simulation appears to be based solely on the number of synapses in the average human brain and their firing frequency multiplied by the number of humans throughout history and the average lifetime of a human. There is no consideration of a consistent, external world and the operations necessary to provide this.

Another Question: Would it be possible for simulated beings, if they are not the center of the simulation but only a byproduct of a simulation of Galaxy movement, would it be possible for them to figure out that they live on a “chip”, if they are not decieved in any way by the simulator?

I’m not sure how, the simulated civilization’s empirical investigations would amount to ours in that they would just be investigating the rules of their universe and they would have no referrant by which to conclude that they are a simulation, unless they observed divine intervention or its like.

[ Edited: 06 October 2011 10:09 AM by Otis ]
Profile
 
 
Posted: 06 October 2011 10:25 AM   [ Ignore ]   [ # 66 ]
Sr. Member
Avatar
RankRankRankRankRankRankRankRankRankRank
Total Posts:  1201
Joined  2009-05-10
Otis - 06 October 2011 08:00 AM
domokato - 06 October 2011 12:32 AM
Write4U - 05 October 2011 05:59 PM

Uncertainty pertains to our ability to predict, but that does not mean the process is random.

But we must be able to predict it in order to simulate it. Is this prediction possible? Maybe Otis knows.

I’m not really sure precisely what it is that you want to predict. If you mean every interaction that makes up our universe then no we can’t predict those, the measurement process in quantum mechanics is not deterministic.

So then we can’t simulate it.

Another Question: Would it be possible for simulated beings, if they are not the center of the simulation but only a byproduct of a simulation of Galaxy movement, would it be possible for them to figure out that they live on a “chip”, if they are not decieved in any way by the simulator?

I think the simulated beings would have to be given ways to interact with the outside world by the programmers (i.e. robotics). Otherwise no.

Fractals can be used to simulate anything down to the Planck scale. It is a fundamental universal geometric property.

This is kind of wooey. Fractals are in no way “fundamental” to the universe, and how can they be used to simulate anything down to the Plank scale? Fractals are great and all, but I think you’re giving them too much credit.

However, after this instant of chaos, natural laws governing the interaction of the physical in spacetime restricted the chaotic expansion and began to bring order and a lot of self simulation at all levels and logically explainable by what we know today from the evidence.

Chaos is the problem, and it hasn’t gone away. Our universe is a chaotic system, which means that a small change in initial conditions (at any given point) can have a large effect on a future state, i.e. the butterfly effect. This is also why I think fractals will not work. Even if you use fractals to simplify the simulation, it will cause such great inaccuracies that the simulation may not be very useful anymore. I guess it depends on what use is trying to be gotten out of it.

 Signature 

“What people do is they confuse cynicism with skepticism. Cynicism is ‘you can’t change anything, everything sucks, there’s no point to anything.’ Skepticism is, ‘well, I’m not so sure.’” -Bill Nye

Profile
 
 
Posted: 06 October 2011 11:06 AM   [ Ignore ]   [ # 67 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02
Otis - 06 October 2011 10:01 AM
Alexander80 - 06 October 2011 09:22 AM

10^36?

No way!
This would maybe work for us alone but what is with the Rest of the universe?
I think it is nonsense to say we (our world) are the only thing which is simulated, that would be to minimalistic for most things I can imagine.
To do serious work it would be necessary to simulate a big portion of the universe, if it is not for pure entertainment, but in this case i would say that our world is too boring.
Also it does not look like the other Stars, Galaxys, Planets etc. are some sort of Wallpaper.

If you look at footnote 10 in the paper, Bostrom’s estimate for the number of operations required to run an ancestor simulation appears to be based solely on the number of synapses in the average human brain and their firing frequency multiplied by the number of humans throughout history and the average lifetime of a human. There is no consideration of a consistent, external world and the operations necessary to provide this.

Thinking on this some more, I recall some research indicating the belief that computational complexity in the brain can be accounted for by the firing frequency of synapses alone to be naive. A popular article on the subject is available here.

Profile
 
 
Posted: 06 October 2011 11:20 AM   [ Ignore ]   [ # 68 ]
Sr. Member
RankRankRankRankRankRankRankRankRank
Total Posts:  207
Joined  2011-09-23

If you look at footnote 10 in the paper, Bostrom’s estimate for the number of operations required to run an ancestor simulation appears to be based solely on the number of synapses in the average human brain and their firing frequency multiplied by the number of humans throughout history and the average lifetime of a human. There is no consideration of a consistent, external world and the operations necessary to provide this.

I dont think this would work, you would need more to simulate a convincing world than just brains.

Profile
 
 
Posted: 06 October 2011 11:26 AM   [ Ignore ]   [ # 69 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02
Alexander80 - 06 October 2011 11:20 AM

I dont think this would work, you would need more to simulate a convincing world than just brains.

I agree, if there is to be any notion of consistency of sensory input then a mechanism is required by which to compute the necessary input. I’m not sure what Bostrom thought he was doing when he came up with this estimate.

Profile
 
 
Posted: 06 October 2011 04:15 PM   [ Ignore ]   [ # 70 ]
Sr. Member
RankRankRankRankRankRankRankRankRank
Total Posts:  207
Joined  2011-09-23

Bostroms simple Idea is complete out of question, but others are more convincing… are they?

Profile
 
 
Posted: 06 October 2011 04:31 PM   [ Ignore ]   [ # 71 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02

10^36 operations

Not that it matters much, but this should be 10^37 operations not 10^36, Bostrom’s calculation in footnote 10 is in error.

Something that isn’t discussed in Bostrom’s paper is the potential energy requirements of a simulation of the type he supposes. Current and foreseeable computational techniques employ irreversible processes leading to the dissipation of energy, there is a fundamental [1] temperature dependent limit to the minimum energy dissipated per irreversible [2] bit operation, the von Neumann-Landauer limit.

A room temperature computer operating at maximum efficiency expends a minimum of ~3x10^-21 Joules per operation (in contrast modern supercomputers operate at around 10^-9 Joules per operation), assuming this and Bostrom’s estimate of 10^37 operations per ancestor simulation the minimum energy cost of an ancestor simulation is ~3x10^16 Joules, equivalent to the output of a 1 GW power station over a year.

Is there any reason for us to think that energy would be so freely available in the future that such an expenditure would be trivial, I don’t think so, and remember this is a maximum theoretical efficiency, whether it will become a practicality or not we have no way of knowing. Factor in the operations required to provide a dynamic and consistent universe that Bostrom omits in his estimate and the corresponding minimum energy cost will be many orders of magnitude larger.

[1] A few people have contested the validity of the von Neumann-Landauer limit as a physical law, but its status as a law is largely accepted.
[2] It is theoretically possible to employ reversible computing techniques which would not obey this limit, but to quote one researcher in this field, “To build a *practical* reversible computer has turned out to be an extremely difficult engineering problem, and might not even be possible.” and “we still don’t even have a truly *complete* and physically realistic *theoretical* model of reversible computing…”. If it were to become possible it would greatly increase the computational complexity overheads of an ancestor simulation and error correction within the simulation would still be irreversible.

Profile
 
 
Posted: 06 October 2011 04:48 PM   [ Ignore ]   [ # 72 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02
Alexander80 - 06 October 2011 04:15 PM

Bostroms simple Idea is complete out of question, but others are more convincing… are they?

I’m not entirely sure what you mean, Bostrom’s idea is really a disjunction of three propositions, arguing that simulations of our own world are impossible or that ancestor simulations in any form are impossible merely supports the validity of the first two propositions in opposition to the validity of the third, it doesn’t dismiss the Simulation Argument, although the certain impossibility of ancestor simulations would obviously make the argument trivial. I think that the argument is incorrect but for different reasons, outlined here and here.

What do you mean by other ideas are more convincing? Do you mean other types of simulation, not full ancestor simulations?

[ Edited: 06 October 2011 04:56 PM by Otis ]
Profile
 
 
Posted: 06 October 2011 06:09 PM   [ Ignore ]   [ # 73 ]
Sr. Member
Avatar
RankRankRankRankRankRankRankRankRankRank
Total Posts:  6055
Joined  2009-02-26
domokato - 06 October 2011 10:25 AM
Otis - 06 October 2011 08:00 AM
domokato - 06 October 2011 12:32 AM
Write4U - 05 October 2011 05:59 PM

Uncertainty pertains to our ability to predict, but that does not mean the process is random.

But we must be able to predict it in order to simulate it. Is this prediction possible? Maybe Otis knows.

I’m not really sure precisely what it is that you want to predict. If you mean every interaction that makes up our universe then no we can’t predict those, the measurement process in quantum mechanics is not deterministic.

So then we can’t simulate it.

Yes that would be a logical conclusion. However it seems that Loll’s CDT (Causal Dynamic Triangulation) points to a fundamental Universal structure and geometry.

Fractals (CDT) can be used to simulate anything down to the Planck scale. It is a fundamental universal geometric property.

This is kind of wooey. Fractals are in no way “fundamental” to the universe, and how can they be used to simulate anything down to the Plank scale? Fractals are great and all, but I think you’re giving them too much credit

Not according to Loll and her associates. The fractal geometry of the universe seems to hold at any level in a non-disturbing and non-conflicting way down to Planck scale (her words). I would call that a fundamental universal property.

However, after this instant of chaos, natural laws governing the interaction of the physical in spacetime restricted the chaotic expansion and began to bring order and a lot of self simulation at all levels and logically explainable by what we know today from the evidence.

Chaos is the problem, and it hasn’t gone away. Our universe is a chaotic system, which means that a small change in initial conditions (at any given point) can have a large effect on a future state, i.e. the butterfly effect. This is also why I think fractals will not work. Even if you use fractals to simplify the simulation, it will cause such great inaccuracies that the simulation may not be very useful anymore. I guess it depends on what use is trying to be gotten out of it.

There may well be instances of a butterfly effect, but that does not negate that even the butterfly effect follows strict rules in its evolutionary path. IMO, it is not proof of a chaotic indeterministic universe, rather the reverse. I will stipulate that a perfect sim of such events is impossible.

True, the universe started in a chaotic way, but clearly we can see geometrical similarities and patterns emerging in universal evolution in all parts of the visible universe from spiral galaxies in the far reaches to Ferns, Broccoli, and neural/vascular systems in higher order living organisms.
Lest we forget, cell division (self duplication) is a fractal process. All we need do is look at our diagrams of molecules and combinations. The fractal nature of such elementary structures and combinations is clearly evident.

Chaos has become ordered,  inevitably, in accordance with the geometric structure and function of spacetime (evolution). The elegance of CDT as a fundamentally simple form and instruction, allowing for and causal to an extraordinary functional complexity, is a very compelling argument.

By that logic, it may be theoretically possible to fashion a CDT simulation which mimics the actual workings of the universe. But as our universal structure “emerged” in reality from a unique state of chaos, it would be impossible to recreate a precise simulation of that chaotic beginning. There can be only one state of chaos, pure energy created in a single mega quantum event (BB) where space did NOT keep things apart and time did NOT keep things from happening all at once. Once inflation created spacetime, everything from that point on was deterministic and in accordance with CDT.
Perhaps an ancestral sim could be fashioned from a point in time “after” inflation.

According to Loll, fractal geometry (CDT) may well be useful in solving “quantum gravity” and is not in conflict with any of the established theories about the structure of and functions in our universal spacetime, as we know it.
Already, CDT allows for accurate simulation and measurement of coast lines (to a high degree of accuracy), heretofore an impossible task.

[ Edited: 06 October 2011 06:34 PM by Write4U ]
 Signature 

Art is the creation of that which evokes an emotional response, leading to thoughts of the noblest kind.
W4U

Profile
 
 
Posted: 07 October 2011 10:30 AM   [ Ignore ]   [ # 74 ]
Sr. Member
RankRankRankRankRankRankRankRankRank
Total Posts:  207
Joined  2011-09-23
Otis - 06 October 2011 04:48 PM
Alexander80 - 06 October 2011 04:15 PM

Bostroms simple Idea is complete out of question, but others are more convincing… are they?

I’m not entirely sure what you mean, Bostrom’s idea is really a disjunction of three propositions, arguing that simulations of our own world are impossible or that ancestor simulations in any form are impossible merely supports the validity of the first two propositions in opposition to the validity of the third, it doesn’t dismiss the Simulation Argument, although the certain impossibility of ancestor simulations would obviously make the argument trivial. I think that the argument is incorrect but for different reasons, outlined here and here.

What do you mean by other ideas are more convincing? Do you mean other types of simulation, not full ancestor simulations?

No, if there would be an unimaginable gigantic Metavers in which our whole Universe could be simulated down to the quantum level.
I dont know if this sounds convincing…

Profile
 
 
Posted: 07 October 2011 12:05 PM   [ Ignore ]   [ # 75 ]
Jr. Member
Rank
Total Posts:  24
Joined  2011-10-02
Alexander80 - 07 October 2011 10:30 AM

No, if there would be an unimaginable gigantic Metavers in which our whole Universe could be simulated down to the quantum level.
I dont know if this sounds convincing…

Entertaining the possibility that our world might be a simulation run in a universe or metaverse with characteristics that permit a much greater level of computational complexity than that exhibited by our own universe is a necessary consequence of the supposition that we might be a simulation, as to consider this supposition is to consider any scenario that might result in a simulation of our civilization. The argument is not conceptually limited to the idea that we can fully simulate this universe within our own universe.

I think there are some misconceptions in this thread about what constitutes a simulation that must be considered by the SA. Remember that this is a sceptical argument, concepts of simulation need not be limited to a full simulation of this universe everywhere down to the scale at which quantum behaviour becomes apparent and beyond in the case of the waffle above about CDT, all that is necessary is a simulation that constitutes an individual’s direct experience. We can argue about what would be necessary under this, but anything that isn’t directly experienced by an individual need not be simulated fully or perhaps not even at all.

Profile
 
 
   
5 of 8
5