Elon Musk is one of the most intelligent and influential men in tech. Yet during his appearance at Recode's 2016 Code Conference, he said, "There's a billion-to-one chance we're living in base reality". In other words, Musk believes it likely that our world, and our lives, are artificial simulations (like in The Matrix).

It's tempting to scoff and wonder how someone like Musk could believe that perhaps nothing we experience is base reality. But when understood properly, it's hard to disagree with the argument Musk relies on.

This line of thought goes back hundreds of years. It was during the 17th Century that Philosopher René Descartes suggested that there is simply no way of knowing that our minds are not "brains in a vat" (again, like The Matrix).

This argument was rekindled and modernized in 2003 when Oxford Philosopher Nick Bostrom published his paper, Are You Living In a Computer Simulation?, the same paper that largely influenced Musk's views on the subject. This is therefore what we need to understand to grasp what it means to say that there's a "good chance we are living in a simulation".

Is That Kind of Tech Even Possible?

Is it even feasible that a computer could be built that could simulate the solar system, the world, and each of our individual lives? Even ten thousand years from now?

Futuristic Server Room
Image Credit: IBM Research via Flickr

The answer is yes. When it comes to computing power, we've barely scratched the surface of what's possible. As Musk explains:

40 years ago we had Pong -- two rectangles and a dot... Now, 40 years later, we have photorealistic 3D simulations with millions of people playing simultaneously, and it's getting better every year. And soon we'll have virtual reality, we'll have augmented reality.

Even if our rate of improvement slows dramatically, it'll only be a matter of time before "games will become indistinguishable from reality". As NASA scientist Rich Terrile says, "Soon there will be nothing technical standing in the way to making machines that have their own consciousness."

In more detail: According to Bostrom's research, based on our understanding of physics, simulating the entire universe down to a quantum level is unfeasible. But compressed representations of distant objects and ad hoc simulations of microscopic objects would dramatically lessen the computing power required. Given this, a rough estimate of a "realistic simulation of human history" would require about 1033--1036 operations per second. Given that Eric Drexler has given plans for a single computer the size of a sugar cube that could theoretically perform 1021 operations per second (among plans by other authors), we can assume that it is definitely possible to create a computer powerful enough to simulate the world as we understand it.

What Is the Simulation Argument?

Knowing that one day we'll probably be able to create these powerful simulations, we should ask... how do we know we're not in one right now?

When we think of how intelligent civilizations around the universe could progress (and have progressed), we have to admit that at least one of the following statements is true. Other than these three choices, there aren't really any other options:

  1. Virtually every civilization will go (or has gone) extinct before developing the ability to create these simulations.
  2. Virtually every civilization that has (or had) developed the ability to create simulations chooses not to do so.
  3. We are almost certainly living in a simulation.

Making Sense of the Argument

Most other articles on the subject brush over this explanation, but it's important to understand.

If (1) is true, practically every civilization fails to live long enough to get to this post-human stage, and that means no one gets the chance to create these simulations. As Bostrom says, "Maybe there is some highly dangerous technology that every sufficiently advanced civilization develops, which destroys them. Let us hope this is not the case."

Old History War Helicopters
Image Credit: manhhal via Flickr

If (1) is false, however, then a good number of civilizations will survive to be able to create these kind of simulations.

But just because they could create these simulations, it doesn't mean that they would. Perhaps none of these civilizations would have individuals wealthy enough to run such experiments. Maybe all civilizations would see such simulations as immoral. But is that really feasible?

We're already building rudimentary simulations. We're already trying to map the human brain. Many historians would love the chance to run an ancestor simulation. In that sense, what are the chances that every civilization would always avoid creating these simulations? (2) seems pretty slim, I'd say.

That leads us to statement (3), which is by far the most interesting.

If one of these civilizations is able and willing to create a simulation, they would probably run many versions of it. And if one civilization is running these simulations, chances are other civilizations will run them, too. Perhaps simulated people would also start creating their own simulations, and so on. Very quickly the number of conscious people inside of simulations would vastly exceed the number of people outside of a simulation.

By default, this makes the probability of me and you being in a simulation much higher than the probability of being in base reality.

And that's the crux of the argument. If some advanced civilizations developed the ability to create simulations, and actually ran them, we are statistically more likely to be in one of those simulations than not.

If in the future there are more digital people living in simulated environments than there are today, then what is to say we are not part of that already? -- Rich Terrile, NASA Scientist

What Are the Chances?

It's widely accepted that the Simulation Argument as described above is pretty water-tight. But that doesn't mean we know which of the three statements is true. There is no real evidence for any of them.

Dice Probabilities
Image Credit: David Lofink via Flickr

Speaking of his own argument, Bostrom thinks that the chances are pretty equal between the three statements. Another renowned Philosopher, David Chalmers, puts the odds of us living in a simulated world at 20%.

One the other hand, Elon Musk thinks the odds of us living in the "real world" -- base reality -- are less than one in a billion. And NASA scientist Rich Terrile says it is "extraordinarily unlikely" that we aren't living in a simulated world.

What Are the Implications?

Imagine that somehow we discovered that we were living in a simulated world. Bostrom believes that this shouldn't change our lives too much. We can continue to learn about our (simulated) world through scientific investigation. Our feelings and experiences will largely remain the same.

Man With a PANIC Sign
Image Credit: Jim Kelly via Flickr

But there would be an undertone to our lives that would change. We would now have a secular version of a "creator", and we could start trying to understand the "motives of the simulators". We would have a different understanding of our place in the world -- just like what happened when Copernicus discovered that the Earth was not the center of the universe.

An element of hope would be introduced. We could at last have the scientific possibility of an afterlife. Perhaps in death, we could continue our life in a new simulation (as was the premise of a Black Mirror episode).

Where to Go From Here

The simulation argument is little more than a philosophical thought experiment that's managed to capture people's attention. But whichever side of the fence you fall on depends on which of those statements you feel is most likely to be true.

No matter which you decide on though, there's very little we can do about it. Whether that's a good thing or a bad thing is open to debate.

Yet if the time ever comes for us to turn on our own simulation, thereby creating conscious, simulated people who do not know they are simulated, we have instantly shown (1) and (2) to be false. This leaves (3) to be the only option: that we are almost certainly in a simulation, too.

What do you think? Are thought experiments like this a waste of time, or is this argument actually on to something?