Why You Should Hope That We’re Living In A Simulation

On balance, simulations have a lot more going for them than “reality” — whatever that is

So far I’m really enjoying Reality+: Virtual Worlds and the Problems of Philosophy, a new book by the philosopher David J. Chalmers. 

I think Chalmers is best known for formulating the so-called “hard problem of consciousness,” which boils down to — of course I’m oversimplifying here — “How can mere stuff give rise to subjective experience”? But as its title suggests, Chalmers’ newest book is about virtual worlds. 

A lot of it is devoted to exploring the philosophical ramifications of the possibility that our universe is simulated. There’s been an explosion of academic interest in this idea since 2003, when the philosopher Nick Bostrom published a pathbreaking paper titled “Are You Living In A Computer Simulation?”

The concept wasn’t new, of course — The Matrix had come out just four years prior, sparking a giant uptick in popular interest in simulated worlds, and hundreds of years before that, Descartes was messing around with the idea of an evil demon with similar deception capabilities. But Bostrom innovated on it in an elegant way. Basically, he showed that under certain reasonable-seeming assumptions about how likely civilizations are to stick around a long time (and to advance technologically while they do), and how likely they are to want to run simulations once they are capable of doing so, it seems at least pretty likely that we ourselves are living in a simulation. That’s simply because at a certain point, under certain conditions, the number of simulated beings in a universe comes to dwarf the unsimulated ones, meaning that any individual sentient being reflecting on its situation should accept that statistically speaking, it is more likely to be in a simulation than “top-level” reality — that is, one that was caused to exist solely by natural processes (or God, if you’re into that) rather than some sort of intelligence running a simulation.

The full, rather overwhelming version of this can involve a giant universe in which countless civilizations are simulating countless sub-universes, which (in some versions) themselves give rise to their own simulated sub-sub-universes, and so on. But for now, just to drive home the numerical intuitions underpinning this whole thing, let’s imagine a simple hypothetical in which — tomorrow — some scientist develops an app that allows anyone with a modern smartphone to simulate a miniature earth with a thousand sentient humans on it. 

If a billion people downloaded the app, then there would immediately be several orders of magnitude more simulated than nonsimulated beings (a billion people times a thousand sim-humans is way more than the 8 billion people who exist in our reality). And this ratio increases, perhaps quite dramatically, as technology improves and simulation gets more affordable and widespread — as more people can run simulations and each simulation can simulate more individual beings.

I’m drawn to Bostrom’s simulation hypothesis for the same reasons I’m drawn to the famous (to nerds) Drake equation, which offers an estimate of the number of advanced alien civilizations in the Milky Way. In both cases, we’re asked to make assumptions about numerical values we couldn’t possibly know, and then out pops an answer that —if true — would shake our understanding of the universe (though some of the terms in the Drake Equation, like the fraction of stars with planets orbiting them, are rapidly becoming more knowable thanks to modern astronomical advances).

I might have more to say about Chalmers’ book after I have finished reading and digesting it. But first I wanted to talk a bit about our intuitions with regard to the simulation hypothesis and why I think they might be ill-grounded.