You are probably not living in a simulation

Nick Bostrom wrote this article called "Are You Living In a Computer Simulation?" for the Philosophical Quarterly, and it has gone viral. I know what you're thinking. You're thinking, "That's old news, Sam." Yeah, but I'm just getting around to writing about it. The reason I decided to write about it now is because I keep running into people on the internet who say they think we're living in a computer simulation. I doubt many people really believe that. I suspect it's just a fad to say that they do. Maybe some people do, though. I don't know. It still seems worth responding to, though, because I find it to be a big distraction when trying to have an otherwise serious conversation with people.

I don't think we are living in a computer simulation. I think we're real three dimensional people living in a real tangible world. I have several reasons for thinking this, but first let me say something about Bostrom's article.

The first thing to note is that Bostrom didn't actually argue that we are in a computer simulation. What he argued, rather, was that one of the following things are true:

(1) The fraction of human-level civilizations that reach a posthuman stage is very close to zero;
(2) The fraction of posthuman civilizations that are interested in running ancestor-simulations is very close to zero;
(3) The fraction of all people with our kind of experiences that are living in a simulation is very close to one.

By "posthuman civilization," he's talking about civilizations that have reached a stage of technological development in which they're capable of running ancestor simulations. The third possibility is the only one that entails that we are likely in a computer simulation, but he concludes that "In the dark forest of our current ignorance, it seems sensible to apportion one’s credence roughly evenly between (1), (2), and (3)." That means we can't know whether we're in a simulation or not. Given three equally viable possibilities, it would seem at best that there's a one in three chance that we're in a computer simulation.

But that isn't as interesting as it would've been if he had argued that we are almost certainly in a simulation, which is the conclusion a lot of people want to run with.

Now, let me give a few reasons for why I don't think we're in a computer simulation. First of all, it's because I give a lot of credence to intuitive obviousness. It's a fundamental part of my epistemology to assume that things are just as they appear to be unless we have good reason to think otherwise (See "It's always more reasonable to affirm the obvious than to deny the obvious"). So the default position any of us ought to take is that we perceive a physical world because there is a physical world. This seems to be the most natural and parsimonious assumption we could have. I suspect this is the primary reason most reasonable people will look at the simulation argument as a mere curiosity--a fun thought experiment--but won't take it seriously. And it's the reason I think a lot of people who claim to believe we are in a simulation are either not being honest with themselves or are not being honest with the rest of us. They're just playing games.

Second, I doubt we're in a simulation for some of the reasons Bostrom mentioned in his article. He talked about the enormous amount of computing power that would be necessary to run a simulation resembling our world. Think of the trillions of stars, galaxies, cells in all the living organisms, sand on the beach, grass in the plains, leaves on all the trees, the complete minds of billions of people and animals, and all the information stored on all the computers and libraries in the world. Bostrom recognized all that, but he was far more optimistic than I am. He imagined computers the size of planets and the discovery of new laws of physics that would make the seemingly impossible possible. These are mere possibilities and not sufficient grounds for any kind of optimism. He argued as if there are no limits to computational progress, but surely there are limits to what physical things can do. However much our technology improves, physical things still have to obey the laws of nature. Bostrom gets around some of these obstacles by suggesting that a simulation need not represent everything we think is in our world, but only what is currently being observed. Or, it may not need to simulate the billions of minds that seems to exist. It could just be you, and everybody else is a philosophical zombie. Of course that raises the problem of solipsism, which is another reason to be doubtful.

Then there are the further obstacles of whether we'll survive to reach the level of technology necessary to run ancestor simulations before something terrible happens to either wipe us out or set us back. And there's the further obstacle that even if we could build computers powerful enough to run ancestor simulations, why would we considering how expensive it could be in terms of resources. And what would be the pay off? There are ethical considerations, too, if we suppose these simulations will contain real conscious beings who will suffer. So there are all kinds of obstacles that, to me, make it unlikely that we'll ever run such elaborate simulations. To make it likely that we're in a simulation, it isn't enough to say that one simulation will some day be run. We have to suppose that multiple simulations will be run. You need there to be more people in simulations than outside of simulations in order to make it more likely that you are one of the people inside of a simulation.

A third thing to consider is whether a simulation of a human being would actually be conscious. This is not the question of whether computers can be conscious, which I'll address in a minute. This has more to do with what a simulation is. A simulation is a representation of something. It doesn't actually have the properties of that thing. It only has information. Any output you get from a simulation is still a simulation and does not actually have the properties of the thing it represents. If you could simulate a brain in such a way that it outputs things like thoughts, sensations, etc., there wouldn't actually be anything that was thinking and perceiving. You'd just get images or data that maybe you could read off a screen. Just as a simulation of a wet sponge wouldn't actually create anything wet in your computer, so also, a simulation of a brain wouldn't actually create anything conscious in your computer. You could simulate water molecules and multiply them in your simulation, and watch on your screen how they behave, but you'd never get anything wet in your computer as a result. The simulation would remain a simulation and wouldn't take on the actual physical properties of the three dimensional object it was representing. A brain is made of three-dimensional cells and atoms, and supposedly consciousness arises because of their physical structure and activity. But if you simulated these cells in a computer, the information in the computer would not actually take on the physical properties of the cells. The information would just be 1's and 0's. Since that information wouldn't actually take on the physical structure of brains cells, there's no reason to expect the simulation to be conscious.

Fourth, I don't think physical things can be conscious. I don't think conscious states arise purely out of physical processes. This reason has a lot to do with why I'm a substance dualist, but since that's a topic all its own, I'll leave it as an assertion here and not argue for it. In Bostrom's article, he begins with the assumption that not only are physical things capable of producing consciousness, but that the underlying physical substrate doesn't matter. If brains can produce consciousness, then so can computers.

And that brings me to my fifth reason. Even if we grant that consciousness arises purely from physical processes in the brain, nobody has a clue how that happens. Consciousness is completely mysterious. We can observe correlations between physical brain activity and mental states, but what we can't do is explain how we get from those physical states to the mental states. We have no idea what the mechanism is. Consciousness is especially mysterious because of its subjective nature. While physical processes are open to third person observation, conscious states are only open to first person observation. That means that if consciousness is a property of physical states, then physical states have invisible properties that can never be observed in a third person way. All the attempts I've seen at explaining this amount to a lot of handwaving. I'm convinced that nobody has a clue. With that being the case, we have no reason to think that a computer made of gold, silicon, and various other materials that work on logic gates, etc. would produce the same effect that a biological object like the brain would. There could be something about brains that is not true about computers that explains why brains are conscious and computers are not. People just assume it has something to do with the level of complexity, but nobody knows that. If computers are not conscious now, there's no reason to suspect that more of the same would make any difference.

The only way it's likely that we are in a computer simulation is if we have good reason to expect that some day in the future, we will be willing and able to create multiple computer simulations of entire worlds resembling our own, populated by real conscious individuals. But since we have no warrant for thinking such a scenario is likely, and plenty of warrant for being doubtful about it, it follows that we're probably not living in a simulation.