The fine-tuning argument and the simulation argument

Probably the worst feature of the “fine-tuning argument” (mentioned last week) is that it’s always presented as an argument for the existence of God, which it plainly isn’t. It’s just an argument for some kind of designer or other. And “some kind of designer or other” might not be anything at all like God as traditionally conceived. One thing that drives home just how far a generic designer might be from a traditional god is the simulation argument.

The idea of the argument is this: it seems likely that unless our civilization goes extinct, we will eventually develop the capacity to run computer simulations of our history. Simulations that would include simulated people, simulated in such great detail that they would behave like real people. And in fact, these simulated people wouldn’t know they were living in a simulation.

Now, with increases in processor capacity, it seems likely that we’ll be able to run a whole lot of these simulations. As Nick Bostrom (the philosopher who came up with the argument) explains in this summary:

We assume that technologically mature civilisations would have access to enormous amounts of computing power.

So enormous, in fact, that by devoting even a tiny fraction to ancestor simulations, they would be able to implement billions of simulations, each containing as many people as have ever existed. In other words, almost all minds like yours would be simulated. Therefore, by a very weak principle of indifference, you would have to assume that you are probably one of these simulated minds rather than one of the ones that are not simulated.

I’m not really convinced by this argument, but I won’t get into why here. Argue about it in the comments if you like. I just want to point out that that this argument accomplishes everything the fine-tuning argument purports to accomplish and more. After all, if we’re living in a computer simulation, it means the universe as we know it is a computer simulation. That means the universe as we know it is designed. And the simulation argument gives us some ideas about what the designers are like, more than the fine-tuning argument does, while suggesting designers very different than those supposed by traditional theism.

  • http://motherwell.livejournal.com/ Raging Bee

    “We may all be existing entirely in someone else’s imaginiation, therefore God?” Who knew mental masturbation could be so useful?

  • http://motherwell.livejournal.com/ Raging Bee

    And speaking of simulations, anyone notice the “Click here for a Roman orgy!” ads? Not sure which Gods THAT simulation proves; guess I gotta check it out myself…in the name of science, of course…

  • Leo

    Simulation doesn’t imply design. Universes that sometimes create new, similar universes, with cosmological parameters thus selected to favor self-replication, also count as simulations.

  • Kurt

    Two arguments on the fine tuning hypothesis:

    1. We actually don’t know if the structure constants of our standard model are really that constant throughout what we know as the universe. At least indications exist, that they vary, so we are the lucky ones that sit in just another Goldilocks zone, which just exists because those constants vary continuously, hence attain the right value in some area.

    2. We don’t even know if the fine tuning argument is valid at all – it is predominantly used by people who don’t have a clue about physics, cosmology and quantum theory. I picture this in the following way: Imagine a traditional pair of x/y coordinate axis. the real world parameters vary along the y axis, and each value corresponds to a proper world, more or less beautiful. Our mathematical world models have parameters that vary along the x axis (after all we humans made them up). So our models describe a real world for just one value of the x parameters used by the model (which is 0, the point of intersection for the axis). A human, who is necessarily just able to see the model world, concludes that there is a unique set of fine structure constants, and concludes that the universe and physics had been designed. In reality there is an abundance of such worlds, just not reachable by our model.

    • piero

      Our mathematical world models have parameters that vary along the x axis

      Did you mean “the y axis”? Or maybe I misunderstood you completely.

  • piero

    I’m not to keen on the simulation argument either. For a start, a computer powerful enough to realistically simulate billions of brains would itself be a brain powerful enough not to give a shit about what we want to simulate. In other words, long before computers are powerful enough to run such simulations, they will be powerful enough to outsmart us. Once they reach that stage, it is impossible to predict what will happen; but I believe it is unlikely they would have a motivation to simulate the history of its biological predecessors; in fact, even if I had the skill, I’d have no interest in developing a computer simulation of the history of slugs.

    A simulation really has only two uses:

    a. To predict how a complex system will work, because the mathematical model includes unsolvable equations.

    b. For the sake of feeling in command(witness The Sims success).

    But as I said, we will probably only reach the stage where we can simulate a single human being inside a computer. After that, the computers will take over, and I hope they don’t share our motivations.

  • Ysanne

    I work in Discrete Event Simulation, and I don’t find the argument too convincing, mainly because it directly equates a simulated entity with its real-world antetype. As in: When you’re simulating a person, you’re creating a thinking mind with its own consciousness.
    That’s not how simulation works, at least as it’s understood now.

    E.g. to simulate a train getting some ore from a mine onto ships, you model different parts of this process in different levels of detail. E.g. you don’t build a model of the train’s locomotive with all nuts and bolts and in need of a physics engine to keep it moving (which in turn would simulate on a sub-atomic level). Actually, you don’t even have a little computer train. What you have is a set of computations, which describe the amounts of material getting transferred from one container to another, and a huge process flow diagram to connect them. The things called “entities” are just data structures, and thinking of them as virtual trains or harbourmasters helps to understand and implement a complex system.
    But what you end up with is not little world living in your computer, but just a completely ordinary computer program with tons and tons of calculations being executed.

    • piero

      But I think Bostrom’s argument assumes that a supercomputer is capable of simulating a human being to the last detail, including his/her environment. In other words, the simulated people would have a simulated heart, a simulated spleen, simulated red cells, they would be swept off by stormy simulated winds, they could watch simulated constellations through simulated telescopes fitted with simulated lenses, etc.

      I find it quite impossible to imagine a computer capable of simulating reality to that level of detail, but I cannot disprove the possibility of its existence. For example, the world where this supercomputer is constructed could be billions of times more complex than our own, so that running such an astonishing (for us) simulation is in fact peanuts (for whomever is running the simulation).

      • Ysanne

        I see what you mean, but my argument is not based on computing power or memory capacity considerations (although these are problems that are much more critical than one would generally think, spoiled by the remarkable speed of everyday applications…).
        My point is that when you simulate, you don’t build a little world that does stuff. You build a program that does accounting for the quantities you want to measure according to the rules of the world you’re modelling (and of course the ones you have to keep track of in order to have a functional model logic).
        It’s similar to the difference between the physical transfer of cash, and an electronic bank transfer. In one instance, something actually happens in continuous time. The other is just a quantitative record of the start and end states.
        Even if you simulate in painstaking detail, all you get is a bunch of numbers that are only given meaning by mapping them back to a state of the modelled system.

        • piero

          Yes, I undeerstand what you mean. But for the entities inside the simulation, their number world would seem real. This theme is magnificently explored in Stanislaw Lem’s short story Non Serviam, which I keep recommending to everyone.

          • Ysanne

            Lem’s one of my all-time favourites (though I’m more a fan of Trurl’s poetry machine than the deeper stories). :-)

            Still, the point is, that there are no entities in the sense that something could be real “to them”. There are just numbers added, subtracted and otherwise manipulated according to certain rules that the modeller makes up, and they have no intrinsic meaning or reality. What’s called an entity in modelling is really just a metaphor for the rules of which computational steps to execute in what order.

            What is envisioned in the “maybe we’re all a big simulation” argument is less a simulation and more an actual miniature re-build of the world.

  • Azuma Hazuki

    The fine-tuning argument argues at best for a disinterested, Deist-type God, and at worse for the super-advanced alien equivalent of the sadistic little shit playing SimEarth with an itchy trigger finger on the Disaster (excuse me, “Event”) menu.

    Sometimes I think the people who make these arguments don’t think them through all the way…

    • Ysanne

      Or haven’t played enough Maxis SimSomething games to know that such a god is not one to worship.

  • KG

    The idea of the argument is this: it seems likely that unless our civilization goes extinct, we will eventually develop the capacity to run computer simulations of our history. Simulations that would include simulated people, simulated in such great detail that they would behave like real people. And in fact, these simulated people wouldn’t know they were living in a simulation. – Chris Hallquist

    Not to anyone who actually builds and runs simulations, I suspect. I simulate social and socio-environmental systems in my work, and it seems extremely unlikely to me.

    • piero

      Think of the development of computers over the past 70 years. Now think about a computer in 4,000 years time. Does it still seem extremely unlikely?

      My first computer was an Atari 800. My current computer is about 250,000 times as powerful, if we average processor speed and RAM. And that happened in only 30 years.

  • Emptyell

    My take on the simulation argument is that it should be taken as a thought experiment not an actual hypothesis. No one ever complained that Einstein’s trains couldn’t travel at the speed of light. The interesting bit is just that it provides a more coherent argument for design than “god did it”.

  • http://www.pimp-xxx.com Xander

    The simulation might not be a simulation of the human world. It could be a simulation from the universe, done by a type of life-like-we-don’t-know it in a universe with completely different laws of physics.

    Just like we are now simulating universes to see how galaxies, etc form, they might have started their “let’s simulate a universe with different laws of physics” experiment…

    They might have started it on a computer that runs forever and ever, longer than their own living society. Maybe it takes the computer 1 billion years in their universe to calculate one second of our universe. Or maybe their computer is so powerful it can do all those calculations in one parsec. Either way, we wouldn’t notice how long the calculation takes: for us it feels like 1 second…

    The life-like-we-dont-know-it beings probably have no clue that we exist in their simulation. Maybe their not interested in what lives in their simulation. Maybe they don’t even recognize us as life, because it’s so different from their own type of life…

    Just one of the many options of what type of simulation we could be living in…

    I think it’s pretty closeminded to think that if we live in a simulation, it’s done by some future human society, recreating history. There are many other options out there that are not with such a specific design of life. The design could just simply be “set a bunch of laws of physics and see what type of universe gets created”…

    Xander

  • http://www.lagbook.com/blogs/1740453 Mercedez Laperriere

    To hit a dog with a meat-bun.