You’re Always On That Computer

One of the advantages of being a well-connected Internet sort is that people are constantly sending me interesting tidbits. From my friend Nick, who also provided yesterday’s Polka Floyd item, is one I hadn’t run across before (don’t know how I missed it, as it is right in amongst all the sorts of things I am always nosing about in).

It’s called the “simulation argument”, and it makes the case that there is a very good chance that we are living inside a computer program. This is not the same as the Matrix, exactly, because that is a trick being played on existing physical brains, whereas the claim here is that we may be nothing more than conscious programs running on a vast computer. It’s a fairly simple argument. We are given three propositions, and the author (Nick Bostrom, of Oxford University) asserts that at least one of them ought to be true. Here they are, as given in the abstract of the paper in which the argument was originally presented:

This paper argues that at least one of the following propositions is true: (1) the human species is very likely to go extinct before reaching a “posthuman” stage; (2) any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof); (3) we are almost certainly living in a computer simulation. It follows that the belief that there is a significant chance that we will one day become posthumans who run ancestor-simulations is false, unless we are currently living in a simulation. A number of other consequences of this result are also discussed.

So: either we die off before we get very far; or for some reason, even though computing power will probably increase immensely in years to come, we never decide to create conscious simulations. If neither of those statements is true, then it is likely that people do create such simulations, and after a while there would be so many more simulated consciousnesses than natural ones that the odds are that any given consciousness is a simulation. Ergo, if 1) and 2) are false, we are probably simulations.

There is a huge assumption here, which is that it is in fact possible to create conscious programs. I was seduced by that notion myself for a long time, but these days I “hae me doots”. The only entities that we know with certainty can create conscious minds are biological brains, and we simply don’t know enough about how they do that do be at all confident that they do so solely in virtue of their computational processing. So I think we must take this argument cum grano salis. Still, though, it’s an interesting idea, the kind of thing that can make you a little uneasy in those unguarded moments. Readers who are drawn to this sort of thing can find the paper here.

Related content from Sphere

5 Comments

  1. Malcolm,
    That this world exists is merely a nostrum
    Said a scholar of Oxford Nick Bostrum
    The truth of our senses
    Is but a concensus
    And a fiction like the cheque in the post.

    A deeply serious person. As well as everything else he manages the ‘Nick Bostrum’ event which we are now participating in. Going to his home page you will read that as an undergraduate he set a record in Sweden which I thought was a brilliant English joke but it was just my lateral Irish mind. In fact he seems to be from Sweden which is non-explanatory but significant. Say ja with a rising inflexion, then say ja with a falling inflexion. That’s about the range.
    Best Wishes,
    Michael.

    Posted May 11, 2007 at 8:06 am | Permalink
  2. Kevin Kim says

    I remember reading this argument a few years back in a book called Taking the Red Pill: Science, Philosophy, and Religion in the Matrix. (The book is on my shelf at home, but I’m typing this at the office. Otherwise, I’d have a page ref. for you.)

    Regarding this:

    There is a huge assumption here, which is that it is in fact possible to create conscious programs. I was seduced by that notion myself for a long time, but these days I “hae me doots”. The only entities that we know with certainty can create conscious minds are biological brains, and we simply don’t know enough about how they do that do be at all confident that they do so solely in virtue of their computational processing.

    I hae similar doots, but if technology is truly expanding exponentially, as some say, then one thing we might see in the future is the eventual erasure of the software-hardware distinction. Perhaps the artificial consciousness will be the software, and the software will be nothing more and nothing less than the hardware (nano-nanotech?).

    A physicalist explanation of the mind-body problem seems to be a variation on that theme: mind, the thing we might call “software” in a hardware/software analogy, is simply a form of matter, i.e., hardware.

    Just a thought.

    Kevin

    Posted May 12, 2007 at 3:09 am | Permalink
  3. Malcolm says

    Hi Kevin,

    Right, but this is different in that we aren’t talking about a single instance of a mind, but a massive hosting environment in which billions of minds are not only instantiated, but interact with each other. If biological hardware is needed for consciousness, then I suppose we might imagine a super-brain (in a super-vat!)that could do the trick — perhaps we grow a new lump of neural tissue to add a new virtual person — but that certainly seems a lot clumsier that what was implied in Bostrum’s example, and anyway, it’s hard to see in what sense the simulacra mightn’t be considered actual, biological persons.

    I quite agree, after spending so much time reading, brooding, studying, arguing, and reflecting on the mind-body problem from all three angles — philosophical, scientific, and introspective/mystical — that the mental/material categories may not be distinct in the way that is conventionally imagined, and that a new understanding of both consciousness and matter will be necessary to get anywhere with these problems in a productive way, either philosophically or scientifically. Easy for me to say, of course.

    Posted May 12, 2007 at 1:31 pm | Permalink
  4. Nick N. says

    I quantified my doubt of this postulate at about 10%

    Is that enough salt for you?

    Or do you just think it is false?

    And granting it for the sake of argument would you pick a #.
    As I said in email I still think 1.
    3 would be nicer though still quite disturbing.

    Posted May 13, 2007 at 4:53 pm | Permalink
  5. Malcolm says

    Hi Nick,

    I’m not sure what you are asking for with “pick a number”…

    I don’t see any reason to assume 1) is true, and I think that yes, if it is possible to make conscious simulations, folks will do so. But I am skeptical that that is in fact possible; the assumption that the brain generates consciousness solely in virtue of something about its computational activity that can then be abstracted into a computer program seems quite audacious to me.

    Posted May 13, 2007 at 10:50 pm | Permalink