I thought this was your quote masque:

"A glass of water is filled to 50% capacity. There are some that see it half empty. There are some others that see it half full.
And then there is a 3rd kind that see it as having enough deuterium (heavy water molecules) that one day will be used to extract enough energy through fusion to fill hundreds of glasses with clean drinking water starting from any dirt
y water source you wish...
I choose to be the 3rd kind not because it is easy, one must remain both a realist and an idealist at the same time and fight the battle that brings change, it wont happen on its own. The choice of rational optimism is enriching one's life through meeting challenges which demand creative synthesis. I choose to be an optimist because i am not fully satisfied with the present status of human condition. I am not an optimist by force or cult infatuation. I am an optimist by scientific inevitability, because it provides a purpose to life.."

The above seems awfully pessimistic.

Do we have moral obligations to sophisticated computer programs we create? How about to highly sophisticated computer programs that can learn and adapt?

What do you mean the above sounds pessimistic. Which part? The deuterium example is pretty realistic within this century. The faithful indistinguishable to reality simulation of a solar system let alone galaxy is impossible even if you are Kardashev order 3 civilization.


Also for anyone interested;

http://en.wikipedia.org/wiki/Lattice_qcd

I don't know, I don't think it's so straightforward. I think morality is fundamentally connected to a being's capacity to feel pain and pleasure. The way I might try to get 'outside' is to say that the badness of pain and the goodness of pleasure is the same for all beings with a capacity for such experiences.

Computer programs don't have feelings, so no. If a highly sophisticated computer gained feelings, we would have moral obligations to it.

But, the rate at which processing power increases is exponential. This doesn't help at all? This is possibly the fastest path that any type of evolution can take, no?

No the Moore's law whatever trend is ridiculously impossible to extend once you hit the quantum limits (so do not expect it to last long more). Of course we may find ways to do more eventually plus quantum computers even but my point is that a faithful enough simulation of the entire system is ridiculously close to having the very system you simulate. I mean it is as "expensive" as having it for real.

For the galaxy example i used i think they only use a few million particles or so maybe (less than 10000 times the real thing and of course this is not even the real thing because the real thing has every "sun" as another ~10^57 particles and times 10^11 more for all visible universe) to see the galaxy structure change over time and establish the patterns seen in nature. Some simulations can tell us a lot about physical systems because the main structure forming effect is not depending sensitively on the vast number of particles, it kind of averages out after some number big enough. Ie if you simulate how a solar system forms planets etc you do not need 10^57 particles, you can again try a few million maybe billion. So you do not exactly need the real thing you can get away with a significantly smaller fraction. But when??? Only when you selectively are after specific patterns. If you want the perfection of the real thing you are hopeless eventually.

By that i mean take CERN experiments for example. In those we have effects that take place within 10^-23 sec or within 10^-16 m etc. For this all to be simulation you end up requiring resources that are more or less nature itself lol. So thats why i said why bother, it will stop being faithful very fast and in order to be very faithful to reflect standard model details etc we check with experiments randomly performing in any part of he world we like, you really are starting to get as tough as creating the damn thing itself.

Plus of course there are other ways to tell that are a lot tougher to analyze but they have to do with how fine structure eventually survives to large scale and you see patterns that wouldn't exist if the system wasnt as refined basically.


Do you see what i mean when i use CERN as example??? We can in principle select any location of the planet to create a particle physics experiment/detector etc. So whoever is supposed to be simulating it all better be simulating all spacetime we have access to, to Standard model detail and beyond even as we advance in order for us to be seeing the same kind of collisions/effects in US, Europe, Russia or Japan etc.

(off topic) this is how progress works:

http://en.wikipedia.org/wiki/Logistic_function

I don't believe that another jump to 3bill in the future will make us much different vs todays example humans vs bacteria. There is only so much we can learn.

(on topic) I always wondered if it was a simulation we would never be able to observe it since there would be no leakage of information. Maybe some strange observation like the quantum mechanic but nothing definite... Something like we will never understand the whole universe with precise accuracy since we can't observe it "from the outside"

you are constraining yourself on the current universe, the energy demand could be minuscule in relationship to the universe where we are a simulation, no wonder why the speed of light is so small :P

@smrk2: Before you ask whether the creator should be concerned with ethics you have to ask whether he should have reason to assume that his simulations could become conscious.



What I see is that our instruments of inquiry have encountered a boundary, the boundary is one of our instruments, we have no definite reason to reject the possibility that the universe may not be discrete at the smallest scale. Consider the alternative, were we in the future to develop an instrument with which we could represent a model where it's 'turtles all the way down' then it's still computational, we could still argue that the universe may be simulated simply by virtue of us having simulated a coherent model of a universe that is non-dicrete at the smallest level.

So, I think we can make the same argument either way. The simulation hypothesis becomes more interesting when we manage to introduce actual consciousness into the simulation, and so far I see no way how we could do that. Consciousness is the actual game-changer, not the discrete vs non-discrete debate.

It is illogical to trust your own thought process to conclude that you are not in a simulation. If you were in a simulation, then what your thoughts make you believe would be entirely dependent on how you were programmed. Perhaps you were never intended to draw correct conclusions about the reality outside your simulation. The fact that you try to do this may be a bug, or it may have been anticipated and deemed to be harmless, or maybe someone just decided to code up a masque de Z as a novelty program for their own amusement.

The problem with all of this is that it assumes a civilization can accurately simulate what would be a meaningful proportion of its universe, a simulation can accurately simulate what would be a meaningful proportion of its universe, etc. There's no reason to rate this as very likely, and without it, the "geometric series" drops off really fast and the argument fails.

Now that I've read the article, something doesn't make sense. They say that certain experiments could be done which could confirm that the laws of physics are superimposed on a 3-D lattice. But why wouldn't that just show that space is really is a 3-D lattice, rather than a discretized version of continuous laws of physics caused by a simulation? How could we possibly know that the actual laws of physics are continuous if we are in a simulation and don't have access to the actual laws? In fact, what sense does it make to say that "the laws of physics appear to be continuous"? They cannot appear to be continuous because we cannot measure space or time intervals shorter than a certain threshold. In writing about the n-body problem, Newton regarded continuity as merely a convenient approximation to a discrete problem. I don't see any reason why all of the continuous laws of physics are not just the same type of approximation.

exactly, we can never know for sure observing from within the space-time...

I have another post i was writing slowly over time about this thread for another response but yes it is true what you say and what i did say was definitely not written as i intended it (although i will try here to even find a small chance for it to be meant as you thought i did). I am not (at the time) saying in the quoted that if you were in a simulation you would have an easy time figuring it out and in fact maybe not possible at all (you will see the other post in detail arguing that). What i meant here is that its easily spotted as simulation if you know the real nature (hence the galaxy simulation examples and then the CERN argument). As a comparison i mean and as detailed lack of universality in laws etc. That in order not to fail the comparison you end up pretty much recreating nature itself so its no longer a simulation its the real thing.

That can also be said as if you were inside a very good simulation that was able to create even conscious beings eventually as a result of its rich complexity and detail, you probably wouldn't be able as this entity to find anything wrong with it, including even random bugs, it would feel as this is your laws of nature. (eg rounding errors energy non conservation slightly becomes your new norm-law)


Basically a simulation is the real thing either in the sense that it has created a very real world that is self aware eventually but unable to see its a made up world , it looks to it very natural, so take that to mean every advanced simulation is a real universe. Or take it to mean that a simulation of the real universe is a real universe not a simulation (it becomes the same thing as the detail needed to replicate reality is as astronomical as that reality).


Furthermore a pending revolution with QM at its core will completely kill this discussion by rendering irrelevant the typical "lattice" logic simulations.


So lets have this talk in a decade for better detail and self consistency.


But for now my claim is that under present day framework a simulation of the real thing is basically a real thing , it cant be as faithful without it being the real thing.


One last thing that may be a very ambitious possibility and of course my preference although i am prepared to not prove true. I really want the theory of everything eventually to show that there was absolutely no choice in the structure of the universe, ie it could never be anything other than that for self consistency reasons (the hell with landscape theories by the way and dont tell Lenny Susskind that i said so yet lol although he doesnt know who MdZ is he could figure it out). That doesnt necessarily mean completely unique parameters etc in everything but the final choice of all is self consistent to many levels meaning that yes you could have had a different universe but the change in all parameters would be coordinated for the same type of self consistency.

That is not another landscape idea because i think the restrictions would turn out to be much more severe than the freedom string theory wants to allow them to have creating 10^500 etc cases or whatever.


If that last is true then maybe a simulation is never able to produce that perfection and the thinkers inside it can eventually find what the theory would be if the world was maximally self consistent and recognize their world is not perfect in that sense (that possible only if the simulation is a very good one that offers a clue about what it would all be about eventually but it fails the perfection consistency test, a very naive simplistic simulation will either failto create awareness or be so far from the real thing that such theory cannot be developed by these entities). But that is a small chance long term result if such consistency is possible to obtain in the sense that all possible universes share the same type of theory of everything albeit with some different initial parameters etc. I would be excited if indeed it was possible to find such perfect theory and then see your world not be that and conclude its false unfaithful simulation of the real thing.

Kind of what Albert Einstein said when they asked him what would he think if the results of the solar eclipse in Brazil didnt confirm General Relativity and his response was i would feel sorry for God.


I think string theory and the rest of the guys have already given up on such possibility of extreme self consistency and are trying the easy way out of course (anthropic principle taken to the extreme etc) using maximal degrees of freedom and complex enough structures like extra dimensions (to avoid problems ,infinities etc) in order to avoid doing the really necessary heroic breakthrough work that a true revolution is always about.

Ok this turns out a bit lengthy. Anyway. (it may also overlap now with another post )


No i am not constraining myself on anything other than rational thinking supported by the science we have recovered so far and leaving room for more too but never taking that room more seriously than the established "facts". I dont care if you think there is a hyper universe simulating us that is 10^100 times larger which therefore sees us as tiny and "easy"/affordable or whatever. (for me such ideas are bs close often but ok maybe not always to mental @$$...@*& actually, but it passes as cool today in the prerevolution last days of the old system due to our scifi and computer dominated era that we view suddenly computers and programming/simulations etc as the mastergame of the universe our cheap gutless excuse of why the world looks like it does, instead of simply putting our butts down to do the hard work and explain the bloody universe we live in the honorable Dirac/Feynman/Heisenberg/Einstein/Newton/Archimedes etc style way). Whatever. Even if that was the case they still need in order to simulate us a system as detailed as nature itself so basically they are not bloody simulating anything, they have created the entire system.

Do you see what i mean???

Every universe is a simulation or the real thing, pick whatever you want to call it. The faithful simulation can only be the real thing itself. So a stupid naive computer simulation in a desktop present day computer of a galaxy will do a kind of good job to simulate structure formation but will bloody hell fail in all aspects involving standard model etc refined solar systems satellites and their craters, magnetic fields, emergence of life you name it, it will require laws for all this too and the fine detail needed to produce the results seen here. Essentially it would start requiring >10^122 or whatever big 4d cubes space for a even a single second of evolution of all solar system. In order not to fail the simulation needs to be as detailed as the very system it is supposed to be simulating. Therefore who cares??? One may choose to see nature as a simulation and another as the real thing. The fact is both the simulation and the real thing are essentially identical if the simulation is faithful enough. It cannot be simulated more economically than the real thing.

So where is it that i am constraining myself? All i am telling you is that if there is some simulation you are imagining its the same as a full creation. We have already probed the system so deep and there is still promise that the simulation is pretty faithful one and basically demands the same resources as the real thing.

So i dont bloody care for things i cannot determine. To me they are the same thing. Some excuse for people to publish irrelevant to true science toy papers that will further prove problematic once physics changes the framework we see a simulation today look like.

Besides as i have repeatedly promised (i dont think i will be proven wrong in my anticipation) the world will prove dramatically more impressive than any lattice simulation or theoretical version of it using current framework very soon.



Do you see what i mean??? A very good simulation that recreates consciousness for example needs to be so detailed that is beyond our current capacity but will eventually happen. We will indeed be able to create a simulated reality that has its own laws or the version of our laws up to a point. (if we insist to be extremely faithful we have to basically recreate a world as big as the simulation of it). If then an entity emerges in that simulation and starts understanding its world it will recover the laws we gave it eventually possibly even all of it over time if its not complex and stochastic enough to be changing all over the place. For that entity in that simulation its world will look perfectly real and will be real enough because the universe is that program running. It is impossible for an entity "living" in that system to come out and say hey i am in a simulation. How would they know it??? To know it requires to know the laws of the real world and to spot your world is somewhere failing a little bit to be faithful. How on earth will you know the laws of the real world when all you have to play with is your own version?????????????????????????????? (* again see other post for a slight chance that still is possible though)

See what i mean? This is why its mostly academic bs (the papers claiming methods to detect simulation etc yeah right) and not valuable science as it is essentially speculation without the benefit of a deeper understanding of nature we still need. Who the hell cares if this is a simulation when a true simulation of the real thing is the real thing!!! And an untrue simulation of the real thing will never recover the complexity of the real thing , it will have nevertheless its own structure and complexity albeit maybe not as rich and the entities inside it that emerge and have consciousness if the system is complex enough will never know what the real world is, only what their world is and it will not be showing anywhere any signs its a simulation because to spot these signs you need to know the real world to compare it with. To tell it differently a simulation that has some rounding errors etc will mean for these guys that part of their laws is those errors. (for you that know the real world this error will scream simulation but not for them). Those errors wont look to them irregular. They will simply think eg energy is not conserved, they will not recover such law. Or they will recover a law that tells them energy is behaving as if it is almost conserved but there are rounding errors that take place that fail to conserve it perfectly always. But they will be unable to know that. Instead they will not recover the conservation theorem as we did, they will recover a version with rounding errors. For them the world is their world, they cant project our world based on the simulation. Even if they found an error in the program ie a process that fails to nearly conserve energy and its losing a lot more than another nearly identical process elsewhere,that will not be indicating them there is a bug in the system. They can easily see it as part of a more complex law, ie the bug is now part of their laws.

They may even recover a method to kill the program, that will too be part of their laws. In fact if we decided we could kill the planet. Its not different than crashing the system. Its our version of crashing it (ie send an asteroid to earth) (over time we may even be able to kill the sun etc that too will feel like a bug???) Only our "bug" feels a little bit more normal. But who is to say its normal??? If we had not experienced this reality and we had a version with strange bugs those bugs would look pretty much the reality.

The fact is before energy mass conservation we had mass conservation but at some point ie radioactivity, that changed! To someone seeing this as a simulation that violation of the law is a bug or in fact a revelation that the law is not that of mass conservation but of mass-energy. If our system had the bugs we imagine now for the simulation we would have called the laws differently to try to describe those bugs even if they appeared somewhat irregularly (lot of things for now seem to appear irregularly in our world as well but eventually are explained in rather economical ways so our world feels not very bug rich so to speak but if it did feel like that we would have no way of knowing this is irregular hence a code error or whatever hint of simulation!)


Today even modern day computers up to a point have some consciousness approximating functionality. That doesnt mean they have a chance in hell to find out what QM is if all the program ever does is run eg windows 8 and never connect to the internet to read news lol!!! It will simply never recover QM or relativity etc based on all it experiences which is the basic desktop functionality.

Yes we could create a compelx enough computer that feels like a real human and put it in a virtual world to start exploring with amazing detail and in order to recover the detail of our world it has to be the world itself not a simulation that is approximation. If it is approximation the entity living there will never know it unless we leave a sign at some place hey , you are a program, this is what the real world is like!!! (and again with a slight small chance in another way in the other post)

Well who cares if i find a sign telling me i am a program!!! I already know i am a complex intriguing quantum program and my organic chemistry and eventually all my quantum level systems and their interactions are the code!!!


People are stupidly imagining that they are in a simulation and they are smart enough to figure it out without us yet even having a theory of everything or anything close to it or even a conclusion it doesnt exist. BS. They cant figure out anything, they will at best recover the laws of the simulation and thats the end of the story. And the bugs will be part of the laws as seen by them!

I don't know much about computers or simulations, so please correct me if I'm wrong, but it seems like it's impossible to perfectly simulate, for example, an atom, using less than one atom. So, any 'simulation' that we could create to perfectly reflect our universe would, at best, simply be a controlled re-ordering of matter. If we wanted to simulate the Earth in a certain state, we would literally have to build another Earth in that state.

Any Earth simulated in a computer (of reasonable size) would be orders of magnitude less complex than the actual Earth. And if we decided to simulate our entire universe it would be so far from the real thing that the approximations and flaws should be obvious (??) to any sufficiently intelligent life in that simulation.

This is also where the infinite simulations falls apart very quickly, because the decrease in complexity from one simulation to another would very quickly not allow for something as complex as intelligent life. And even if I'm misunderstanding and there is some way around this, the original simulation would just crash as soon as a secondary simulation is created in the first.

@masque de Z funny and interesting at the same time, seems you have a tendency to say things in a long way for what I reserve a sentence or two...hope you see that :P

maybe we could run our own simulation to analyze bugs in our system to see what kind of results we get if we change our parameters (like the speed of light to not be a constant) and in that way find what universes are stable and what are not...

Yes. Supposing that the simulation hypothesis is true then the simulated universe must necessarily be extremely less complex than the parent universe. The reason is simply that you have to 'fit' the simulated universe into the parent universe. Where would you put the simulated universe? You would have to dramatically reduce complexity or size, otherwise it wouldn't fit. Same goes for every subsequent universe (simulation within simulation). You can't just reduce the size, because the simulated universe must be made up of stuff that is the same as in the parent universe. The stuff out of which the simulated universe is built must necessarily abide by the same laws as the parent universe's. One can't change the laws by which the essential building blocks behave. Therefore the simulated universe has to be the same as the real universe albeit with extremely reduced complexity. But, as we see already with our universe, we can't possibly simulate a universe that we even remotely could describe as an accurate approximation of the universe we're living in.

obviously running the simulation is not used to simply replicate the world but to assess your hypothesis in a controlled, scalable environment where similar fundamental logic applies but complexity is greatly reduced.

even if it is a 1:1 ratio in size/energy a simulation of space-time created in parallel with original space-time but where our "simulation" is with a faster time (speed of light). An excellent model to predict the future outcome and to prove theories....

I know i use length often that can be cut for sure but recall i try to "talk" to everyone and even few people that know less on some topics so detail and examples matter and also i am very sure a lot of what i say is not possibly covered in few lines. The problem with my posts is that due to length trying to cover many things (mostly for my own benefit often) i am tiring , not ideally organized i am sure and efficient editing would have helped but then again i have limited time as i do other things in parallel and proper editing takes more time than the writing itself which is already too much. It takes design and planning and rewriting. Most authors probably write a topic and then rewrite it on top of it and rearrange it many times. Can you imagine me doing that here??? Ideally i should have!

Also unfortunately here i was writing something else to you as i noticed BruceZ get a different interpretation of what i intended to mean so a double post resulted.

Apologies to people for some repetitive results.

If consciousness can't be simulated, then let them run any simulations they like. If consciousness can be simulated, then I think they have moral obligations to not the run the simulations at all, or at least they must prove that none of their simulations can possibly turn out to be terrible for the beings created in their simulations before they run their simulations. If they're not sure whether consciousness can be simulated, then they should err on the side that it can be and only run simulations which don't come close to simulating beings that may or may not be conscious; stick to simulating galaxy collisions or planetary geology or strategies for games.

I don't see why it would matter how much simulated pain and suffering we allow as long as the simulated beings can't retaliate against us in any way.

You have me there, that's a sound moral principle.

I don't think morality has anything to do with it. Morality is just a set of cultural conventions that we adhere to in order to get what we want. If we already have everything we want from the simulation, then we have no need of morality.

I don't agree with that view of morality, but what if we merely wanted to reduce or eliminate suffering in the universe (in lieu of a moral principle), and on the basis of that want we refuse to run the simulations?