Philosopher describes the Simulation Argument

Discuss

119 Responses to “Philosopher describes the Simulation Argument”

  1. Palefire says:

    Sounds like a resurrection of Descartes Evil Genius Argument on some level.  However, if I recall correctly, Descartes utilimately defeated the Sense Doubt, Dream Doubt and Evil Genius arguments with the Cogito, “I think, therefore, I am.”

    Of course, I could be totally wrong in how I’m understand this.

    • Yavar says:

      You have misunderstood; but no worries. Cogito Ergo Sum is true even in the face of the Doubts. Descartes however, only defeated his Doubts (Sense, Dream and Evil) with: “God is by definition good, so he wouldn’t deceive me”.  

      The Light of God is unfortunately Descartes’ solution to his skeptical epidemiological exercise in The Meditations.

    • CLAVDIVS says:

      Yeah, but I seem to recall an unchallenged assumption that a benevolent god exists somewhere in that chain.

      EDIT: And I shoulda kept reading. Someone else covered that.

    • Lobster says:

      But the Cogito doesn’t solve the problem; you can exist and still be deceived.  In fact you MUST exist to be deceived, which is the basis of the Cogito entirely, that the only thing of which you can be sure is that you “are.” 

      Descartes solved the deceit problem by saying that God is good and would not allow him to be deceived.

      Not terribly satisfying.

    • Sam Tope says:

      Cogito ergo sum does not ‘defeat’ any of those possibilities.  It was a conclusion he came to when examining the full extent of what we can know to be unfailingly true. 
      His final response to his original triad of doubts was his rather circular ontological argument, wherein god was perfect by definition, existed because perfection requires existance, and a perfect god would not decieve us in such a manner.

      Cogito ergo sum is no less applicable to the hypothesized simulation scenario, but all we can take from it is that we are at worst an independant equivalent of a self contained complex subroutine of the simulation.

  2. Raederle says:

    The “religion” Iain Banks postulated for the Mercatoria in his novel The Algebraist is similar.  Called simply The Truth, it holds that if some percentage of the total population of the universe come to believe in their heart of hearts that we’re living in a huge simulation the people running it will be forced to take some action.  Not necessarily a positive action from the perspective of the people inside the simulation, but it would at least be an acknowledgement of the situation.

  3. Jer_00 says:

    I haven’t had a chance to listen yet, but I think there’s an option in the traditional formulation that’s missing above.  Namely “it’s actually not possible to create a simulation so advanced that we could be living in a simulation.”  That may be a subset of what you have as (2) above, but I think there’s a distinctly different argument to be had between “not interested in” and “not capable of” in this respect.

    • Mark_Frauenfelder says:

      I agree with you Jeremy. The comments at Philosophy Bites discuss this option.

    • Brian Bishop says:

      I think the “civilizations are not capable of creating advanced simulations” option could be encapsulated within option #1.

    • Joshua Ochs says:

      In particular, there are physical limits to data and processing density at some point. For instance, suppose we can perfectly harness every quark for storage and quantum effects for processing. Doesn’t it follow that it would take more than a single quark to simulate a quark? From that basis, any given civilization could only simulate realities of (significantly) lesser complexity than its own.

      I haven’t read what the exact details are of an “advanced civilization” and “advanced simulation”, but there are definitely limits involved.

    • First Last says:

      There is also no “no civilisation has progressed far enough to invent the technology required for these advanced simulations.” Not as in die out before they invent it, but just that the real universe is physically not old enough for any existing civilisation to yet have that ability.

  4. hassenpfeffer says:

    If we’re in a simulation it’s even more upsetting to imagine a cruel operator who DOESN’T flip the switch to eliminate pain and suffering for all creatures therein.

    I’ll be back to this thread after going home, tending to the “herb garden,” and rereading a bunch of Phil Dick.

    • It’s not a given that the simulation is about us . Maybe we’re just some static noise…

      • hassenpfeffer says:

        Fair enough. I can settle for being kipple.

      • vonbobo says:

        “It’s not a given that the simulation is about us . Maybe we’re just some static noise… ”

        This is hillarious, and poignant. Humans need to stop pretending we are special, maybe we could get along better.

    • janusnode says:

      It’s not _real_ suffering. An operator of our simulated world would no more flip the switch because we were suffering than we would stop playing ‘Call of Duty’ because too many people were getting shot.

      • l337n00b says:

        If we are simulations then we can still really suffer.  Comparing a simulated person in the kind of complex simulation we are talking about here to a character in Call of Duty doesn’t work.

        Call of Duty does have something of an AI that controls the moving parts of the characters.  Odds are it is not sentient, but imagine that video games get to such an advanced level that the AIs controlling them become sentient.  What would the motives and feelings of such a being be?

        First of all, the sentience may not even exist in each enemy character, but may exist instead collectively between them or as the entire game.  Killing an individual “person” in the game is not necessarily harming or killing an AI.

        Second, the AI did not evolve the way that organic beings did, and so there was no evolutionary pressure to live long enough to reproduce.  Without that, there is no reason why suffering and trying to avoid death to ever come about.

        Third, the AI does not even have evolutionary pressure to “win” the game.  If we had powerful enough computers to make AIs that could play the bad guys in Call of Duty really well and we started putting them through an evolutionary process to make them better and better, there is a good chance they would quickly become too good and players would not be able to play the game anymore.  Because of that, the pressure on AIs (pressure created by the humans who program them) would be to be good enough at the game to create an enjoyable experience for the player, which ultimately means their goal would be to lose in the end.

        Therefore, even if Call of Duty is sentient, there is a very good chance that it wants us to win.  It is the Dungeon Master – trying to make things exciting enough that everyone can have a good time collectively telling a story, rather than simply crushing an opponent that has no chance.  The AI may be incapable of suffering, or may suffer only when the player frustratingly attempts the same task over and over, being incapable of winning (since this would be the condition that triggered the death of it’s less fortunate cousins who were too good at the game).

        If *we* are a simulation then we are obviously a simulation of a species that evolved through sexual reproduction.  We suffer greatly.  Assuming sufficient computational power, we should be able to simulate beings that have some level of self-awareness and genuinely fear their own destruction.  In that case, the operator would have every capacity to be cruel.

      • travtastic says:

        But the evil foreigners in Call of Duty aren’t artificially intelligent, as we would be if simulated.

        Let’s put it this way: if you could make a perfect, digital copy of a human being, would it be okay to torture it?

      • peterblue11 says:

        ingenious argument :)

      • John Aspinall says:

        For a counter-argument, see iain M Banks’ Culture novel Surface Detail.
        Synopsis: Hells (plural) are computational simulations, running the captured mental states of the (now deceased in the physical base world) inhabitants.  Their suffering is just as meaningful as it was when their mental state was running on its original hardware.  Because it’s the same mental state.
        So no reason to shut down ‘Call of Duty’ yet, but as the level of detail in the simulation increases, you might want to think about it.
        Or not.  Both sides are represented in the novel.

    • John Delaney says:

      If you’ve played sim city you probably understand the impulse to unleash Godzilla or an alien invasion…

    • If we’re in a simulation it’s even more upsetting to imagine a cruel operator who DOESN’T flip the switch to eliminate pain and suffering for all creatures therein.

      Neuromancer might disagree.

    • Jon Konrath says:

      That’s like saying everyone who installs Angry Birds will eventually stop playing because of the pain and suffering of the pigs.  Granted their civilization might be radically different from ours, but I’d imagine a scenario where the operator makes it more painful and cruel for the creatures inside, because that’s more entertaining.

      • hypnosifl says:

        That’s like saying everyone who installs Angry Birds will eventually stop playing because of the pain and suffering of the pigs.

        Not really, because we don’t believe the pigs are conscious, they don’t have complex simulated brains. If the simulation hypothesis is true the creators of these simulations would presumably be AIs themselves, so if they had any ethical belief that causing suffering to conscious beings is wrong, hopefully they (or most of them anyway, hopefully they would have some sort of policing of the sociopaths among them!) would avoid causing us pain just for fun. But the question of pain occurring “naturally” in the course of a “realistic” simulation might be another story, perhaps analogous to how we don’t make any attempt to alleviate the natural suffering of animals in the wild…

  5. user1234567 says:

    What if the first civilization to create advanced simulations added a rule to prevent sub-simulations and then placed everyone inside of the simulation?

  6. trefecta says:

    It’s simulation all the way down…

  7. Allan Self says:

    or number 4. – it is meaningless to talk about simulating the contents of mental phenomena when those contents are not reducible to computation.

    • The mind, like anything else with structure, is a just a pattern– a particular way of reducing the complexity of data.  You’d need computation to predict it or have it respond to new stimuli, but to replicate it, all you’d need to do would be to observe it.   If the observation was detailed enough, then that itself would be a conscious mind, indistinguishable from the original except for context.

      Of course, context is extremely important in how we humans value things– I wouldn’t treat some sort of hyper-detailed EEG scan the same as I would a person.  However, if I was certain that I was going to have every detail of my brain scanned and recorded, then I don’t think I would be able to tell at that instant whether my mind consisted of neurons or scan data. 

      Not that it would necessarily matter– what sort of action would being a simulation imply?

      You know, I’m not sure you’d really need a hypothetical civilization for this sort of “simulation”.   The universe is vastly more complex then our minds. I wonder how likely it would be for the structure of a mind at a particular instant to just occur randomly.   In the vectors of every fourth iron atom in a sun or something  (all that would really be required for it to be a pattern would be that it compressed the data).   You wouldn’t really need an infinite number of monkeys to write Shakespeare, after all.

    • John Aspinall says:

       “or number 4. – it is meaningless to talk about simulating the contents
      of mental phenomena when those contents are not reducible to
      computation”

      I regard this as a non-starter (or more pejoratively, “the WOO
      argument”) as mental phenomena are a consequence of the behaviour of
      atoms and molecules in our brains, and the behaviour of atoms and
      molecules is reducible to computation.

      If you think otherwise, (and many do, I must admit) you are making an
      extraordinary claim;  to wit:  there is something (my term is “WOO”)
      going on in our brains outside the usual laws of physics and chemistry,
      that has never been observed or characterised  in terms of how it
      departs from standard physics and chemistry.

      Extraordinary claims require extraordinary proof.  And quantum mechanics
      (a la Penrose) doesn’t provide that proof.  QM is still a piece of
      well-understood physics, to the extent that we can perform computation
      about its behaviour.  You still have to explain why regular old QM as
      seen outside the meat-between-the-ears, stops working and is replaced by
      WOO-QM.

  8. John Eikenberry says:

    I think all these sorts of ‘arguments’ fall apart when you start asking what are the differences between being in the physical universe vs being in a simulation. It usually quickly falls back to Kantian style metaphysics which, IMO, were thoroughly trounced decades ago.

    • trefecta says:

      That’s what I’ve thought since I first read the paper. Though that style of thought has just gone out of mode, perhaps it’s coming back in a new fashion?

  9. betatron says:

    I’m reading this on a METRA train inbound to chicago.  Someone spilled coffee on the guy across the isle.  A surly fat guy makes fun of my turtleneck…

  10. Brainspore says:

    The “cruel operator” theory would certainly be one way to explain a universe capable of producing a film as bad as The Matrix Revolutions.

    • Maybe the Matrix sequels HAD to go bad, because somewhere, some system operator said “Dammit, it was one thing when lone fruitcakes like Phil K. Dick and Buddha talked about simulation, but we just can’t afford to let this stuff get mixed up with bullet-time action choreography and Keanu Reeves, for Chrisakes….no, killing them wouldn’t work, lets brainwash them to make really bad sequels….and Speedracer.”

  11. digi_owl says:

    The idea of a simulation operator sounds like god to me…

  12. What if only a small number of people in the simulation are actually “people”?  Maybe 95% of the population are just cleverbots. Doesn’t it seem weird that all the Republican candidates all eventually start talking about God?

  13. Ambiguity says:

    A little off-topic perhaps, but if you like philosophy podcasts, Diet Soap is often amusing, entertaining, or informative.

  14. Stefan Jones says:

    There’s a classic SF novel — Star Maker by Olaf Stapledon, that finds a far-future multi-galactic civilization harvesting the mental powers and experience of minds across the span of time to get a glimpse of the power beyond the universe.

    SPOILERS AND IRREVERENT PARAPHRASING

    They find . . . God, I guess. A creative spirit who runs endless universe simulations, tweaking the parameters and refining things each time.

    The galactic group-mind gets a terrible bummer when it senses the great spirit looking it over, thinking “Meh, OK, but I can do better” and tossing out universe in the bin.

  15. flagler23 says:

    What really is the difference between living in a simulation and being “real”?  Ultimately everything in a simulation and sub-simulations relates back to phenomena in the underlying physical world.  The interpretation of the agent of design as a sentient, or even moral agent obscures this equivalence.  The operator itself experiences no more free will in a fixed universe than do we in the simulation.

  16. Pentashagon Pentashagon says:

    Modal realism or mathematical realism would imply that it doesn’t even matter if we’re in a simulation or a real universe because ultimately there’s no difference.  All that matters is the existence of entities and relationships between them, not what ultimately causes (or doesn’t cause) them.  For instance, given the initial conditions of the simulation and the rules used to change it over time, the entire history of the simulation is predetermined.  An evil overlord who flips a switch to torture everyone does nothing to the original simulation but only creates a new simulation using the old one as its set of initial conditions.  The original simulation still has a mathematically certain future, and much like one can prove that there are an infinite number of primes without enumerating them, the unaltered beings inside the original simulation still provably exist in the future of the original simulation.  Is it sufficient for things to exist without actually flipping bits in some gigantic computer to hold their values?  No one would argue that there are a finite number of primes simply because we can’t enumerate an infinite list, nor deny the mathematical relationships between the ones we can’t enumerate.  Who would further deny that the mathematical relationships between particles of an imaginary (but mathematically well-defined) universe don’t maintain their mathematical relationships over all time, including the relationships in any imaginary brains and bodies and societies and worlds that result from the rules?  If those relationships are mathematically fixed, aren’t the imaginary thoughts and actions just as real to the inhabitants as our existence is to us?  Are we really anything more than the result of just such an initial set of conditions and rules?  Do we need to be?

    • chenille says:

      I like your analysis. If you want to list it as an option, I’d like to suggest:

      5.There are so many “basic” universes the chances we are not inside one are vanishingly small – to the point where we would exist in one even if we are also in a simulation.

      Why Bostrom speculates about building realities inside one another but assumes there should be only a single one that counts as “physical”, I’m not sure. Usually we use physical to mean the world we live in, but if there are more, I don’t see why there must be one so privileged.

      • hypnosifl says:

        Why Bostrom speculates about building realities inside one another but assumes there should be only a single one that counts as “physical”, I’m not sure.

        I don’t think he’s really assuming this, I think the idea is that most simulations are restricted in time, so they don’t bother simulating past a certain date. If simulations lasted indefinitely so that simulated civilizations could spread out throughout their simulated universes, then the same doomsday argument I discussed above, which says we should expect that an “average” mind in the higher-level universe would likely have a much higher birth order than our own (assuming the fraction of civilizations that experience a singularity and spread throughout space is not vanishingly small) would also suggest an “average” mind in a simulation (running in the higher-level universe) would have a much higher birth order than our own, so in this case the simulation argument would be of no help in “explaining” our relatively small birth order. It’s only if you assume a lot of simulations get cut off before more than a few hundred billion minds are born within them that the simulation argument can help with this puzzle. 

  17. xzzy says:

    “They really hate it when you do this!”

    *Raul grabs his universe farm and shakes it up*

    Really, that’s what the big bang actually was.

  18. lknope says:

    This reminds me of Geoffrey Miller’s Runaway consumerism
    explains the Fermi Paradox. (Partway down the page here:  http://www.edge.org/q2006/q06_9.html)

    Basically, the argument is that the reason we haven’t seen any evidence of extra-terrestrial intelligent life is because they are too busy playing video games to explore space.  I think from his essay you could say that he would think 2 is true.  

    But if there are different civilizations and they advance at different rates, more than one of the statements could be true at once.  Some civilizations blow themselves up or go extinct, some create advanced simulations and others choose not to.

    • AnthonyC says:

      No, they idea is that if even a few do, they within the simulations of those few, a few *more* do, and so on, then the # of simulations will soon vastly exceed 1, the number of physical universes we started with.

  19. Frank W says:

    The level of sensory input I’m getting is nothing that can’t be built in another hundred years.

  20. hypnosifl says:

    As I understand it, part of Bostrom’s argument is related to the Doomsday argument which says we should most probably expect ourselves to be a fairly “average” member of the set of all intelligent beings throughout spacetime. Only about 100 billion humans have existed on Earth so far (which leads to the kind of mind-blowing observation that about 6% of all humans who have ever existed are alive today, see here), so the conclusion would be that most intelligent beings are part of civilizations that don’t last long enough to produce vastly more than a few hundred billion beings before dying out or experiencing a huge reduction in population (presumably due to a permanent collapse of technological civilization). And yet there seems no basic physical obstacle to the idea of a civilization that uploads their minds into computers (or just creates new AI beings), then spreads throughout the universe in self-replicating probes that convert large amounts the heavier elements in different planetary systems into computational devices which can house more simulated minds…if that was the case the number of intelligent beings in this civilization would be vast (quadrillions? septillions?) and the vast majority of members of this civilization would have a much higher “birth order”. So if the principle of averageness (which Bostrom calls the “self-sampling assumption”) is correct, either it’s astronomically improbable for any civilization to experience a technological singularity and spread throughout the galaxy like this (so that the increased population of the few that do is balanced out by the combined populations of all the ones that don’t), or else our perception of our “birth order” is actually incorrect, and we are part of such a old and vastly populous civilization, but we just don’t know it because a large proportion of minds birthed by such civilizations are in “ancestor simulations” who think they are part of a relatively young civilization.

    • JakeHamilton says:

      I’ve read about the Doomsday Argument, and am sort of at a loss to explain how it could possibly square with any sort of realist metaphysics. What is this “we” which preexists and can expect something about itself apart from its temporally specific, individual identity? Are the souls of all the sapient entities that will ever be born or made waiting out in the aether to be incarnated, precisely identical, being without contingent attributes, yet still mysteriously distinct and quantifiable, such that each should “expect” to be dropped into an average point in the history of sapients in the universe? And if the Doomsday Argument holds, does it not hold for every population of sapients present (apparently) early in its species’ history?

      As far as I can tell, its proponents seem to be making the rather incredible proposition that a probabilistic calculation based on Kantian dualist metaphysics constrains, without apparent causal mechanism, the history of intelligent species in general, so as to arbitrarily shorten their lifespan.

      Christ, at least the mathematical Platonism+many-worlds crowd have gotten their ontology as far as Hegel.

      • hypnosifl says:

        Jake, the self-sampling assumption isn’t a metaphysical claim about some kind of bare awareness randomly “choosing” a given observer to determine what the awareness’ conscious experiences will be (although I would disagree that idealism is necessarily incompatible with “realism” or even naturalism, look at David Chalmers’ speculations or the many-minds interpretation of quantum mechanics), it’s just a claim that if a lot of observers reason this way, most of them will reach correct conclusions. Look at this thought-experiment from Bostrom’s FAQ:

        A firm plan was formed to rear humans in two batches: the first batch to be of three humans of one sex, the second of five thousand of the other sex. The plan called for rearing the first batch in one century. Many centuries later, the five thousand humans of the other sex would be reared. Imagine that you learn you’re one of the humans in question. You don’t know which centuries the plan specified, but you are aware of being female. You very reasonably conclude that the large batch was to be female, almost certainly. If adopted by every human in the experiment, the policy of betting that the large batch was of the same sex as oneself would yield only three failures and five thousand successes. … [Y]ou mustn’t say: ‘My genes are female, so I have to observe myself to be female, no matter whether the female batch was to be small or large. Hence I can have no special reason for believing it was to be large.’ (Leslie 1996, pp. 222-23)

        If you had been part of this experiment, and the fiendish experimenter forced you to bet your life on whether the larger batch was male or female, are you saying you would see no reason to bet the larger batch was the same sex as you, as opposed to the opposite sex? After all, if everyone bets the larger batch is the same sex as themselves, only 3 will lose their life while 5000 will be spared, so it seems like a pretty smart bet.

        Also, it seems to me that in any multiverse scenario we sort of have to use this sort of reasoning to make any scientific statements about probabilities…given quantum randomness, somewhere in the multiverse there will be a version of Earth where the observed statistics depart wildly from those predicted by the true laws of nature, or they conform to the laws of nature up until a certain date and then depart afterwards…assuming you don’t completely discount multiverse theories, what makes you think the statistics of observed experiments reflect the actual laws of nature, or that you can make predictions about what will be seen in future experiments, if not some version of the self-sampling assumption? After all the inhabitants of these statistically anomalous Earths are just as real as any other Earth, without the self-sampling assumption I don’t see any reason to confidently claim we aren’t living on one of these anomalous Earths.

        • Pentashagon Pentashagon says:

          So long as the correct information is properly hidden, the doomsday argument and the sex batch experiment have their simple answers.  For instance, the experimenters could have just picked 5003 humans at random and asked each whether they thought they were in the first 5% or last 95% of the group to be born, and killed the ones who said 5%, making it roughly equivalent to the doomsday argument.

          The real world problem with the doomsday argument is that information about the experiments leaks into the measurements the experiments can make.  Birth order should be an obvious weakness in the argument; there is (nearly) full information about the birth order of everyone alive today, and also an approximate number of people who have lived and died before us.  From this, one can calculate that he or she is the Nth person to ever live, out of a current total of M.  The probability, P(X), that X is the total number of humans who will ever live is 0 for X<M, and the sum of P(X) for X from M to infinity must be 1.  The cumulative probability summed from M+1 to infinity can be calculated relatively easily by observation.  If there is currently a pregnant woman in the process of giving birth, the sum of P(X) from M+1 to infinity is very close to 1.  If a huge asteroid is seconds from obliterating the earth, the sum is very close to 0.  Similarly, the sum of P(X) from M+J to infinity for arbitrary J can be calculated based on estimations of current pregnancies, population growth, natural disasters, the eventual likely states of the universe, etc.  From this, the probability of being in the first L% of humans who ever lived is the sum of P(X) from (100/L)*N to infinity.  The doomsday argument makes some simplistic assumptions about what P(X) has to look like.  In reality, we have far more information.

          • hypnosifl says:

            Bostrom’s version of the Doomsday argument does take into account the issue of information (like knowing a huge unstoppable asteroid is about to collide with the Earth)–Bostrom uses Bayesian reasoning, so you start with a prior probability distribution that intelligence will last for a given amount of time based on historical information and plausible predictions about the future but ignoring your own birth order, and then you use your birth order to update the Bayesian prior. See the discussion on this page, particularly the paragraph beginning “So the modern form of the Doomsday Argument, which was formalized by Bostrom, is the Bayesian form…”

          • Pentashagon Pentashagon says:

            Say you have a Petri dish and a small bacterial colony.  There are N bacteriums, and the nutrition in the dish is sufficient to grow a total of M.  Say N is 2^8 and M is 2^30.  Barring advanced, rapid evolution the bacteria will never number more than M, and will most likely be slightly less due to waste products, but not significantly (perhaps half, for a properly designed dish.  I’m not a microbiologist).  It is virtually deterministic how many bacteria will exist in the dish regardless of what arguments an individual bacterium might use to estimate its chances of being in the last 95%.  The original colony will grow exponentially and produce dozens of generations who can answer the 95% doomsday argument wrong.  The very first (anthropomorphic) bacterium could calculate its energy needs and measure the dimensions of its Petri dish and the available nutrients and have a nearly 100% chance of answering the doomsday argument correctly despite being in the first 9e-08 percent of bacteria.

            Now look outside at night and tell me how many humans will fit in this little Petri dish called the Universe.  My descendants will be legion.

            A stronger doomsday argument could be constructed by considering the vast number of species that have gone extinct compared to the number of species alive today.  I have no doubt that genetically pure humans will disappear relatively quickly on cosmic timescales, but I think that our descendants will consider themselves “human” or at least rational beings subject to the doomsday argument for all practical purposes.

          • hypnosifl says:

            Do you really think it’s a sort of deterministic certainty that your “descendants will be legion”? You don’t think there’s any significant chance that humans (or posthuman intelligences) will be wiped out on Earth before establishing totally self-sufficient industrial bases on other worlds? Or even a significant chance our civilization will collapse to preindustrial levels before then? (I have read that since we have already mined the most easily available sources of various resources, it’s unlikely there could be a new industrial revolution following such a collapse). There are a lot of ways such a thing could happen (many discussed in astronomer Martin Rees’ Our Final Hour), from global nuclear war to a really catastrophic environmental disaster to a genetically engineered disease epidemic to a nanotech “gray goo” scenario.

            In your petri dish analogy, suppose the bacteria know that some bacteria populations in petri dishes are allowed to grow to consume all available resources, but in some dishes human experimenters add antibiotics or other factors which destroy or decimate the bacteria population before they can reach these levels (they don’t know the exact percentages for each case). I would think that in this case, if all bacteria tried to guess the probability of reaching a large population by first picking a plausible prior and then updating using their own observed “birth order” within their own petri dish, they would likely end up with better estimates than if they just stuck with whatever prior seemed initially plausible to them.

  21. cymk says:

    If we can simulate the forming of the milky way, http://www.youtube.com/watch?v=VQBzdcFkB7w, just how far off are we from simulating entire worlds or civilizations?

    • Lane Yarbrough says:

      “simulate the forming of the milky way”
      Interesting. Let’s replace Milky Way with any number, let’s use 100. There are many ways of coming up with 100 using numbers found within 100. Now add numbers outside of 100 and you have infinite possibilities. Now let’s use a prime number, 7. Still two possibilities within 7 exists. 

      Now grab a bag of 7 marbles and dump them out. Take accurate measurements of their position. Now create a simulation so that you can dump out the marbles into the EXACT same pattern. Think of all that is involved. Did you put the marbles in the bag exactly like before? Did you measure the rate of your pour? Is the humidity the same level? You have a better chance of creating a new universe. 

      • John Aspinall says:

        Asking to simulate the EXACT same pattern is invoking the Impossible Conditions Fallacy.
        (http://theautonomist.com/aaphp/permanent/fallacies.php#impossiblecon).
        If we modify your argument so that you have to specify the degree of accuracy to which you will, or can, compare the simulation result to the original, and then cymk responds with the accuracy with which he must measure the initial conditions, and the budget for his computation, then his argument survives just fine.
        Remember, in the simulations we are talking about we have no “gold standard, measured to infinite precision” description of “what should have happened”.  Instead we only have the inhabitants’ estimates of “yes, that seems consistent with previous observations of my world”.
        If the 7 marbles hit the floor and stay there, I’m not going to give them a second glance.

    • Spriggan_Prime says:

      So we’re all just purple stuff swirl in the toilet bowl?

      Pretty, but with 8 months couldn’t they have come up with a better sound track?
      Bleep bloop bloop beep.

  22. anansi133 says:

    I would tend to believe that (2) was the most likely, but then I remember the tendency for economic bubbles to arise and then pop. Planning anything for any sort of future must necessarily involve some sort of simulation. I suppose for some, the map is so much better than the territory that they’d rather remain in the map.

    Reading about a tasp made me wonder if I’d rather live in a painful reality or just become a wirehead.To choose reality, one must have reason to think that it offers something that the comfy chair cannot.

    This kind of theory could actually go someplace interesting, if “simulation” could have a percentage value. If it truly were all or nothing than there would never be an economic crash.

    • been thinking about this stuff for a while, personally. IMO, the human brain already is the best possible universe simulator known to exist. that the “you” within it is lost when it stops functioning is already a given, so it should be obvious, yet not everyone seems to make the connection. our bodies already are the sufficiently advanced civilizations created by RNA and DNA to better represent, inerpret, and predict the state of the universe…

      this reaffirms why most people already live further in their head than in reality; unless they’re buddhists most people are born and bred to be “wireheads” and never know there is a territory outside their map. the stuff this guy is talking about is the map of the map, or baudrillard’s “3rd level” of simulation. in that there is a “we”, it is already simulated, so you’re not really replacing the territory with the map (which you couldn’t do if you wanted to), but one map with another.

      i think the reason noone else really pushes this viewpoint is that the implications are both too common sense to ignore and too fantastical to comprehend at the same time.

  23. JakeHamilton says:

    Bostrom is clever, I suppose, but in any serious terms his “argument” is fairly ridiculous. He presupposes, as a matter of course, that we currently have a sufficiently complete understanding of the mind, its relationship to the brain, the metaphysics of computational simulation, and a dozen other things, to be able to throw out these massively speculative, thinly defined statements as if they were determinate propositions from which he can make confident deductions about future knowledge and events. This is especially painful when the man is billed as a “philosopher.” This is worldbuilding for the setting of a philosophical hard scifi novel…not an actual philosophically rigorous line of reasoning.

    Earth to Bostrom: Aquinas covered “how many angels can dance on the head of a pin” a few centuries ago. Also, you should read Popper on the historicist fallacy.

  24. Cicada Mania says:

    I would like to thank my overlords for creating ceramic panther lamps and placing them randomly at garage sales and swap meets… also, thanks for Martin Denny albums, the new Bleach Resurrection video game, Korg keyboards and whiskey. No thanks for hurricane Irene. Please add the ability to fly and “shoot waves of destructive energy from a magic katana” to my simulation. Thank you.

  25. SCAQTony says:

    First off we CAN do advanced simulations from plotting weather models to “747 pilot simulations.”

    His cold & heartless simulation “idea” as been proffered before – please compare:
    “Remember that thou art pixels, and to pixels thou shalt return.” — Nick Bostrum”Remember that thou art dust, and to dust thou shalt return.” — Genesis 3:19

    • Lane Yarbrough says:

      But aren’t those based on scenarios, like a fighting a forest fire? The winds often change. How many times have we heard, “We didn’t perceive that happening”.

      So it’s perception perhaps? Simulations take great care in setting up, plotting and documenting. Still, the 747 pilot will experience something that was never a part of his simulation training. 

      There’s too much going on in the world, we are decision makers and can bob and weave with what life throws at us. A true simulation takes place in a severe controlled environment, which the world is not.

      It’s about acclimating.

    • SCAQTony says:

      We have created physics simulations on the computer that respond to their presets with extraordinary precision. I mentioned a 747 Sim. I will concede that weather models was a bit to ambitious but needless to say, we have created simulations and Bostrum’s simulation analogy does not wash since how can a simulation know it is a simulation? It’s like a “Call of Duty” video game soldier suddenly getting PTSD and refusing to fight.

  26. Recluse says:

    Brian Greene did a nice explanation and discussion of the Simulation Hypothesis in his book: “The Hidden Reality”. There was also an interesting novel called “The God Hater” by Bill Myers http://www.thegodhater.com/book.html  which attempts to leverage the Simulation Hypothesis as a conduit to some (IMHO unconvincing) Christian propaganda. 

    I think I want a T-shirt that says: “We are all NPC’s”

  27. Singularité says:

    “As I understand it, one of the following three statements must be true:”

    “1. Civilizations go extinct before they are able to create advanced simulations.”

    Nothing dies, nothing exists. Informations lives. Individuals, civilisations never exists.

    “2. Advanced civilizations are not interested in creating advanced simulations.”

    Simulation is real, even if reality is real. Advanced civilizations are not interested in RE-creating advanced simulations.

    “3. There are so many advanced simulations that is is far more
    likely that we are inside a simulation than in the physical universe.”

    We are the simulation of ourselfs. And maybe we are the simulation or the puppets of other civilisations.

    The only way to explain life on this planet is 1) a game 2) other civilisations just don’t care at all : even if we are not in a computer.

    http://tumblr.9gag.com/post/522877118

    I hope one day other people will understand fuzzy logic and complex thinking

  28. MrEricSir says:

    Just like how so many folks in the 50′s thought alien abductions were real, this seems like another case of rationalizing a delusion based on reading a little too much science fiction.

  29. atimoshenko says:

    How about:
    4. Few civilisations ever emerge.

    Also, it is not an issue of creating advanced simulations per se, but of creating persistent advanced simulations in which all or almost all sentient participants are actually unaware that they are within a simulation.

  30. ZikZak says:

    Perhaps that explains some of the seemingly inconsistent and bizarre sub-atomic phenomena we’re observing as we increase the resolution of our observations.  Maybe our laws of physics are just a quick and sloppy implementation of the real thing.  Or bugs in the simulator?

  31. To paraphrase Wittgenstein, we speculate about these things because we can. They are meaningless mental constructs beyond that. Mind Fucking. Bottom line: It doesn’t matter… Put another way, this kind of investigation seeks to provide a window to the meaning of existence. If existence has no walls or doors, why are windows even relevant?

  32. Bobby Martin says:

    4.  Consciousness (self-awareness) is not simulable.

    I don’t believe that, but I know perfectly intelligent people who do.  It deserves to be on the list.

  33. Daemonworks says:

    God is a 14 year old girl with a really high-end version of The Sims.

  34. politician says:

    I suppose it all depends on what is encompassed in the simulation. At it’s barest level, a simulation need only be effective for a single observer/simulant (ie. all of the “others” can be just projections conjured to bolster the believability of that individual’s world to that individual)- so it can be fairly economized for whatever that singular individual is experiencing locally (including percieved memory recollection and semiconscious states, or dreaming).

    So there doesn’t have to be a “we”, just a single engineer who can be the sole recipient/participant of the simulation.

    Anyhow, all the above stolen from an old Twilight Zone episode, I believe.

    • surreality says:

      Yep. One of the early ones. Man finds himself on completely abandoned Earth; when he wakes up we see he’s been in a training simulation since he’s going to space and will be alone for a long time. I think the episode is aptly named “Where Is Everybody?”

  35. codesuidae says:

    The argument is invalid because it projects the rules of our universe outside it’s context.

    If this universe is a simulation, then the concepts that exist within it do not necessarily apply to whatever it is that is running the simulation, including the idea of universes, civilizations, simulations, thinking beings, physics, information theory, math, etc.

    Similarly, if we create a simulation of a universe, while we have various limits on our ability to manage the complexity of the simulation (size, speed, consistency, etc), the experience of that reality to a being within it would not necessarily need to reflect these limits (excepting perhaps as the workarounds we might use change the nature of that experience).

  36. vonbobo says:

    I’m actually more concerned with fractals. We (our “known universe”) could actually be a germ on the tip of a child’s finger, ignorantly waiting for the antibacterial soap to wipe us all out.

  37. Lane Yarbrough says:

    I don’t like it one bit, it’s statements like this that create cults and religious falsifications. It’s a riddle meant to confuse; people who are confused are easy to lead. Remove all the subjectivity. 

    1. “Simulation” is ambiguous, what kind of “simulation” are you discussing. Isn’t “acting” a form of “simulation?”

    2. And who decides what “an advanced simulation” is? 

    3. “Advanced” isn’t a valuable word in technology, it’s a moving benchmark, like “perfection” or “infinity”. (If you told someone five years ago they would be able to by a 64g thumb drive with a 5 year warranty, they would have taken your car keys away)

    4. It all comes down to MATH, and math is never-ending, there will always be an undiscovered equation. 

    But I’ll humor you…What if the Simulation itself is a Simulation of a Simulation……?

    END SIMULATION

    • dnebdal says:

      A “simulation” in this sense is something like “a system that built within one set of rules that lets you see what would happen given a set of rules that might be different”. As a simple example, a computer program that lets you play with newtonian physics is a simulation -  e.g. one where you can place objects with different masses/positions/speeds and then see what would happen as time passes. These exist, and are used for things like physics classes – and testing engineering designs. Given the context here, this (rather common) meaning of the word is the only one that makes sense.

      The point of the thought experiment is that it’s theoretically possible to make such a simulation so detailed and large that it could cover the entire observed-by-us universe and all the interactions in it – and again, it should be obvious from context that this is what’s meant by it being “advanced enough”.

      As for math being never-ending …  that’s a claim I’d like to see substantiated, given that it could equally well be said that math is the study of all the things that follow from a limited starting set of axioms. It’s also kind of irrelevant to claims about the nature of the physical universe.

      And indeed, what if? If you start with a simulated universe, it should be possible to simulate a comparable universe within it, and another one inside that, and so on and forth. It seems like each nested simulation would have to be more limited, given the overhead of running such a simulation. If we assume that there’s a minimum amount of matter needed in a (possibly simulated) universe before it can produce such a simulation, then the maximum number of nested levels will be finite as long as the amount of matter dedicated to the outermost simulation is.

      The big problem with that argument is that it depends on matter and information storage working roughly like it does for us. If there is an outer, non-simulated universe that operates on completely different rules, then it’s impossible to reason about it.

    • DavidPursel says:

      Yes! Well stated, Lane.

      –> Begin Edit

      Also, an additional statement is needed:

      4. The following proposition is false:

      “…one of the following three statements must be true:

      1. Civilizations go extinct before they are able to create advanced simulations.

      2. Advanced civilizations are not interested in creating advanced simulations.

      3. There are so many advanced simulations that is is far more likely that we are inside a simulation than in the physical universe.”

      Yes, Nick, you are vastly oversimplifying things here.

      End Edit <–

      • hypnosifl says:

        The version you call “oversimplified” is not really Nick’s argument, it’s Mark’s imperfect summary of the argument. Nick Bostrom is a philosopher who has done a lot of work on anthropic reasoning, and his actual argument depends critically on this sort of reasoning–did you read my earlier comment on the relation of the simulation argument to the doomsday argument? And for a more technical version you could also read his own original paper at simulation-argument.com (though to follow it you probably need to have read some of the material at anthropic-principle.com for background on the doomsday argument and Bostrom’s “self-selection principle”) A lot of commenters on this thread seem to be responding to Mark’s summary as if it really is the entirety of Bostrom’s argument, seemingly completely unaware of the key importance of the doomsday argument and the self-selection principle.

  38. kmoser says:

    This explains the occasional BSOD I see in the mirror.

  39. OgilvyTheAstronomer says:

    I guess the difference between “reality” and “simulation” from the point of view of the simulated is that a simulation is subject to arbitrary intervention, and reality isn’t. The sudden onset of phenomena that run counter to established physical laws would lend credence to the simulation argument. Until that happens, it doesn’t make a blind bit of difference whether we’re simulated or not. To quote Burroughs:

    “If life is illusion, then I am no less an illusion, and being thus, the illusion is real to me.”

    • codesuidae says:

      “the difference between “reality” and “simulation” from the point of view of the simulated is that a simulation is subject to arbitrary intervention, and reality isn’t”

      Possibly, yes, but not necessarily. A simulation could potentially be built in such a way as to prohibit arbitrary intervention (or to make it too difficult to realistically accomplish). 

      Also, the simulation need not be a simulation of /our/ reality, it could be a simulation of a completely different physical system, which would then be quite different.

  40. Pag says:

    Can’t it be a mix of all three options? Most civilizations never become advanced enough to build such simulation, many other civilizations have no interest in creating them, some civilizations do build them. If this is true, then statistically you would still be much more likely to live in the real world than in a simulation.

    • Lane Yarbrough says:

      You bring up a good point. What if a caveman is hunting a boar? And he takes a split second to weight his options: go left or right, charge or stalk, spear or club? Are not all of those simulations? 

      Thanks for posting this comment, right now I’m editing this post because I’ve automatically created simulations on how certain readers may perceive this. Interesting. 

      • dnebdal says:

        … I just realised that I sound really passive-aggressive in the post I submitted a moment ago in response to you. My apologies; I shouldn’t be posting this late at night.

  41. miasm says:

    Any simulation you care to run will, in some location within the multi-verse, have an analogue; real, actually happening universe.
    If pattern conservation is how stuff goes about the process of actually ‘happening’ then your simulation, which is perfectly identical to SOME universe out there IS that universe out there.
    Problem Solved.
    You can turn your ‘window’ onto the simulation off, but the universe it is a representation to your consciousness of will continue. All active modification/destruction arguments should be easily defeated with some deft use of the plenum.

  42. pltz says:

    as usual, when a scientific explanation exists it is way more insightful than any philosophical one:

    4) the resouces required to symulate a system of N interacting sentient beings increases exponentially with N

    (4) it is true for fundamental particles, I suspect it will hold for humans as well

    • John Aspinall says:

      “the resouces required to symulate a system of N interacting … increases exponentially with N”
      Nope.  That’s true if you have infinite-distance interactions that cannot be ignored or approximated.  But most systems can be approximated with logarithmic cutoff of detail.
      So, for particles, if I am simulating the Earth’s oceans, I might want to put the gravitational pull of the Andromeda galaxy into my computations.  But it would be a complete waste of time for me to put each individual star in the Andromeda galaxy into the calculation.  The Andromeda galaxy can be approximated, in my ocean simulation, by a point mass.
      Not that the results of the two hypothetical simulations wouldn’t differ, in the gazillionth decimal place.  They would (c.f. Butterfly Effect).  But both would be plausible, valid simulations within the uncertainty imposed by my incomplete knowledge of all the initial conditions.

      The situation is a little more complicated for interacting sentient beings, but my own suspicion is that approximate-yet-legitimate cutoffs would be even easier to find, simply because most beings will not be communicating different information to every other being.  If you have an aunt and uncle in Australia, let’s say, will there be a noticeable difference which of them you tell first about the new baby?  From the point of view of a simulation of your household, we can approximate by “we told the Australian relatives”, and the results will be realistic.

      • pltz says:

        “Nope. That’s true if you have infinite-distance interactions that cannot be ignored or approximated.”

        That’s what “interacting” means, if particles (be them humans or electrons) do not interact or interact negligibly they can clearly be simulated separately.

        “The Andromeda galaxy can be approximated, in my ocean simulation, by a point mass.”

        This is not an approximation:  thanks to Gauss theorem even the Moon can be exactly treated as a point mass.

        “simply because most beings will not be communicating different information to every other being.”

        Interaction does not require direct bi-directional communication; e.g. the 2000 US presidential election where about 101M people interacted to generate an event that affected the entire Earth population. I.e. your life is not influenced by a single Tunisian person, but it is influenced by the collective behaviour of Tunisian people: the revolution and all the events that it triggered. This is a sort of mean-field interaction, which is a very real interaction.

  43. sally599 says:

    I have no doubt that I’m living in a simulation…recently had a psychotic break and let me tell you it all seemed real—-all of the output you get for your version of reality is filtered through your own mind and heaven forbid the translation gets messed up and you start seeing leprachauns in the “real” world.  Sometimes seeing is not believing.

  44. cmpalmer says:

    There was an interesting Scalzi column a while back about his cat being fundamentally unable to comprehend where the bubbles in its water bowl came from and he extends the analogy to humans. The interesting point to me is the possibility that it is possible we can’t formulate the question, much less find the answer. It made me think because normally I lean toward the idea of a technological singularity and the inevitability of being able to eventually create simulations containing conscious intelligences.

    http://whatever.scalzi.com/2011/06/13/zeus-the-cat-confronts-the-greatest-mystery-of-his-life/

    The “why” of the universe, simulation or “reality”, or the ability to create such a simulation may be to us like the bubbles to his cat.

  45. Waffle_Man says:

    What I don’t understand, possibly because I’ve never investigated this theory closely, is why you’d go for the Simulation Argument when the Book Argument is stronger. Think about it, one of these statements must be true:

    A. Advanced civilizations go extinct before they become capable of writing books.

    B. Advanced civilizations are not interested in writing books.

    C. There are so many books that it is much more likely we are characters in a book then that we are in the physical universe.

    We can guess that A. and B. are false with a much greater degree of certainty then we can with the Simulation Argument, both because we’ve already invented writing ourselves, and because some form of writing would seem to be a prerequisite for making advanced simulations.

    The Dream Argument, which does not require any kind of civilization at all, would seem to be stronger still.

    What I’m asking is, what work is being done by the word “advanced” in this argument?

    • hypnosifl says:

      There are so many books that it is much more likely we are characters in a book then that we are in the physical universe.

      I think the assumption is that characters in books aren’t conscious, whereas really detailed simulations (of the kind that would simulate every synapse in a simulated brain) would be.

  46. Christopher says:

    Crumbs.

    Confusing messages about password resets = stupid double post.

  47. ultranaut says:

    Being an NPC is awesome. Now what?

  48. foobar says:

    I would argue for a subset of #2. Civilizations sufficiently advanced as to create such a simulation would consider turning one off to be genocide, thus significantly limiting the number that they could run.

  49. robuluz says:

    Isn’t this argument simply missing option 4: “No civilisation has yet acheived the capacity to create a sufficiently advanced simulation”, thus ending the speculation? What am I missing?

  50. SkylerNelson says:

    As a philosopher who studies physics, I have to frown at this sort of thing, because it makes us philosophers look silly.

    In these kinds of arguments, you’re always going to have to assume the existence of things outside our universe that we have absolutely no evidence for.  That’s going to be your main problem right there, and there’s no way around it(*), even in the best case.

    The way that this is supposed to work is that if you can simulate our universe (or one very much like it) within itself, then you could make a continuously expanding chain of simulations within simulations, and because there are a lot of simulations, you can make the argument that if you randomly select one of those simulations, it’s statistically not the first one. That trades on you knowing that there’s a chain of universes but forgetting your place in it.  You’re aware of a number of universes (because they’re contained in simulations you can observe), and if you imagine that you’re randomly placed in one and don’t know where it is in the chain, you’d not be able to figure it out.  But you do know where you are in the chain, because the only chain you have any evidence for the existence of is the one you can observe, and you’re at the first position there.

    It’s sort of like this: suppose you have a very large stack of blocks, and you pick a random block.  You know it’s a big stack, so you can assure yourself that the probability that you picked the first block is low.

    But, suppose you have a single block, and for some reason you can’t see what it’s resting on.  You can start adding blocks on top of yours, and continue as long as you want, making the stack very tall. If you randomly pick a new block in this stack, again, sure, it’s not likely to be the first.

    But none of that tells you anything about what your original block is sitting on!  Making the stack you can see taller never gives you any information about what you can’t see, it only allows you to make arguments that are valid in about what you can. What stacking blocks does demonstrate is the logical possibility that your block could be stacked on top of another, but gives you no reason to think that’s more likely than any other logical possibility you can come up with. Thus, even the possibility of constructing a simulation of our universe within itself doesn’t give us any reason to think that it’s in any way likely that our universe is simulated.  All it would do is demonstrate to us that it’s a logical possibility to do so, among all the other logical possibilities we can’t gather evidence about (creation by various gods, crapped out by giant space anus, emerged from the void, etc.)

    That said, what we know about physics makes even that unlikely!  You can’t embed the amount of information you’d need in an accessible volume.  Even if you’ve got some amazing compression scheme, there’s an algorithmic limit to how many bits you need to represent anything interesting at all, and there’s a physical limit to how much information can be contained in a given volume  The more nested universes you have, the more extremely crazy this becomes.  The best you could even imaginably do, given what we know, is a much smaller and simpler universe.  We could always be wrong about physics, but “we could be wrong about physics in a specific way that I can imagine and happen to find intriguing” arguments don’t buy you much in my book.

    (*) Assuming our simulators don’t poke around inside our universe and try to communicate with us in unmistakable ways.  There’s a point where you have to start considering that evidence, but as long as that doesn’t happen, we’ve got nothing to go on.

  51. The argument demands we presume too much.
    If “advanced civilizations” can manage to exist, AND if those “advanced civilizations” consist of entities sufficiently like humans to wish to have “advanced simulations” of such breadth and detail, AND we are not the first ever civilization, “advanced” or not, to appear, this argument could be valid. However, where is the evidence? What evidence do we have that any civilization can become so “advanced”? What evidence do we have that these “advanced civilizations” are sufficiently like us to make such broad simulations (perhaps they are very much drawn to only making extremely specific simulations). What evidence do we have that we actually are not the first civilization to arise in the universe?

    “Argument from infinity” is nothing but vacuous handwaving.

  52. mayer501 says:

    I wouldn’t be too quick to eliminate the possibility that its impossible to simulate the universe without a >= source of energy. Of course, quantum physics may or may not throw a wrench into that hypothesis. Still, our universe could merely be a pathetic echo (in code or whatever wibbly wobbly you want) of a much greater universe. I guess its up to us to make our own observations. Always remember that while simulating the complete universe, or rather existence is unlikely, simulating a controlled program with a limited number of observers affecting a much larger number of observation dependent functions may be possible… But hey, life is life, if we’re just programs at least that means we’re a finite sequence of 1s and 0s that will invariably be recreated at some point in space and time so…

  53. You can use a similar logical clusterf**k to prove that in all likelihood, we’re time travelling aliens from the future.

    The whole thing falls apart right at the start (of the original manuscript) though, when he decides to simply “take for granted” that human consciousness can be simulated. If he put away the Matrix DVD for a while and watched Short Circuit for a bit, he’d see that the question of how you prove something is alive and conscious is rather central to the plot. Though I don’t claim to be an expert in the matter, laughing at a bad joke is not a scientifically sound indicator of consciousness in a third party.

  54. isomorphic says:

    A universal Turing machine can simulate any other Turing machine.  If our universe is infinite and computable, then we can in turn simulate child universes of comparable complexity–albeit more slowly, from our point of view.  From the point of view of the child universe, everything would proceed according to its own rules.

    If our universe isn’t infinite (but is still computable), it would certainly impose limitations on our child universes.  However, if our universe is finite and computable, it might lend credence to the idea that we were being simulated.

    It is unreasonable to impose any limitations on the physics of a parent (simulator) universe.  That universe could be simulating ours with simpler laws of physics to test a theory.

    Also, if Someone (capitalized) is out there simulating universes for sport, perhaps one of Their selection criteria for continuing simulations is that the child universes eventually create their own simulations (grandchild simulations).

    Combining these two thoughts, the point of the whole Exercise might be to find the simplest set of rules and initial conditions that create a universe capable of creating simulations.

    Could our universe be an unoptimized quine?  (I’m assuming we aren’t the Solution.)

    All of this depends upon the universe being computable.  I hear Gödel shouting something from the grave about that, but his voice is muffled.

    (PS, if our universe isn’t computable, then I have some questions for Physics about what it hopes to accomplish.  Perhaps Physics would be happier if it were Scalzi’s cat.)

  55. Glen Able says:

    OK, so if this is a simulation, I’d expect that whoever is taking the trouble to run it wants to watch us doing fun stuff like running around and smashing the custard out of each other.  Sooooo…what about if we all go on strike?  We could paint our demands in huge letters on the moon, then simultaneously curl up into a ball and then refuse to move until they give us our goodies/upgrades.

  56. Jason Gots says:

    Boing Boingers, see also my post on this at Big Think: 

    http://bigthink.com/ideas/39548

  57. l337n00b says:

    It’s entirely fair to make assumptions about the “parent” universe and the sort of beings that would create a simulation within it.  After all, the argument does not show that we are in a simulation, argues that one of three things must be true.  The statement of them in the interview is a little loose, so let me tighten up the loopholes with some of the arguments from this thead:

    1. Civilizations don’t make it to the point that they can make “advanced” simulations (this could be because they die out first, because the universe isn’t old enough for a civilization to progress that far, because such a simulation is actually impossible, or because civilizations are simply too likely to die out beforehand)

    2. Civilizations that make it to that point aren’t interested in making such simulations

    3. We are probably part of a simulation

    If you don’t assume that the “parent” universe is like ours then you get stuck on point two.  The argument relies on the idea of “historical” simulations.  After all, given a universe like ours and people like us who advance to the point of being able to simulate the entirety of their history on the cheap, it seems a pretty good guess they would do it (for entertainment or for science).  However, given a universe not like ours and people not like us, the idea that they would simulate a universe like ours is suddenly unreasonably unlikely, and easily counters the idea that they would have many simulations.

    In the real world the question of “why?” may be a pointless one.  There may ultimately be no “why” to our existence.  In a simulated world there is definitely a “why” even if it is well beyond our ability to comprehend.  If we don’t assume that the “parent” universe is at least something like ours then “why?” is just too big a question to be ignored, and we stop at 2.  So, just as I’ve clarified #1 above to show the several valid reasons to think it is plausible, here is a re-formulation of #2:

    2. Civilization that make it to the point that they can simulation a universe/civilization as complicated as ours aren’t interested in simulating a universe/civilization like ours

  58. simonbarsinister says:

    No, I have subjective experience of my self, which means at the least my portion of a simulation has been instantiated and isn’t just implicit in the initial conditions and the logical outcome of the rules.

  59. Kieran Owens says:

    The argument in favor of existence as simulation is specious. It can’t even be discussed outside of metaphysics. It seems merely a restatement of the existence of God argument. If the “amount” of unreality is greater than the “amount” of reality then chances are that we exist in unreality i.e a simulation. Spot the problem? It only works if all are equally likely. Many would argue that reality is the only existence that is possible, all others are nonsense. It’s faith based.

  60. Ito Kagehisa says:

    I hired these blokes Zeno and Achilles to knock off Freeman Dyson, but they never got the job more than halfway done.

  61. The biological brain – most suitable host of bioware diversity available on Earth realms, to inhabit a microwave matrix imposed trough telepathical type of suggestion from higher intelligence minds, using not silicon circuitry, but biological material and not electronic or light impulses for data transmission, but radiation capable to cause mathematical/logical matrixes to be impsoed on brains, and call those “minds”, plus the ability to alter genetical code for biological upgrade of the host vessel meatware.

    Simulation and control, cyber loop. The elecrtonic worlds we build now are the product of a higher intelligence civilization puppeteering the human race as workers who “invent” and create the next storage/retrieval host for the grand simulation – cyberspace as it is, not yet announcing the awakening of what humanity would call “Artificial Intelligence”, but is just the newborn offspring of that higher intelligence entity inhabiting the human biological vessels trough radiational layering of mathemathical/logical matrixes, we call “minds”.

    or in short “Electronic Minds”. I wrote a book with hypothesis and speculations on the natural evolution to cyberspace.

  62. Martin Rix says:

    “If we are in a simulation (and I don’t think we are) it is upsetting to imagine a cruel operator who could flip a switch and send all of the people in the simulation into agony for all eternity”

    Have you ever played The Sims? Sim City? Any other Maxis game?
    How often do you play seriously and make their lives great? And how often do you mess around with their lives, burning their house, deleting doors, causing natural disasters?

Leave a Reply