The Simulation Argument

Are you living in a computer simulation?

That is the title of a 2003 paper by a guy named Nick Bostrom, who is a Swedish philosopher, researcher, and expert on AI safety, among other things. In short, he's a super smart guy.

I recently read this paper because I've been going deeper down the simulation theory rabbit hole, and I came to know that this paper is an important piece of the literature and of the modern-day simulation theory discussion. So I wanted to check it out. I've also been listening to some podcasts and a few interviews with Nick Bostrom.

To be honest, some of it went completely over my head. Some of the terminology, some of the equations that Bostrom lays out, and some of the other computer science and philosophy concepts just went a bit over my head. I'm sure if I spent the time, I could dig in and understand a little bit more. I feel that sometimes the terminology and language used amongst academics, scientists, and researchers is a natural barrier that prevents someone like me from understanding. But if you can translate the terminology, then I think those things can be understood.

Nevertheless, I think I understood the main points of the paper. And really, what the paper lays out is something that's known as the simulation argument.

So there's simulation theory, which proposes that we might be living in a computer simulation. Then there's the simulation argument, which, as I understand it, is more of a thought experiment. It's kind of a tool for thinking about the probability that we might be living in a computer simulation. So I'll attempt to break down the simulation argument and my understanding of it, and try to explain it in a simple way.

The simulation argument is made up of three different parts. Really, they're three different statements. And the thinking behind the simulation argument is that at least one of the three statements must be true.

The first statement says that civilizations like ours across the universe never reach technological maturity. They never reach the point of advanced technology at which they could create hyper-realistic simulations of people like us, with our experiences. So basically, the first statement is saying that civilizations across history and across the universe always die out before they reach a certain level of technological maturity. Before they reach a singularity-type moment, or a certain threshold of technology, something happens. They wipe themselves out, or something else wipes them out. They go extinct before they ever reach that level.

The second statement says that civilizations do reach that point of technological maturity. In the paper, Bostrom calls it "posthuman." So the second statement says that civilizations are able to reach that point without wiping themselves out, and they do have the capability to run advanced, hyper-realistic simulations that include conscious beings. But for whatever reason, almost all the civilizations that reach technological maturity choose not to run those simulations. Whether it's because they're not interested, or because there are ethical reasons not to, or they just don't find it valuable or entertaining, they just don't run them.

And then the third statement says that we are almost certainly living in a computer simulation. The third statement is quite straightforward.

So that's the simulation argument in a nutshell. To my understanding, anyway; I think I'm close to the mark on it. The overarching concept is that at least one of those three statements must be true. And that allows some interesting analysis to occur. It's an interesting thought experiment to think through each of those statements and to ponder them one by one.

One of the overall themes is that if you reject the first two statements, then you have no other option than to accept the third: that we are almost certainly living in a computer simulation. Because if civilizations do reach technological maturity, and even a small fraction of those technologically mature civilizations choose to run simulations, then, following that line of thought, the total number of simulated minds and simulated conscious beings in the universe would vastly outnumber the minds that exist in base reality.

So it's a fairly exciting and intriguing thought experiment to run through. And I've been doing that over the last couple of weeks.

A few thoughts on the first two statements.

The first statement is that civilizations die out before they reach technological maturity. I'm inclined to believe that that is not true. And I'm mainly inclined to believe that because if it were true, it would be a bummer. I feel that it would be too dark and too bleak of a scenario to actually be true. Now, I don't know if there's any principle behind that, but that's my gut feeling. My gut feeling is that there must be a way for civilizations to move up the Kardashev scale, so to speak, and to harness the full power of planets, and of stars, and even galaxies and beyond.

I think there's a greater mystery beyond what our current understanding can comprehend. If I think about where technology has gone in the last 50 or 100 years, and I extrapolate beyond that, and I think about the vastness of the universe and how many other civilizations might be out there, then I'm inclined to believe that the first statement is just not true. There are many great mysteries to come in this universe, for humans and the human condition, wherever we end up. So that's why I don't think the first statement is true, and why I don't want to believe that it's true.

Now, as for the second statement, I think that's an even more interesting one to think about than the first: that civilizations do reach technological maturity, but they're not interested in running hyper-realistic simulations of conscious beings.

I'm also inclined to believe that that is not true. Mainly because I really don't think, ethically, that advanced civilizations would be opposed to creating simulations. And I think there must be some sort of value that would be gleaned from creating simulations of conscious beings.

The reason I say that is because if I think about myself as a simulated conscious being, if we are indeed living in a computer simulation, then I feel incredibly grateful that someone or something or some intelligence created this simulation that I'm existing in. Because I get to exist. I get to experience all of this. I get to feel things and learn things and have kids. I have two young kids and a wife, and I have interests and hobbies and projects that I'm working on.

So me, myself, I feel incredibly grateful that I get to exist, even if this is a simulation, and regardless of the suffering that I've been through and the suffering that I might go through in the future.

So if anything, I think there is perhaps a moral obligation for intelligent, posthuman beings to create simulations of conscious beings. They're creating life. They're creating new consciousness from nothing.

So I guess I'll leave it at that. Those are my main thoughts on the matter. Obviously, we cannot know definitively what the answer to the simulation argument is, but I find it quite a tantalizing thing to think about.