No, We Are Not in a Simulation

Last week I wrote about the overlap of intelligent design with the simulation hypothesis and how belief that we’re living inside of a giant simulation seems like disguised theism. Yet interest in the simulation hypothesis doesn’t seem like it’s going away anytime soon. There must be some reason why that famous poseur Elon Musk believes we live in a simulation, right? I mean, he didn’t just read a random Reddit comment and change his entire view of the cosmos, right? Right…?

Nick Bostrom put forth his simulation argument over two decades ago based on a probabilistic trilemma, which is a fancy way of saying Bostrom used wordplay to present three false alternatives to gullible South African billionaires. In this trilemma, Bostrom argues that at least one of these three propositions is true:

1. Early Extinction: Humans (and other intelligent species) almost always go extinct before they reach a “post-human” stage (where they have the technology to run massive simulations).
2. Lack of Interest: Technologically mature civilizations have almost zero interest in running “ancestor simulations” (simulations of their own history or beings with consciousness).
3. The Simulation is Us: If the first two are false, then we are almost certainly living in a simulation.

Here’s the sleight of hand. Bostrom argues that if a single real civilization can create millions upon millions of simulations, then there are billions of sim-people in the universe yet only one real base reality behind the simulation. Thus, the odds of you being a sim-person are billions to one or greater. Basically, as soon as any civilization is sophisticated enough to produce simulations, simulated beings will vastly outnumber the real ones to such a degree that the odds of you being a real person are almost nil.

If your Spidey senses are telling you something is off in this argument just from a pure logic perspective, you’re right. It sounds kind of convincing and I’m guessing Musk’s entire education in logical reasoning came from reading the back of a cereal box so who can blame him for being hoodwinked. Let’s begin to dissect these false alternatives.

First, all kinds of other plausible possibilities exist. Maybe civilizations go extinct after reaching a post-human stage but before running simulations. Maybe simulations are run in limited fashions. What even counts as a simulation? Does Minecraft count? We have no way of estimating how likely it is that there even is an advanced civilization anywhere in the universe (we clearly are not it) so to jump from a funny suggestion that there could be many simulations to there must be billions of them is a leap too far. It is impossible to support or falsify any statement in the trilemma. Finally, the same arguments against intelligent design around the anthropic principle apply here as well. That is, these premises make sense to us because they are based on our own limited understanding of our lives as human beings and so become a kind of circular prediction. The trilemma is more of an interesting thought experiment than a sound philosophical argument, even if the rhetoric seems convincing to one famously bigoted windbag.

So much for the logical failings. There are reasons involving math, physics and computational complexity that suggest any kind of simulation at this level is outright impossible. One example of this comes from quantum mechanics where things get a little unruly when you look at small particles like electrons. Quantum mechanics indicates that simulating our universe would be much harder than it sounds on the back of Musk’s cereal box. I’m not a physicist, so if you’re interested in a better explanation of this than I can provide, do a search on the “sign problem,” and Physicists Zohar Ringel and Dmitry Kovrizhin. But briefly as I understand it, you need to add more particles to your simulated universe but when you do that, the amount of computer power increases exponentially not linearly. So, your memory requirements for the simulation grow so fast that you quickly need a computer with more memory than the total number of atoms in the observable universe. Do you see the problem? There isn’t enough stuff around in our universe to build a computer big enough to simulate the universe. Again, the trilemma isn’t based on any observable facts, it’s kind of just saying stuff to be provocative. No wonder it’s so appealing to the person who turned Twitter into a cesspool of click-bait misinformation.

In other words, the base reality that is creating our simulation would have to be much larger than our actual reality. Simulation proponents would argue there are workarounds. For example, the universe doesn’t actually exist anyway, it’s just being rendered for us when we need to see it like a video game. You know how when you play Minecraft and you quickly run to the edges of your world and the screen starts pixelating. That happens to save processing power. You only need to render what is in front of the sim people at any given moment. But it does seem like the Hubble Telescope complicates this argument. There would be a lot of the universe under some observation at any given point in time. Well then, maybe you don’t need to simulate everything, just the overall effect. You don’t need to simulate the atoms in your coffee cup unless a scientist is measuring it at the moment. You just need to provide the effects of hot, brown fluid upon the sim observer. Still, with billions of sim people, that’s a lot of rendering going on even if they don’t all have electron microscopes. OK, fine, then maybe nothing exists and you are a kind of one-person simulation, a brain in a vat as they say. The simulation only needs to create your personal experience of reality. Then why does my experience seem to match yours so tightly? Or are you just a nonplayer character in my game? Or am I?

Because simulation theory is really just a disguised and weakened form of classical theism, it is subject to some of the same arguments atheists make about God. One such argument is turtles all the way down, otherwise known as infinite regress. The phrase, “turtles all the way down,” come from a story where the world is supported on the back of a giant turtle. Then someone asks, “but who is supporting the turtle?” The answer is another, bigger turtle. You can keep this up forever until someone says, “it’s just turtles all the way down.” Replace “turtles,” with, “simulations,” and you have an idea of why the simulation trilemma fails. Many philosophers would disagree with me, but one of the reasons I believe theism is true is that you cannot have infinite regress, you need a first cause. The upshot is that if it’s so probable that we are in a simulation based on the third proposition of Bostrom’s trilemma, then it is equally possible that our own creators are themselves inside of a simulation, and so on up and down the chain ad infinitum. Yet, each layer of those simulations would require more computational power than before.

Wherever we happen to sit in this chain of simulations, I can tell you that the number Pi has an infinite string of nonrepeating decimals in our simulation right here, the simulation in which you find yourself reading this blog post. Yet, computers are all zeroes and ones and require the precision of discrete bits of information. The world doesn’t seem to work that way. If Pi, not to mention the rest of the universe, is infinitely continuous, then that would require infinite memory. The sim argument boils down to this: If humans ever become capable of running a simulation at this scale (note: they cannot), they will (note: they may choose not to), so there will be billions of simulated people (note: there may not be, we might limit them and you’re really testing the definition of what “person,” might mean), thus we are certainly inside of a simulation right now (note: we certainly are not).

If you find logic unconvincing, or if you suffer from a critical reasoning deficit like a Nazi-saluting transgender-hating tech bro, the argument about consciousness and intentionality is even more convincing. I’m not going to argue why I think consciousness is evidence of God here today, although I think that is true. I just want to argue that consciousness and intentionality, regardless of where you think they come from, help disprove a simulation.

The philosopher John Searle wrote about something he called the “Chinese Room,” to demonstrate why even the largest computer, even a computer larger than our known universe, cannot actually understand what it is doing. Computers do not possess any kind of consciousness or intentionality and despite what AI snake oil peddlers will tell you, computers will never become conscious. Imagine you are locked inside of a room with a lot of Chinese symbols on the lock and you don’t read or understand any Chinese. Luckily, someone left you a rulebook in English that helps you match the symbols together. Thanks to the rulebook, you can put the right symbols in the right order to open the door and escape. Your captors might think that you understood enough Chinese to open the lock even though you understand nothing. You just followed the instructions of your rulebook without any regard for what the symbols on the lock actually say. The symbols might say, “open this lock and the known universe will end,” but you don’t care, you just did what you were told in the rulebook.

This is what computers do. They follow a rulebook that tells them how to process symbols. They are just moving through a set of instructions without any idea what the symbols mean. I’m typing this blog out, but my computer is just putting symbols in order and based on what I type it is rendering some characters for you to read. It cannot make meaning from those characters, but you can. How are you able to derive meaning if we’re inside a simulation?

Consciousness also includes qualia, or a subjective experience that a computer simulation cannot reproduce. Imagine an AI robot that has been programmed to appear human. You stab the robot with a needle and the robot says, “%&$ you moron, that hurts,” or because AI chat has been sanitized in our culture, perhaps it says, “your actions are out of line, and I expect better from you.” But there is no actual feeling of pain. Or consider a plant. We can create plants in a model to simulate photosynthesis, but there is no resulting oxygen.  

The one thing I appreciate about some people’s instincts around the simulation hypothesis is that the universe seems more like a mind than a machine. I call that the mind of God, but you don’t have to. I’m out of energy for argumentation today, so you could alternatively say it is the mind of John Malkovich. I like that idea much more than thinking that consciousness somehow emerges from matter, because again, meaning does not emerge simply from a string of characters on this blog page, it emerges in your mind only because you are already a conscious being. I happen to think that consciousness is the ground and the point of creation, not the other way around. And if the universe itself is mind, then a simulation of it purposelessly redundant.