Saturday, March 02, 2013

We don't live in a simulation

In 2011, I already wrote a text about the theories that our universe is a computer simulation:
Rebooting the Cosmos
Needless to say, a regular reader has seen lots of other criticisms of discrete physics, a more general concept. Let me return to this topic – and change the focus somewhat – in the wake of a sensible text at a German physics blog,
The simulation hypothesis and other things I don’t believe
As David Tong explained using different words in his Silver Prize Winning Essay written for a crackpot foundation whose basic mission does include the promotion of the "Simulation Hypothesis", there is strong scientific evidence today that the world isn't discrete (and it isn't simulated).

We do encounter integers and discrete mathematical structures in physics but in all the cases, we may see that they're derived or emergent. They're just limited discrete aspects of a more general and more fundamental underlying continuous structure, or they're a rewriting of a continuous structure into discrete variables (eigenstates in a discrete spectrum) which makes it impossible to understand the value of certain parameters.

Quite generally, if the Universe were fundamentally discontinuous, it couldn't have continuous symmetries such as the rotational symmetry, the Lorentz symmetry, and even descriptions in terms of gauge symmetries (which aren't real full-fledged symmetries but redundancies) would be impossible. In a fundamentally discrete world, many (or infinitely many) continuous parameters would have to be precisely fine-tuned for the product to "look" invariant under the continuous transformations.

I am not saying that it is "impossible" for all these parameters to be fine-tuned for the product to look e.g. Lorentz-invariant; I just say it is very unlikely. The scientific theory that requires this huge unexplained fine-tuning is less likely than scientific theories that work with no fine-tuning (or less fine-tuning). It's less likely simply because given any sensible, quasi-uniform a priori distribution of the parameters, it is insanely unlikely that they will have the "right" values if there are so many wrong values.

EPR-style experiments show that local realism is incompatible with the observations. Locality must be fundamentally respected because it may only follow from the Lorentz invariance, the Lorentz invariance is observed to hold as well, and making it accidentally hold in a fundamentally Lorentz-violating theory would require the nearly infinite amount of luck and fine-tuning, as explained in the previous paragraph. So the reality is local but non-realist. It is quantum, stupid.

It means that the proposal that our Universe is a simulation run on a classical computer is ruled out immediately. This is a very naive idea, indeed. Again, I am not saying that it is "impossible" to write a computer program that behaves in a way that resembles the reality. It may have a wave function treated as a collection of classical degrees of freedom; it can make the wave function "objectively collapse" at various points (although this guarantees that some disagreement with the observations is inevitable, as I have often emphasized). But it may look "more or less OK".

But when a simulation may "look OK", it's a different statement than the statement that the simulation may "be" the reality. These are totally different questions. The simulation may do a good job in emulating Nature around us. However, we may still prove by the scientific method that it must be a simulation, not the real deal!

I wrote a lot about the flawed idea of discrete physics. But in this blog entry, I want to write a few more words about a more ambitious aspect of this reasoning, the idea of an "Overlord". We are a simulation that a more intelligent "Overlord" is playing with. This is such a childish and mathless idea that we are really leaving proper physics. The first thing I would say – in agreement a Western European blogger – is that this idea is just religion in "modernized clothes" and its details reveal unusually naive anthropomorphism that we usually associate with very primitive forms of religions.

But some people apparently think that if they replace the old Gentleman floating on the cloud by a modern (or slightly futuristic) computer with its data floating in the cloud, it's not only much cooler but also much more scientific. Well, it's not. The decoration and makeup may look modern, technological, and geeky but the central idea in this scenario is still the same primitive religion.

(Some IT companies may think that their cloud technology is the newest state-of-the-art invention but clouds have been a basic tool used by the Overlords for thousands of years.)

Brian Greene promoted these ideas in the last 2 chapters of his latest major popular book, The Hidden Reality. It's a fun reading that tries to suggest that it's connecting physics with the ideas of science-fiction movies and books. However, it is not. These ideas are pure science-fiction. They don't contain any "physical added value" whatsoever. They show a physicist trying to make an excursion to the world of science-fiction.

Don't get me wrong. I do like many of these books and movies. I think that some of them are clever in the ways how the rules of the game are sketched. I have watched some of these movies many times, thought about them, and I have perhaps tried to write my own science-fiction novels, too. Those activities may require some thinking. However, it is important to realize that this thinking is certainly not scientific thinking.

The Ultragirl who plays with the simulation is usually visualized in a totally anthropomorphic way. She is a girl so she just wants to have some fun. She may be bored. She doesn't want to be bored. She may have bought the device with the simulation in the mall. She may even be sexy and tease you. She may even look like a human babe.

Needless to say, all these ideas are preposterous. If there were a totally new fundamental layer beneath our Universe, and I have argued it can't exist, it would not only have no reason to include beings that are physically similar to humans with the same kind of lust. Even broader feelings such as fun and boredom would almost certainly be irrelevant. They're as silly as they are in the case of Moses and the Lord who apparently created the Solar System within a week some 6,000 years ago but who needed to relax on the sixth or seventh day (sorry, I am really uncertain at this point!) because all the workers want their weekends to be relaxing.

But the anthropocentric delusions don't stop with fun and boredom. Some people, including Brian Greene, are even promoting a scenario in which the simulation runs on a limited device that only simulates something or uses rougher, approximate algorithms to simulate things that aren't too important because the simulated beings aren't observing these aspects too accurately (or they don't observe them at all).

It may be a good idea for you to buy a cheaper computer if you don't need the most powerful one. You may have good reasons to save some electricity and run simpler programs if you don't need the state-of-the-art ones. You may pay smaller bills and you may save some money for other things. But to imagine that the underlying Overlords are trying to save the money in the same sense as we do is the ultimate anthropomorphic naivite. It's not just anthropomorphic: it's an idea that envisions a layer underlying the whole observable Universe to resemble the early 21st century culture of the U.S. urbanite teenagers.

Holy Christ and Saint Simulator, we must ask: Why? All these assumptions about the Overlords are completely irrational. There is absolutely no reason to assume that the beings underlying our existence would resemble the U.S. urbanite teenagers of 2007. In fact, if we have some experience and if we can imagine a reasonable "landscape of possible ideas", we must agree that the probability that the underlying reality respects these principles is virtually zero.

Brian Greene faithfully yet uncritically sketched many other arguments in favor of the "Simulation Hypothesis" that I find utterly idiotic. One of them is a sort of the anthropic propaganda. He says that in the asymptotic future (of our particular Universe or perhaps the whole multiverse), there will probably be so many computers with so much power that the number of simulated TRF readers will be vastly greater than the number of biological TRF reader. For this reason, Brian thinks, it's far more likely that you are a simulated being, not a physical or biological one.

There are several reasons why this argument is dumb. One of them is that there is no such a law that the probability that you belong to a subset S of the whole set W is equal to the ratio of the number of elements N(S)/N(W). In most cases, the ratio isn't even well-defined because both the numerator and denominator are infinite. But even when they're finite, there's absolutely no reason why the probability should be equal to the ratio and in almost all cases, it is not equal. Only if all elements of W are equally likely – and that pretty much occurs only when there is a reason why they are equally likely (thermalization balancing all the microstates, for example) – the probabilities may be represented by the ratio. But it is a negligible fraction of the situations. There isn't any symmetry or democracy between biological and simulated beings because they're qualitatively different objects so there can't be any justification of the "equal odds" assumptions.

Regardless of the number of simulated beings "somewhere", we may still present scientific evidence that there is no "simulated layer" in our reality. The continuous symmetries are such an example. If you want to deny this argument, you may deny it by saying that your meme that "the number of simulated beings is vastly greater than the number of genuine beings" is so powerful (because the number of simulated beings is so high) that it may render the smallness of the "we are simulated beings" prior probabilities irrelevant.

But such a method to render rational arguments and calculations irrelevant is just a sleight-of-hand. It is exactly equivalent to the statement that all scientists and atheists are deluded a*s*oles because God is infinite and by His infinite powers, He may neutralize any scientist, his arguments, and inflate an arbitrarily unlikely possibility into the most likely outcome. To argue in this way means to totally abandon rational discourse and replace it by a kindergarten boys' pissing contest trying to find out whose God or Father is more infinite. You're just showing that your faith is infinite and won't be affected by any finite evidence, however extensive. But this only proves your bigotry. It doesn't prove anything about Nature.

Moreover, one may conjecture equally convincingly that in the future world, the number of biological beings will be vastly dominant; yes, I am just saying that the other side could be active in the pissing contest, too. I don't want to spend too much time with these anthropic arguments because they're just totally unjustifiable. In fact, there aren't any infinite numbers in physics that could ever "strengthen an argument". All physically meaningful numbers are always finite (some renormalization etc. is sometimes needed to see it) and if you think about possible infinite numbers such as the volume of our not-only-observable Universe, they never affect probabilities of observable propositions. The latter are only affected by intensive or local, and therefore finite, quantities. Someone's infinite size as the "ultimate argument" belongs to religion, not to physics.

Another would-be scientific yet unscientific line of defense that the champions of the Simulated Universe Hypothesis often employ are the demands of loyalty towards our Overlord. Some of them are as dead serious about this line as the most hopeless religious fundamentalists. They say that the Overlord may be deliberately trying to mislead us and we must just stay humble and respect Her great power because we're such tiny insects.


Well, let me say that I know that this Overlord doesn't exist but even if I had doubts about Her existence, as a scientist, I don't have and can't have any respect towards Her "authority" that would affect my reasoning. Such an influence of fear on reasoning is just not kosher for a scientist – and I would surely extend this moral principle to North Korean scientists and simulated scientists, too. It's just about the most important, defining features of science. So even if the Overlord exists, she may politely have a romantic relationship with Her own openings, or whatever other parts of the body She is expected to have. Give me a break with this stuff. Worshiping a being, especially one that doesn't exist according to all the rational reasoning, is just plain stupid.

When She is presented as omnipotent in this way, the champions of Her dominion over the world have no problem to say that She may fine-tune things to values I find unlikely because it's Her right. She has the right to confuse me, too. Because these folks live in a mental world that assumes that She exists as if it were the most reliable and universal axiom of science – in a complete analogy with religious bigots in Christianity, Islam, and perhaps other old-fashioned religions – they consider their uncritical faith to be a sign of their being scientific.

But it's exactly the opposite. In religion, one can make assumptions (such as the Simulator's omnipotence) and require that everyone else worships these assumptions and treats them as facts. All other facts must be distorted, adjusted, or completely overlooked so that they become compatible with the basic religious axioms and of course that if you're willing to say awkward things, it's possible to a large extent. However, this approach is unacceptable in science.

In science, we always question the assumptions. We're comparing competing theories whose assumptions differ. That's really another defining feature of science. So feel free to create a personality cult around the Female Simulator but a person who is thinking scientifically will always question it and will always try to estimate the probability (and probable truth values of various assumptions) using the most rational, indirectly empirically based, arguments that are available.

Let me phrase this point about "Her right to confuse" differently. Some champions of the Simulation Hypothesis (and religions) think that if they postulate that "She" can arrange things in such a way that scientific strategies to investigate the foundations of the Universe become misleading or impossible, they make their Simulation Hypothesis more likely because they "hurt" the other hypotheses. But the answer is the opposite one. They only hurt their own, Simulation Hypothesis because they are adding additional unnatural assumptions to it: they are making it less likely. The probabilities of the competing (scientific) explanations can't be lowered by axioms done within the Simulation Hypothesis because – with apologies to the Overlord – the competing scientific explanations lie beyond the Overlord's power. So if the Simulation Hypothesis contains additional axioms that "hurt the Overlord-free scientific explanations" and if someone believes that the scientific explanations really become less likely because of that, he is a victim of circular reasoning. He can only prove what he wants (scientific explanations are disfavored or impossible) if he assumes it from the beginning, anyway.

For the question of fine-tuning, a rational argument is that the fine-tuning of 40+ parameters of the Lorentz-breaking Standard Model to the Lorentz-preserving values is almost infinitely unlikely, so a fundamental theory that doesn't respect the Lorentz symmetry is almost fully excluded. That's true for the Sexy Simulator Hypothesis, Loop Quantum Gravity, or any other silly fairy-tale of this sort. It doesn't matter that you may imagine that such an unlikely scenario is true and you may collect a billion of gullible simpletons around the religion that it is actually true. Science is still saying what it is saying and it is saying that the probability that such an assumption about the Simulator holds is basically zero. It's the strongest way how science may disfavor a hypothesis. So science is against you in the most violent way that the cautious and peaceful structure of science admits.

To summarize, I think that if you admit that there is a difference between science and religion, the proponents of the Simulated Universe Hypothesis are squarely on the religious side and all the feelings that their theories are modern, geeky, technological, and advanced are just about the makeup, not about the essence of the hypothesis that is the same primitive religion that the mankind has played with at least for thousands of years.

And that's the memo.


  1. Is there anything coming out of Moriond yet?

  2. Ok, if the world is continuous, how come string theory, a TOE, can be formulated using a finite set of of discrete symbols, and every theorem of mathematics is just a combinatoral statement about manipulating finite strings of symbols?

  3. The laws and propositions are discrete but they are about continuous objects.

  4. I think as one understands the nature of spinors more, the differences between continuity and discreteness become more clear. When one understands that the solutions to the Dirac equation are continuous functions, but have discrete features, some of the confusion is alleviated. Chirality is a interesting property that highlights how discreteness emerges from a objects that are otherwise continuous. Who would normally think of nature operating in this way? It is much easier I think for people to ignore relatively simple and elegant answers and insist on cruder ones.

  5. Dear Lumo,

    have you actually written a science fiction novel or something like that which you could upload here? Just thinking about reading it and what it potentially could contain makes me LOL ... :-D. I am sure such a thing would be a hell of a fun and cool reading !

    These discussions about the universe as a simulation etc, reminds me of the cozy final round after a good students party with enough beer and maybe something to smoke available too at about 3:00 in the morning, people starting to talk about ever crazier philosophical ideas and world views, etc ...

    And sorry but I dont see the difference between the overlords and the Santa Claus in white floating on a cloud :-P

  6. Simply superb. I did not only have fun reading your article but also I learnt a good number of arguments against this pseudo-science that is infecting us as always religion did. Thank you . I will keep reading you.

  7. Great post Lubos! The simulation hypothesis is fun, like a good comic book that would blow your mind or should I say knock your socks off... But why do guys always take it to the letter ? Nerds... :-)

  8. Thanks Dilaton! The novels were in the times when "writing" surely meant "writing in Czech" in my case. Even today, I would almost certainly prefer such a novel to be translated into English and not written in English (and fixed by a language editor).

    Right, this is exactly the 3am after-party stuff when everyone reveals what *actually* intellectually excites him or her. ;-)

  9. I am not so shure. There is a paper saying that Moor's law can in principle hold yet another 600 years.

    At that point having gone down in scale as much as is possible, we may be able to manipulate the
    very fabric of spacetime and then, who knows .... Of course, simulating a universe on a silicon computer as
    some people tend people believe, that's at best bullshit - sorry that I'm polemic too. But paradigms tend to change...


  10. Dear Dan, apologies but this kind of gibberish leaves me speechless.

    First, the process of making transistors exponentially smaller must end at most in a few decades because one gets to the scale of atoms - and the barriers will probably arise long before that. The total number of transistors may only keep on rising then if the computers get physically larger again and I don't think there will be a reason for that.

    Second, what does the increasing computation power of computers has to do with "manipulating the very fabric of spacetime"?

    If you want to manipulate a fabric of spacetime, you need a Planckian collider or another device to compress mass or energy to Planckian densities. Computers have nothing to do with it.

    You may try to do calculations/simulations that are - plausibly - relevant for the predictions of Planck scale physics but that's something else than "living" the Planck scale physics. Moreover, this simulation, even if one had the right theory and algorithms with everything (we don't quite understand everything yet), would have no implications on "our identity" because our thinking and life processes are almost independent of ultrashort distance physics such as quantum gravity. It's just a low-energy atomic physics, biology etc.

  11. Surely there's no Overlord in any anthropomorphic way, but can't Nature itself be seen as the suprem Overlord ?
    I mean, incompleteness is in some sense, a way for nature to "protect" herself from what we could discover on her, isn't it ? ;>
    So, tell me, do you have any respect for Nature when it affects your thinking ? Cause it surely does, free will or not.

  12. Isn't Wolfram pushing something like this with his cellular automata theory about the universe ?

  13. Nature is amazing, smarter than us, and so on, but these are totally different superlatives than anthropomorphic ideas about Nature as a woman who tries to protect Her sensitive parts from our looks, and so on.

    This is just silly. The laws of Nature have no correlation with the question whether they're available to us. The success of physics surely shows that they're not being completely hidden - in this sense, Nature is more of an exhibitionist than a shy girl.

    In some respects, it may look harder than expected to understand certain things about Nature; in other respects and contexts, it may look surprisingly easy or easier than expected. There is no universal rule and there is no reason why such a universal rule should exist. These comparisons depend on our expectations - which are not terribly important things as they're ultimately shown to be inaccurate.

    I have lots of respect for the wisdom hiding in Nature and the way how it works. But if you ask me whether I have a similar respect to Nature as I could have towards a distinguished woman who has privacy, human rights, and who may have the right or desire to hide her sensitive parts or anything of this anthropomorphic sort, my answer is a resounding No. A scientist just can't have any respect of this sort, it would be called bias.

    I think that your question is, by itself, the example of the unscientific emotional demagogy that is often used. You want to a priori suppress or ban my answer - the only answer that a scientist may pick - by presenting this answer as borderline immoral. There is nothing immoral about investigating things without prejudices and without anthropomorphic sorts of "respects".

  14. Of course the classical transistor will not serve the purpose any more.

    You are doomed to go down to the Plack scale to have access to the "fundamental bit" in nature on the scale $\approx \ell_{Plck.}^2$ (see e.g. see t'Hooft, S. Loyd).
    There is a high energy demand, yes, this may well be THE ultimate obstacle.

    I thought about it in the philosophy of Feynman, Deutsch, etc. that you can't really distinguish a computation from a physical process.
    In the end, I think computation, simulation, emulation and all that will converge with pure physics. (Quantum computation, it seems to me, already points in this direction).

    If you don't believe me, maybe you believe this (crazy :-)) string theoretician:


  15. Of course the classical transistor will not serve the purpose any more.

    You are doomed to go down to the Plack scale to have access to the "fundamental bit" in nature on the scale $\approx \ell_{Plck.}^2$ (see e.g. see t'Hooft, S. Loyd).
    There is a high energy demand, yes, this may well be THE ultimate obstacle.

    I thought about it in the philosophy of Feynman, Deutsch, etc. that you can't really distinguish a computation from a physical process.
    In the end, I think computation, simulation, emulation and all that will converge with pure physics.
    (Quantum computation, it seems to me, already points in this direction).

    If you don't believe me, maybe you believe this (crazy :-)) string theoretician:


  16. Actually, I just wanted to point out that the Overlord does exist in a non anthropomorphic way. Be sure that I totally agree with the rest of your answer and that I'm as convinced as you that people who trusts in a kind of simmulated universe theory are retarded.
    Yet, these questions are not physical but metaphysical ones, and to be honest I'm also convinced that a good physicist is not "just" a scientist who does science as if it was just an empty game.
    There's for sure more than just a set of models free from metaphysical interpretation.

    So, well, lets be fair. People who trusts in simmulated universe can't be true regarding our current knowledge (and some common sense). Yet, what do you think about all these guys convinced by the multiverse interpretation of quantum mechanics ? Isn't it as retarded as trusting in a kind of simmulated universe ?

  17. PS Great post by the way. I am quite at a loss to understand why a highly reputed guy like Brian Greene seems to love pushing such flaky ideas as simulations and MWI. I fear that media physicists want to "sex-up" physics because they think it will interest people more. So they look for the weirdest and wackiest theories. But in fact it only makes people think that in physics we just make random stuff up and it's a sort of free for all like art where every idea is sort of possible and valid and is more original wins. Sometimes I long to see a media physicist do something like write down the Schrodinger equation and then solve it slowly and carefully for the hydrogen model and discuss how the results compares to reality and then bring in relativity etc.... I am fed up hearing about Schrodinger's cat. A programme like I suggest would break this myth and show physics for what it is. And I think it would excite many more serious and smart potential physicists than all of those programmes with the special effects and hocus pocus that we get from the two Brians! But then the liberal arts programme comissioners would never commission it. End of rant.

  18. I've previously only ever read here as I'm spectacularly ill-equipped to comment. However, for once, I wonder if there's any mileage in my anti-Simulation thought: that if we "exist" within a simulation, why not the simulator also? ...and so on, ad-infinitum, turtles all the way down. Which for me at least, gets filed under: 'obviously nonsensical'.

  19. There is one problem with simulation-hypothesises I havent seen mentioned before, suppose one does 'simulate' some world X or universe or process, ok, is this simulation of X still not something completely different from X? Suppose we wrote down the equation of a one celled organism, then computed it step by step, on paper, does anyone think that this simulation constitutes a real organism? There is a difference between an equation on a computer, and what that equation describes, even if the equation describes it fully. As Stepehn Hawking said, what is it that breates fire into the equations to make them come alive? I dont see anything special about 'computing' in any form that can make some equations come alive. The only possibility is that the universe is simulating itself, which is a quite a meaningless statement.

  20. Yeah, I think what some media physicists do does not help but it rather gives the "real" physicists working outside the attention of the media a bad image ... :-/

    If you want to see a physicist slowly deriving the equations, solving them, and discuss what the equations and their solulions mean by just writing and drawing stuff on the white board, you can look at these gems:

    Maybe they will get more popular too with the book corresponding to the first course that has been released an reviewed by Lumo recently.

  21. Retarded ?

    "Most important of all is the simple statement that once inflation happens, it produces not just one universe, but an infinite number of universes."

    - Alan H. Guth

  22. What if the Universe is a quantum computational simulation? For example, if M-theory is hiding Cayley plane fibers, the plane itself can be interpreted as the state space of pure octonionic qutrits. The unitary gates acting on these qutrits are of type F_4 and the SLOCC gates are of type E_6(-26). Perhaps it is a matter of interpretation, or maybe Nature does employ a type of quantum ternary logic. Either way, the simulation hypothesis might be testable by further study of black holes.

  23. I'm also convinced that a good physicist is not "just" a scientist who does science as if it was just an empty game.

    I can't possibly understand what you may mean by this sentence so that it's not nonsense. Physics is a science and science done really properly is physics - up to some subtle differences, these words mean the same thing or at least a thing that employs pretty much the same methodology, the scientific method. The scientific method is not empty and it leads (and it has led) to huge insights that are not empty at all. If someone thinks that physics or science are "empty", well, then they dislike physics and/or science. But physics done properly *is* pure physics, it is stripped of all the unphysical gibberish. Of course that identifying this gibberish and removing it *is* a feature that makes great physicists great.

    I don't understand what you mean by "multiverse interpretation of quantum mechanics with simulations". It seems like you have mixed several silly things into a thing that would obviously be even sillier. I don't think it's appropriate to spend more time than the time I needed for this paragraph with your bizarre construct.

  24. Even if one could pump Planck-scale energy into elementary particles, I find it sort of obvious that one couldn't construct a viable and working IT device from Planck scale components.

    This claim has several subclaims such as the fact that Planckian colliding objects are inevitably black holes; that minimal, Planckian black holes instantly evaporate so they can't hold information for too long; and that the information coming from the Hawking radiation becomes impossible in practice to be decoded and this problem gets more severe if we deal with larger - and therefore more stable - black holes. So there's some fundamental level information exchange going on at the Planck scale but it's also in the regime that may be least exploited.

  25. I do not believe in the simulated Universe for the same reason as Zer0th pointed out. At the end of the day, this does not explain anything - i.e., is the Universe of the simulators more fundamental in some way and so what are its basic laws or is it also a simulation etc. ad infinitum?

    However, if our Universe were a simulation, I disagree with you Lubos that our science would help us find out about it. Our world could be simulated on a lattice and so at some fine scale would look discrete and be incompatible with continuous gauge symmetries. But to know this we perform a finite number of experiments - at any time the ones running the simulation can hit the pause button and modify the results of the experiments so that they conform with our view of the world. In fact for all you know you could be yourself a simulation that was started 5 minutes ago, with a set of initial conditions that provide you with all you know about physics and the world, but which could have nothing to do with the actual reality (that you are in fact a program running on someone's computer 8) ). Ah well, don't loose any sleep over it.

  26. Lubos MotlSo there's some fundamental level information exchange going on at the Planck scale but it's also in the regime that may be least exploited.

    I find this an interesting point. The idea here is that we head in that direction? What would information communication look like. Is this possible?

    However, Kapusta also notes that a sufficiently advanced civilization might use pulses of neutrino
    superfluid for long-distance communications.
    The right spin for a neutrino superfluid


    Galactic Neutrino Communication by John G. Learned, Sandip Pakvasa, A. Zee

    We examine the possibility to employ neutrinos to communicate within the galaxy. We discuss various issues associated with transmission and reception, and suggest that the resonant neutrino energy near 6.3 PeV may be most appropriate. In one scheme we propose to make Z^o particles in an overtaking e^+ - e^- collider such that the resulting decay neutrinos are near the W^- resonance on electrons in the laboratory.

    Information is encoded via time structure of the beam. In another schemewe propose to use a 30 PeV pion accelerator to create neutrino or anti-neutrino beams. The latter encodes information via the
    particle/anti-particle content of the beam, as well as timing. Moreover,
    the latter beam requires far less power, and can be accomplished with presently foreseeable technology. Such signals from an advanced civilization, should they exist, will be eminently detectable in
    neutrino detectors now under construction.

    If everything is energy? Is there a better way to describe such an Overlord?:)


  27. Dear Lubos,
    I am not sure how you decided we need a discrete grid to be able to simulate a continuous universe and continuous symmetries of that universe. I wonder how many universes you have simulated in your life and came up with that conclusion but obviously it is just plain wrong. And we also don't need a huge computational power to do it, the only required thing required is lots of memory. The cpu time can be 10^100 seconds for only 1 second of the simulation time, it obviously doesn't matter since the perceived time would be independent of that.

  28. Sorry, Henry, but you apparently deliberately misread. I wrote - and I emphasized it extremely carefully - that you may indeed make various simulations.

    But we may still prove, using the scientific method, that our world is *not* a simulation.

  29. Scooby, you're wrong. Science *always* helps (it is even a weak word, just help?) a thinking creature to find out about all meaningful questions in the world where the thinking creature lives.

    The idea that the scientific method would break down if the laws of physics obeyed just some special or unusual property is just a micron away from the paranormal claims that science actually breaks down in the real world. It never does.

    Your finite-number-of-experiments comment is totally irrelevant. Science (even science in this very world) is *always* built on a finite number of experiments. That doesn't mean that it's impotent to find important physical things out.

    There are important basic facts about our world there make science more straightforward. One of them is repeatability that boils down to the constancy of the laws or time-translational symmetry. So when we learn about the laws that held an hour ago, we also learn about the laws that hold at another moment.

    But even if this weren't the case, science wouldn't be impossible. The laws could be changing but we could still study the laws that determine how the "first laws" are changing. After all, cosmology and many other "historical sciences" in the real world belong to this category.

    If our Universe were a simulation on a computer - and we've proved it's not - we could still in principle find out about every physically meaningful question we could ask. We could ultimately find out the number of microprocessor cycles that individual instructions that run in our Universe take.

    If the simulation were running on a laptop, it could happen that the brand of the laptop - Dell or Hewlett-Packard - would be impossible to find. But again, science has very clear rules on whether it's interesting or not. If the brand could in principle show up somewhere - we could connect to the webcam and see an image of our laptop; or we could find a hardwired ID of the laptop in some chip accessed somewhere - it would be possible to find out. If there weren't any way to find out, even in principle, then the question whether we're in a Dell or Hewlett-Packard would be utterly unphysical. If the influence of such a question on our experiments is demonstrably zero, the scientific interestingness of this question is zero as well, regardless of the amount of (religious) hype that some people could create about this question.

    Even in real science, sensible folks/scientists are abandoning tons of such unphysical questions. We don't ask whether our world runs in the Heisenberg picture or the Schrödinger's picture because they're exactly equivalent. That's exactly analogous to asking whether our simulation - if we were a simulation - is running on a Dell or HP laptop.

    Science has no limits and a thinking creature in any context or hypothetical world who follows its rules properly and who is sufficiently smart and creative may find out how things - including important things - in his world work, including the most fundamental ones, as long as we're talking about things that are scientifically meaningful. Believing that science isn't possible is a sign of one's supernatural delusions, a quasi-religious stupidity.

  30. I agree, this may well not be not doable.

  31. I have a question.

    Does the simulation argument imply a reversible or an irreversible computation ?

    If I remember that correctly, S. Loyd said that the universe has made ≃10^120 flops since the Big Bang.
    If it is a reversible computation, I conclude that this leads to an energy consumption of the simulator of the order of 10^80 Watts !!! due to Landauer's principle.

    That's the energy it would take to simulate just one inflationary pocket universe.

  32. Dear Dan, at the level when one studies the "useful information" in a computer only, computation may be reversible. But a physicist studies all degrees of freedom, not only those useful for the IT guys, and processes involving these degrees of freedom, as long as they're doing something useful, can't be reversible due to the second law of thermodynamics. As you say, lots of heat had to be created as well, due to Landauer's principle.

    The simulation champions would complain against your argument by saying that they may use rougher simulations for less important things.

    But I don't get the point of all these arguments about the consumption. Our Universe is clearly not a simulation run on computers in the *same* Universe. It would be like Jesus' being a thought of Jesus. If we talk about a simulation at all, it's a simulation in a different, higher-level Universe where the concept of energy, it it's relevant at all, surely has different units and different relevance. It seems anthropomorphically preposterous that this higher-level Universe would have to deal with questions how to save energy. Why? Except for the reason that people love to imagine Overlords with properties similar to ours?

    Even our Nature - without Overlords - ultimately has no problem to compute with infinitely many continuous numbers completely exactly. All the evidence surely indicates that is doing nothing less than that. Much of the discrete physics and simulation hypothesis bullshit is driven by the anthropomorphic idea that it must be hard for Nature to calculate with precise numbers because it's hard for the humans. But Nature isn't humans. Humans have trouble to calculate with continuous numbers because they have a finite number of neurons and they work in a *certain* way they cannot change. But Nature doesn't have a finite brain and it is not a human. It doesn't have to work in the same way - and it almost certainly doesn't work in the same way. It just has no limitations of this sort, no problem to use laws that are absolutely precise and start with infinitely many continuous degrees of freedom. Whoever has a problem with this ability of Nature is trying to project anthropomorphic properties on Her - it's just stupid and absolutely scientifically unjustifiable.

  33. Thanks a lot for your exhaustive answer. It all makes perfect sense to me what you are saying.


  34. Hi Dilaton - I have all these on my ipad. Susskind is a dude! I did not know about the book. Must have a look.

  35. From the point of view of the paranormalist I'd argue that the evidence from psychic research and other extra-curricula sources, denigrated as they may be by the ignorant and orthodox, do suggest to the thinking modernist that the physical reality is analogous in some respects to a simulation in some way or other. This is because it is readily apparent that the physical laws are in fact pretty rubbery when the rubber actually does hit the road.

    Given my deep misunderstanding of quantum mechanics in every respect, I'd nevertheless speculate that it does provide a sort of basic infra-structure for this to happen. Since everything fundamental is said to be probabilistic, I am assuming nevertheless that the probabilities require adherence to basic conservation laws, the law of least action etc, in order to model physical reality effectively. (This may be an incorrect assumption on my part)

    Therefore, in order for supernatural or paranormal phenomena to occur, which they do, you have to postulate some sort of further dimension with its own laws of physics that connects with the mathematical box of quantum mechanics by manipulating the conservation laws in some way.

    So for example the photons normally travel from source to target in aggregate by following the law of least action. But if something in this further dimension says to the photons, stop following the law of least action and now follow the law of Such and Such so as to produce paranormal effect A, then this would be seem to be allowed in my opinion in the quantum mechanical scheme.

    So I would argue that in the normal physical state, the law of least action rules as per usual, the reality is not a simulation as such. However in supernatural states, this rule has been over-ridden by dictates from the further dimension, so the reality then takes on some of the characteristics of a simulation.

    Obviously the evidence suggests that the 'further dimension' is analogous to a mental structure of some sort, but since we don't even understand our own basic brains very well, I'd hesitate to make any anthropomorphic projections as to of what this mental structure might consist.

  36. "So the reality is local but non-realist. It is quantum, stupid."

    Nice line.

  37. "This is such a childish and mathless idea that we are really leaving proper physics. . . [another good line] . . . "and its details reveal unusually naive anthropomorphism that we usually associate with very primitive forms of religions."

    Shouldn't that be "advanced" not "primitive" forms of religion? The primitive forms were anamistic, not anthropomorphic. Not that there is anything wrong with anthropomorphism in the social (ie, non-physical) sciences, where man is the measure of all things.

  38. Dear Luke, physics is fortunately not a social science.

    Anthropomorphism is a special case of anamism because people are animals. At any rate, both ways to imagine the fundamental layer of laws of Nature are equally infantile!

  39. In a simulated universe, science would operate within the limits imposed by the simulation. Believing anything else is wishful thinking.

  40. strictly speaking...Mar 5, 2013, 12:33:00 AM

    This is actually an interesting statement. You could assume that all
    mathemathical objects in physics are defineable, in a way similar to the
    defineable numbers.

    form a countable subset of the real numbers that is closed under a well
    defined set of logical operations. Unlike other systems of interest
    such as rational numbers or algebraic numbers, these actually form a
    countable subset of the real numbers that should be completely
    indistinguishable from the real numbers for the purposes of constructing
    a physical theory. This assumes that the universe isn't horribly
    pathological, having physical states that are not explicitly definable
    using set theory. Should that be true, it'd be an at least equally
    worthy cause for killing yourself to the existence of a spin-2
    Higgs-like boson. ; )

    These countabillity properties would imply
    that at least, our universe(and it's entire time evolution) could
    technically be described entirely in the memory of a turing machine
    using gödel numbering.

    While the simulation hypothesis is far fetched, at least it isn't demonstrably false. It is only an unfalsifiable proposition similar to religion, as you nailed in your post.

  41. Dear strictly speaking,

    right, in everything that makes a physical sense, real numbers could be replaced by (countable) constructible numbers.

    But the laws - that are also formulated using a finite number of countable symbols - still describe real numbers. What matters here isn't really that real numbers are uncountable and one may find non-constructible real numbers. What matters is that the laws of physics preserve the continuity.

    If you enumerate the constructible real numbers and identify them with integers, the ordering will be arbitrary gibberish that isn't compatible with the laws of physics. What I want to say is that if you convert the countable real numbers to their integer-valued IDs, the laws controlling these integer-valued IDs - pretending to be real numbers - will have a completely unnatural form. The laws will always require you to calculate these numbers, convert them back to real numbers with a sufficient accuracy, and then deal with them.

    But if you count "purely technical" limitations, there's still the limitation that real numbers can have any accuracy but your computer has a finite number, not countably infinite memory. So the error of the fake real numbers is always finite and nonzero and an arbitrarily small error has some implications in some contexts.


  42. It is complete nonsense. Intelligent brains, whatever the microscopic fundamental laws make them operate, ultimately work in the same way and they're capable to simplify the observed information, generalize, and induce the underlying laws.

    One may set "limits" what is accessible to the simulated beings but that doesnt affect the previous paragraph. If certain things are in principle accessible to the simulated persons' experience, they may be cracked and understood. If some things aren't accessible at all, even in principle, they're just complete unphysical and no scientifically minded person, simulated or otherwise, should consider them important because they're meaningless.

  43. Hmm, I have been pondering about this very interesting statement:

    "Humans have trouble to calculate with continuous numbers because they have a finite number of neurons and they work in a *certain* way they cannot change. But Nature doesn't have a finite brain and it is not a human."

    How can the last statement follow from the first one ?

    This boils down to an old question of mine: How can a brain if it is finite make any statements about the countable or even uncountable infinite. (E.g. take mathematical limits, etc.) - I am completely stuck here.

    This issue pops up in many contexts in physics, QFT - infinite # of dof's, thermodynamical limit, non-analyticity in phase transitions, renormalization, etc.

    I would like to better understand your *certain*

    Is there any experiment conceivable that could unambiguously tell us if a phenomenon is finite (or "digital" in the language of the simulation adepts), countably infinite or uncountably infinite.
    (E.g. is a continuous spectrum *really* continuous or just continuous "for all practical purposes").


  44. Dear Dan, I have obviously never claimed that the "second claim follows from the first one". I only made two claims. Both of the are right. Obviously, Nature's not having "something" doesn't follow from humans' having it. They're independent questions.

    Finite brains have been able to predict quantities that take values in infinite sets - whether they're countable or not - because in Nature, all the observables follow rules that ultimately may be "written on the T-shirt" (with a finite number of symbols), or at least the effective laws have this property, and the damn finite brains just reconstructed what's written on the damn T-shirt! What's so hard about it?

    But the laws written on the T-shirt - like the Standard Model Lagrangian - are describing the properties and evolution of a priori continuous degrees of freedom, and infinitely many. What's so hard about it? This is how physics works all the time. If something seems shocking or surprising about those things to you, could you please ask a particular example what you're puzzled by? I guess that when you try, the confusion will look as silly to you as it looks to me.

  45. I think you could do very well in English, Lubos, with a human editor to correct a few nits. Then you would have to go over the editor’s corrections to make sure the meaning remained intact. Your English prose is really very good.

  46. Dear Lubos,

    Let me try with an example (if I fail, I will follow your advice and stay silent about it :-)):

    Chaos, which is sensitivity to initial conditions. I claim that any classical theory that exhibits chaos is an effective theory (e.g. chaotic behavior in Newtonian mechanics).
    But if one goes to the more fundamental quantum mechanical description where the phase space is cellular, chaos should go away and there is no "true" quantum chaos (i.e. sensitivity to infinitesimally small changes of initial conditions).
    My more general claim is that any infinity, infinitesimal etc. is a manifestation of an effective
    physical description which should disappear in a more fundamental description. Example, GR and singularities.

    I read somewhere that string theory is the only fundamental theory found so far - so then, is there (quantum) chaos in string theory ?


  47. Dear Dan, you would have to define more accurately what counts as quantum chaos for you and what doesn't.

    Very simple systems such as quantum billiards display quantum chaos. One may embed quantum billiards within string theory, at least in some approximation, so of course that string theory - as a theory of everything - contains everything else that occurs in valid physics before that, including quantum chaos.

    The right predictions of the quantum billiard are modified in the extreme, Planckian, stringy regime. But the sensitivity doesn't really disappear. This sensitivity is described in very different terms in the "usual" systems that deal with the information and the Planck scale - like the black hole evaporation. Black holes scramble the initial information much more thoroughly than any non-gravitational quantum billiard. It's so thorough scrambling that the resulting radiation looks exactly thermal not only in the leading approximation but to all orders in perturbation theory.

    With this information, does it answer your question? I am not really able to convert it to a sharply defined, meaningful question.

  48. Dear Lubos,

    thanks for your answer.
    "you would have to define more accurately what counts as quantum chaos" - I will not even try. There is a huge literature on that, where smarter people than me try to come up with a proper definition. I'm still sorting through this jungle and try to understand, for instance, what they mean by "signature of quantum chaos".

    "Quantum billiard" is a very interesting hint. I know little about that, so I'll stop here and do some further readings, before saying anything unqualified.

    This discussion has made me rethinking several things, thanks a lot.


  49. I'm not sure where this comes from, since I'm not a religious person at all. To return to your analogy with the computer, today we commonly run programs in virtual machines. We can design a virtual machine in such a manner that it would be impossible for a program running on it to know that it is executing in a virtual machine, let alone map detailed features of the underlying OS. It's not even difficult.
    So while I don't believe in the simulated universe, I also don't buy so-called proofs that we aren't living in one.

  50. Dear scooby, you don't believe the proofs because you don't consider them to be what they actually are - rock solid evidence about natural sciences, not rigorous mathematical proofs - and because you're not approaching this question scientifically at all.

    If you're willing to adopt an arbitrarily unlikely and contrived assumption that is needed to reproduce the same predictions by a wrong theory that a correct theory explains naturally, you will be unable to understand this "proof" - but also every single other proof in natural science.

    Natural science can never 100.000000000000000...% eliminate all conceivable alternative theories. It may just prove that if an alternative theory exists, it is much more hugely fine-tuned than the established one, and therefore unlikely.

    This is the case of evolution theory, relativity, quantum mechanics, role of DNA in inheriting genes, every single other important advance in science, and this one is no exception. You apparently think that your reasoning is less unscientific than the common creationist thinking that God could have constructed the species exactly in this way, so that it appears that they evolved - but there is really *no* structural difference between your reasoning and creationism of the most naive kind.

    Creatures in a simulated universe that are independently intelligent in the same sense as we are will be capable of finding the discrete nature of time in their simulation and many other things by doing sufficiently accurate and clever experiments. If these things aren't there even in principle, then the statements that they exist at all are just untrue. Physically speaking, their world is *not* a simulation in their case.

  51. I agree, just happen to have a different opinion about what is likely and what is not in this instance.

  52. It's ironic that you allow yourself to insult people who disagree with your stupid superficial thinking and not allow any intelligent person who can argue and refute your claims to respond on this site .

  53. It doesn't strike me as any more bizarre (actually less so) than many worlds theory

  54. Hey. I'm afraid I still don't quite understand how Bostrom's Argument is debunked (probably since I lack a background in physics, math or any kind of science),but I find the prospect IMMENSELY depressing, so I would really really appreciate it if you took some time to explain this to me. It would really help me

  55. 'IF' we are in a simulation which would mean we are copied people living
    out someone else's life (because this is what 99.9% of simulations
    actually are) then there will absolutely be easily deducible observable
    evidence that this is the case.

    So, why hasn't anyone looked for
    observable MACRO evidence? I ask this because there is actually loads of
    directly observable evidence AND the only way you can explain why all
    'speculation' is stuck in 'experimental' evidence rather than HAVING A
    LOOK AROUND is that we are being directly managed to NOT think of
    stupidly obvious possibilities.

    For some points of observable evidence then read:

    a page discussing 'the simulation, attempting to simulate self aware
    free thinking people will directly manage the simulated population' to
    NOT think of obvious possibilities then try here:

    There are 30+
    pages on that site focused on the simulation argument / earth as a
    simulation and I'm putting a new page up just about every day