back to article Forget the 'simulated universe', say boffins, no simulator could hit the required scale

That “we live in a simulation" trope being advanced by Elon Musk and some folk on the fringes of science? Fuggeddaboutit, because it's impossible to build a simulator that would reproduce what humans already know about quantum systems. That's the confident conclusion drawn from a paper that revels in the title “Quantized …

Page:

  1. Pascal Monett Silver badge
    Coat

    "To model just a few hundred electrons needs a computer bigger than the universe"

    Okay, I know I'm the village idiot next to these guys and I'm sure they know what they're talking about, but that sentence really bugs me. It is probably due to the sentence further down that says that adding another particle doubles the size of the model, so they just did an exponential of 2 to the 100th power and yeah, for sure that's big, but is that really how the simulator would work ?

    What if it was a quantum simulator and each particle was thus analyzed by a quantum dot that considered all of its possible positions simultaneously ? Isn't that how it should work ?

    I don't know. It's way too early for this anyway. I'm quitting this before the headaches start.

    1. DropBear Silver badge

      Re: "To model just a few hundred electrons needs a computer bigger than the universe"

      I'm finding it remarkably bold to attempt to infer any conclusions about our universe's simulated nature from any findings regarding complexity considering that IF we are indeed in a simulation we have not the faintest idea about a) what kind of hardware does the simulation / what exactly a unit of it can simulate (one would expect a very different performance from an ALU and an FPGA implementing the same complex logic after all), b) how mind-bogglingly large it might be compared even to the vastness of our own universe and c) how much of unobserved complexity it does simulate down to its most accurate level and how much of it gets fudged by ingenious shortcuts and workarounds when we're not looking at it and finally d) how much it redirects intractable complexity from the space into the time domain - ie. how much it slows down (not something that we would be able to detect as long as the rest of the simulation is kept in sync with the slow bits) whenever it needs to recurse deep into some detail (nobody said we're running at a stable 100FPS in "outside world" terms - we might be a slow and unsteady as fuck scientific simulation...)

      Let us not forget for instance that any detail of interest of a fractal image down to arbitrary resolution CAN be calculated on-demand never bothering with calculating "all" (for whatever arbitrary limits we care about) the image at the same resolution, working from a very much uncomplicated formula. FFS, we don't even have any idea how many dimensions the outsiders might have at their disposal to expand their mainframe into - it's not like all simulations we run are in full 3D!

      So yeah, I'm fully prepared to believe those guys that were we trying to simulate ourselves, especially in a naive brute-force fashion, the ramping complexity would prevent it really quickish-like; what I'm much less prepared to believe is that they know what they talk about when they say "impossible" (or more specifically, that they don't need to rely on a massive heap of embarrassingly arbitrary assumptions to say that).

      1. Doctor Syntax Silver badge

        Re: "To model just a few hundred electrons needs a computer bigger than the universe"

        e) If I (and this has to be expressed in 1st person terms) exist in a simulation then none of you, el Reg or anyone else exist except as inputs presented to me. That includes the papers you write which I'd have no way of verifying as I have no access to the observatories that don't exist. So if one were to consider oneself as living in a simulation one doesn't have to believe that everything an entire universe is being simulated all the time because nothing outside one's immediate experience needs to be simulated all the time; it can just be instantiated as needed.

        1. Yes Me Silver badge
          Angel

          Re: "To model just a few hundred electrons needs a computer bigger than the universe"

          Actually it needs a computer which is exactly the same size as the universe. The universe is, existentially, an exact simulation of itself. That doesn't mean the paper is wrong. Those electrons don't exist in isolation; because of entanglement, it's quite artificial to separate them from the rest of the Universe, and if you do so, you're going to have simulate the effect of the rest of the Universe on them... so you're going to have to simulate the whole Universe anyway, which of course needs a simulator that contains more bits than the whole Universe, even if it's a quantum computer, and even if quantum theory is the true theory.

          All we can ever simulate, even in principle, is an approximation.

        2. fishdog

          Re: "To model just a few hundred electrons needs a computer bigger than the universe"

          "If I (and this has to be expressed in 1st person terms) exist in a simulation then none of you, el Reg or anyone else exist except as inputs presented to me."

          This is a possible consequence, not a necessary consequence.

          Don't get me wrong, I agree with you, such a solution would be n to the nnnth power easier to simulate, but if you're going to critique others logic you have to accept criticism of your own, as well.

          My programming told me to say that.

      2. Muscleguy Silver badge

        Re: "To model just a few hundred electrons needs a computer bigger than the universe"

        Musk's original claim was that we can simulate reality to details that look photorealistic so therefore simulation is possible ignores the fact that simulation games use things like ray tracing and are built from triangles etc which mean the Planck length in the 'photorealistic' simulations is HUGE compares with the Planck length we have measured in this universe.

        It would seem to me that the Planck length would ineluctably have to increase for each level of simulation compared to the previous one to avoid the problems this paper details. The shortcuts you describe in effect very greatly, orders of magnitude increase the Planck length. There is no getting away from that.

        BTW my wife was finishing up her CompSci degree in the department of the University of Otago which did the Americas Cup simulations and graphics in the early '80s which broke the ground of what had been possible so despite being a mere biologist I have an inkling of the technologies, methods and issues involved. The effective Planck length of those simulations would have been of the order of whole centimetres.

      3. tekHedd

        Re: "To model just a few hundred electrons needs a computer bigger than the universe"

        "any detail of interest of a fractal image down to arbitrary resolution CAN be calculated on-demand"

        It seems to me that the very nature of quantum physics suggests that perhaps the system is designed intentionally such that nothing needs to be known until it is observed/needed. But what do I know? Nothing until I observe it, apparently. :)

    2. Aitor 1 Silver badge

      Re: "To model just a few hundred electrons needs a computer bigger than the universe"

      It is bullocks, it can be demonstrated that of course you cannot model the universe, but you can model a relatively small part of the universe.. that we do not know HOW to do it does not mean that it CANNOT be done..

      Anyway, it matters not if it is simulated or not, it makes 0 difference. I vote for non simulated, but who cares?

    3. Alex Read

      Re: "To model just a few hundred electrons needs a computer bigger than the universe"

      That, and the title bug me... there's no need for the comma before the AND

      1. Sam Therapy
        Headmaster

        Re: "To model just a few hundred electrons needs a computer bigger than the universe"

        Yep, there is. Read up on the Oxford comma:

        https://en.wikipedia.org/wiki/Serial_comma

    4. John Smith 19 Gold badge

      " for sure that's big, but is that really how the simulator would work ?"

      IIRC the point about a quantum computer is that it computes all answers to a problem EG simulating a universe at the same time.

      IOW it's not just the best computer available, it would be the best computer that could ever be available. *

      And it's still not powerful enough to handle this problem. :-(

      *Keep in mind that what physicists call a computer can just as easily be an analog computer for a specific task, which is how I'd describe a lot of the reports of "quantum computers" I've seen over the years. Great for the class of problems it's built for, but not programmable in the sense of actually writing a program, loading it and running it.

    5. Just Enough

      Re: "To model just a few hundred electrons needs a computer bigger than the universe"

      The correct answer to this statement should be; "So?" If we are in a simulation, we have absolutely no concept how much bigger the environment is outside of the simulation.

      Plus the whole point of models is to simplify things. Electrons are models. Effective models, because the mostly explain the behaviour of what we know of electrons in terms that we understand. But you don't complicate a model with information that you don't need for the immediate task at hand. Simulations are no different.

      1. nijam

        Re: "To model just a few hundred electrons needs a computer bigger than the universe"

        > If we are in a simulation, we have absolutely no concept how much bigger the environment is outside of the simulation.

        And perhaps more relevant, what the laws of physics are in that environment. There's certainly no requirement for them to be the same as those we (or our experiments) experience.

        But as has already been mentioned, it really doesn't matter.

    6. scrubber
      Boffin

      Re: "To model just a few hundred electrons needs a computer bigger than the universe"

      By definition a computer EXACTLY the size of the universe IS able to model somewhat larger than just a few hundred electrons.

      Strikes me that they may be good at maths, but they're pretty poor at logic. Might be all the quantum in their heads.

  2. Michael H.F. Wilkinson Silver badge

    Doesn't really surprise me

    I once heard an astronomer explain that it is already impossible to compute all possible binary images of 2048x2048 pixels, not because the code is hard to write, but because the 22048x2048 bit flips needed cannot be computed if the energy budget for computation is the rest energy of the visible universe, and the time available is the Hubble time (13.8 billion years or so). This was true even if you brought the cost of a bit flip down to the quantum limits posed by the quantum uncertainty. I would expect simulating the entire universe down to quantum scales requires a tad more that 22048x2048 bit flips.

    1. DainB Bronze badge

      Re: Doesn't really surprise me

      Why do you need simulate all atoms in a star few hundred LY away if all observer will be able to see is just one pixel ?

      1. Yes Me Silver badge

        Re: Doesn't really surprise me

        "Why do you need simulate all atoms in a star..."

        You don't, if all you care about is one pixel down here. But you do, if you care about how the star works internally in detail. However, it's a pretty uninteresting question in terms of what is actually possible.

    2. Aitor 1 Silver badge

      Re: Doesn't really surprise me

      Of course, but there is no need to do that in order to simulate the universe.

      It is also impossible to render all the possible images out of a 3D MMO game using all the energy of the universe, yet people play quite ok.

      So this is just bullshit, and the ppl producing it are mathematicians at heart, so they know it.

      1. TechnicalBen Silver badge
        Headmaster

        Re: Don't need to simulate that star a few billion light years away...

        But you do... as in, currently we can test to see if it matches the laws of physics. We have eyes/telescopes and other things effected on the macro scale (gravity) and quantum scale (QM effects).

        If it is not mathematically simulated, then we would get contradicting results on multiple tests/observations/detectors.

        It COULD be procedural simulation, using an algorithm to pick out the number of photons from a stars light. In the same way the original Elite did on just a few K of data and computation... except one big problem.

        Quantum Mechanics and the Bell's inequality prove that this cannot be the case for our universe (but could for others, for example those in the computer game Elite ;) ). The actual results are given at that time, and cannot be pre-calculated.

        1. defiler Silver badge

          Re: Don't need to simulate that star a few billion light years away...

          Is this the bit where I say "Right on, Commander"?

          1. CrazyOldCatMan Silver badge

            Re: Don't need to simulate that star a few billion light years away...

            Is this the bit where I say "Right on, Commander"?

            I prefer "smoke me a kipper, I'll be back for breakfast"..

    3. John Smith 19 Gold badge
      Unhappy

      "already impossible to compute all possible binary images of 2048x2048 pixels,"

      Which is really not very big at all, especially if they meant a pure 1 bit per pixel image. That would be 1/2 MB.

      Which is sort of depressing

  3. malle-herbert Silver badge
    Facepalm

    Really ?

    If this universe is a simulation, who's to say that the physics in this simulated universe is the same as in the world our simulation is running in ?

    1. Christian Berger Silver badge

      As well as...

      our universe being in the same order of magnitude of space than the universe the computer is in.

      This is simply an undecidable problem.

      1. TechnicalBen Silver badge
        Thumb Up

        Re: As well as...

        The usual computational questions might give some answer. Such as the NP/P problems and such. We can give certain logical statements that are true/false for example. Such as "at least one universe exists". :P

    2. Mark 85 Silver badge

      Re: Really ?

      I remember an SF story along these lines (Asimov? Heinlein? Or someone else, not sure) where the entire universe was a young persons simulation in a glass globe for a school science project. Sometimes I wonder at "what is the reality here?" but not often as it boggles my mind.

  4. d3rrial

    So basically...

    ... the simulator needs to be in a universe which has either solved entropy or follows different laws of physics?

    1. Lysenko

      Re: So basically...

      If the Universe is simulated the the laws of physics could be as well so entropy might be an artificial "problem". Once you allow the possibility of a simulation then the Heisenberg Uncertainty Principle might just be an externally injected random number generator and entropy might be a deliberate inversion of a natural tendency towards greater organization.

    2. disgustedoftunbridgewells Silver badge

      Re: So basically...

      Or the simulator works the same way computer games do - if a tree falls in the woods and nobody is around, is the destruction rendered?

      1. TechnicalBen Silver badge
        Headmaster

        Re: So basically...

        Yes it makes a noise. "A tree falling in the woods" and "rendering (it falling in the woods)" are the same things. Simply "falling" and "sound" are the same thing, objects moving.

        What you mean is, if the game is not rendering the forest, will the tree even fall? That is only knowable if you know the code. Is the tree falling a pre-determined event (in code or causality) or a random event?

        If it is random (true at the event unknowable noise), then you never knew it was going to fall, as the forest was not rendered/calculated/simulated. It never falls and you see a tree standing up.

        If you are using a pseudo random number generator (or a known algorithm for the random effects) and know the seed, then you can always recreate the event after the fact, without it needing to "fall".

        If it is pre-determined and you do not render it, how do you then know where to put the tree after the fact? Our universe see no distinction between "observed" and "not-observed" for ensembles such as trees (though this changes for individual quanta ;) ).

        We could retry the question as "if we see a tree lying on the floor, did it fall, or did it appear spontaneously there?" That is a better question to ask.

        1. DainB Bronze badge

          Re: So basically...

          Sound is an interpretation of observer, not a property of a falling object.

          1. disgustedoftunbridgewells Silver badge

            Re: So basically...

            Why play back the MP3 * (or whatever) of the free falling when there's nobody to hear it?

            * randomly selected from a group of MP3's so you don't hear the same sound every time, like that single creaking door sound they use on American TV.

  5. Anonymous Coward
    Anonymous Coward

    Does the fact that we can't simulate the universe not prove that we live in a simulation because if we could then it would be simulations all the way down?

    1. TechnicalBen Silver badge

      It may...

      Prove Turing correct...

  6. Anonymous Coward
    Anonymous Coward

    If we are in a simulation then it is generally taken to imply something must be running the simulation. That leads to recursion - as at each higher level something must be running that particular show.

    Our existence on the other hand is a simulation. Ourselves, our senses, and the universe are an interaction of some elementary physics that produces the effect of us existing as entities at a particular macro level of perception.

    1. DropBear Silver badge

      "If we are in a simulation then it is generally taken to imply something must be running the simulation. That leads to recursion - as at each higher level something must be running that particular show."

      I don't see how that follows at all. That would be true only if we were to posit that existence is only possible in a simulation - where indeed someone must be running a higher level one then for each level - but I don't see why we'd need to go that way. Why would it not be possible for an un-simulated form of existence to simulate another?!?

      1. Anonymous Coward
        Anonymous Coward

        "Why would it not be possible for an un-simulated form of existence to simulate another?!?"

        At which level the question is still "who/what created that creator?".

        The religious would say that their ineffable god "has always existed".

        The Universe could be said to have "always existed" in some ineffable form of time and space - and the need for a recursive god/creator becomes unnecessary by Occam's Razor.

      2. This post has been deleted by its author

  7. ShelLuser

    What if it isn't a simulation but reality?

    Einstein once said that "The micro cosmos is the macro cosmos". What if that were true? It would basically refer to the fact that we're not inside a simulation but that our universe is part of something much bigger, something not easily comprehensible.

    We already know that although the knowledge about atoms and electrons is mostly correct there's much to them than simply a nucleus and electrons spinning around them, thanks to the expanding knowledge on quantum mechanics. We also know that some things cannot be destroyed, energy can't just "disappear" for example. If you burn a piece of paper then all the basic elements which made up the material are still there, but in a completely different form. Effectively maintaining those basis aspects which made up for the piece of paper or which were 'inside' them.

    If we are part of something bigger, why can't it be something which is pretty common in the grand scale of things? So basically theorizing that we, in comparison, are so small that our universe can't "just" be destroyed by any external events.

    1. Kane Silver badge
      Boffin

      Re: What if it isn't a simulation but reality?

      "It would basically refer to the fact that we're not inside a simulation but that our universe is part of something much bigger, something not easily comprehensible."

      Obligatory Discworld quote:

      OH, I'VE SEEN THE INFINITE, IT'S...NOTHING SPECIAL.

      "Don't be daft, you can't see the infinite, it's...infinite!"

      I HAVE.

      "Alright then, what did it look like?"

      IT's BLUE.

      "It's black".

      IT's BLUE

      "It's black".

      FROM THE OUTSIDE, IT'S BLUE, BELIEVE ME.

      - Soul Music (Animated)

      .

      .

      .

      .

      .

      .

      We gonna get our Pterry icon, or what El Reg?

      1. tekHedd

        Iconage

        "We gonna get our Pterry icon, or what El Reg?"

        We have a beer icon. That works nicely.

        Death /holding/ a beer, well, that would cover it 100% of course. And be appropriate for El Reg as well.

    2. Doctor Syntax Silver badge

      Re: What if it isn't a simulation but reality?

      "It would basically refer to the fact that ... our universe is part of something much bigger, something not easily comprehensible."

      I don't know about you but I find the universe itself not easily comprehensible let alone anything bigger.

    3. Francis Boyle Silver badge

      Citation needed

      Ninety per cent of all quotations on the internet attributed to Einstein were in fact Churchill misquoting Mark Twain.

  8. Filippo

    You don't run a simulation of every single electron. You run a gross approximation that's computationally cheap and good enough for the simulated humans' senses, and you only run the very fine simulations when you detect that some of the sims are performing quantum physics experiments.

    1. Matthew Taylor

      By that metric, one could argue that Cern scientists are actually endangering the universe. If their particle probings cause the simulator of our universe to have to work extra hard, whoever is running the simulation might decide our universe is over budget, processing-wise, and shut it down.

      1. disgustedoftunbridgewells Silver badge

        If the framerate drops, nobody inside the simulation would notice - only outside observers.

        I did hear an amusing theory that people sleep for 1/3rd of the day because the simulator can't keep up with everybody being awake at once.

        1. Matthew Taylor

          Or perhaps it can, but with a much coarser simulation fidelity (lack of sleep causes groggy, muddle-brained behaviour) In a similar vein, the greater the world population gets, the less simulation time is available per person, so the stupider we get. :-)

      2. disgustedoftunbridgewells Silver badge

        Actually I've thought about this more - pervasive CCTV is arguably more dangerous to the simulation, assuming it only does full on calculations for things that are directly observed, with lazy physics for everything else.

        Eg: You don't model each subatomic particle in a tennis ball to calculate how it bounces off a wall. It's a predictable behaviour that we already model in computer games in the abstract.

      3. tekHedd

        "If their particle probings cause the simulator of our universe to have to work extra hard, whoever is running the simulation might decide our universe is over budget, processing-wise, and shut it down."

        Who is to say this has not already happened, perhaps more than once?

    2. Doctor Syntax Silver badge

      s/some of the sims/the sim

      What makes you think there's more than you?

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019