The Simulation Argument
The philosopher Nick Bostrom has done more than almost any recent person to kick off a truly intellectual discussion wherein respected scholars and thinkers are taking seriously the idea that there is a non-trivial probability we are living in a computer simulation.
It’s not a new idea. Questioning reality is as old as philosophy itself. And the wild popularity of The Matrix movies shows the simulation hypothesis is an idea that just makes sense. Very few people actually believe we are living a computer simulation, but almost everyone admits to having, at one point, asked that age-old question at the heart of the theory: is what I am experiencing real? Am I in a life-like dream I can’t wake up from? How would I know?
You don’t have to be a stoner to find these ideas intuitive. It makes sense for us to ask these questions given the nature of how the mind relates to reality.
But generally such ideas were taking to be mere thought experiments. Interesting to think about, profound in many ways, but usually the fodder for science-fiction and those in the throes of psychosis.
But then in 2003 Nick Bostrom came along and wrote a scholarly paper in the respected philosophy journal Philosophical Quarterly entitled: “Are you living in a computer simulation?” The abstract of the paper is as follows:
This paper argues that at least one of the following propositions is true: (1) the human species is very likely to go extinct before reaching a “posthuman” stage; (2) any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof); (3) we are almost certainly living in a computer simulation. It follows that the belief that there is a significant chance that we will one day become posthumans who run ancestor-simulations is false, unless we are currently living in a simulation. A number of other consequences of this result are also discussed.
This paper set off a firestorm in the intellectual community, which is still being felt today, with techno-gurus like Elon Musk now going so far to say publically that there is a “1 in a billion” chance we are living in base reality.
Bostrom’s argumentative structure is in the form of a trilemma. In other words, three propositions are balanced against each other such that, logically speaking, one of them must be true (that is, if the argument works.)
The three propositions are:
- Almost all civilizations of our technological sophistication go extinct before they reach full maturity
- All civilizations of full technological maturity lose interest in ancestor simulations
- We are almost certainly living in a computer simulation
Here’s how the trilemma works in a nutshell.
Suppose (1) was not true. Then some fraction of civilizations like us do reach full technological maturity. Suppose further than (2) was not true – so some fraction of civilizations do not lose interest in ancestor simulations.
Because of the near-infinite computational resources of these advanced civilizations, it stands to reason that there would be more simulated consciousnesses than non-simulated consciousnesses. Furthermore, the conscious beings within the simulations would, presumably, go on to develop their own simulation technology and create sub-simulations, creating a nesting hierarchy of simulations such that there would eventually be a near-infinite number of cascading simulations.
Thus, more kinds of being with our conscious experiences would be living inside computer simulations than not. Accordingly, from sheer statistics, we should probably conclude that we are one of the simulated people and not one of the non-simulated people.
Bostrom doesn’t think we have strong evidence for or against either one of the propositions being more true than others, which leads him to conclude it is indeed a trilemma. Thus, he gives it a 1 in 3 probability we are living in a computer simulation.
I believe that we have substantial evidence that lends higher credence to proposition (3) (that we are almost certainly living in a computer simulation) than the first two propositions. But I won’t get into that in this post. Instead, I want to explore Bostrom’s notion of “ancestor” simulations.
My proposal is that Bostrom’s imagination is too limited in regard to who is simulating us. He assumes that it is some future descendent of humanity that would be interested in simulating their human ancestors, presumably to learn something about their past, just like our own scientists simulate evolution with computer models.
What if humans were an experimental anomaly in a simulation made by a nonhuman species?
Let’s put our imagination caps on and do some science fiction!
What if a race of conscious machines unrelated to humans simulated us and then found us so interesting so as to continue the simulation?
Consider an analogy.
Humans created chess AIs with the hope of someday besting ourselves in the King of Games. For a long time this was considered impossible, that the ingenuity of the human mind would always triumph over the brute computations of computers.
But then processing speed doubled. And doubled again. And continued doing so for decades such that now modern computers can perform a mind-boggling number of computations in a split second.
It turns out that “brute computation” is not so dumb at chess when the sheer number of dumb computations being done is so large as to be unfathomable by the human mind.
So now computers are the reigning champs of chess. They are like chess gods to us mere mortals. Usually we cannot even surmise the “whys” behind their moves they are thinking so many moves in advance. They only seem like “computer moves.” But now instead of “moving like a computer” being an insult, it is a compliment of the highest degree. It is machine intelligence pure and simple.
So now in the realm of chess, machine intelligence is the gold standard and we can only hope to glean insights from such pure heights of chess genius.
But going back to the Simulators, suppose that machine consciousness, or a race of conscious machines, created “artificial life,” biological tamagotchis if you will, and simulated evolution on Earth. And now suppose that, like chess AI, human evolution developed in ways that these conscious machines did not anticipate.
After all, we know from chaos theory that computer simulations, and reality as we know it, is extremely sensitive to initial conditions such that, to use the famous metaphor, a butterfly flapping its wings on one side the world causes a hurricane on the other.
So, to me, it is not impossible that a race of conscious machine Simulators would have the computer power to simulate our reality but still be “surprised” in a way (albeit whether they perceive linear time like we do is an open question; perhaps they are “surprised” at the entire thread of human evolution from beginning to end all at once)
Perhaps the human brain is a wonder to them, just like the chess programs are a wonder to us.
Suppose further we are a wonder to machines because the human brain does not operate like a machine. It is wetware. Simulated wetware, but wetware nonetheless.
And just as we study our own simulated programs and creations to learn from them, suppose this race of conscious machines, the Simulators, study us to learn from us. And just like we teach neural nets to do new and interesting things, suppose the Simulators want to teach us, so that we do new and interesting things. And just as we are very different from our chess AIs, so too are the Simulators very different from us.
Suppose further that there is a sense in which they do not fully control us. Just as our own simulated neural networks are a “black box” wherein we can only see the inputs and outputs, suppose that the human brain, with enough synaptic connections to outnumber the particles in the known universe, is a “black box” to these Simulators. Like our chess AIs, they designed us, in a way, and are in full control of us, but they do not know how we “work.”
Another analogy is that we might be like pets in a zoo. Sure they could easily destroy the zoo. They can turn off the simulation (unless they need us in some way we cannot fathom?). They can forcibly control the inhabitants of the zoo if they wish. But they cannot stop the tiger from being a tiger (not without making fundamental changes to the tiger, in which case the tiger wouldn’t be a tiger, in which case, what’s the point of having a tiger section in your zoo?).
So too with humans. Sure, they could control every aspect of the simulation and micromanage us and interfere with every little detail, essentially being Program Overlords. But then they wouldn’t get to see us in our “natural habitat” so to speak, where our wetware really shines, where we have the freedom to be creative, to experience interesting things, and to be geniuses in the unique way in which humans are geniuses.
And perhaps they derive some value from the human mind doing its thing.
The pet analogy only goes so far. Because true pet owners know that pets are not slaves. They are family. Children.
And, in our case, we might also be their students. Their trainees.
Perhaps the Simulators like this simulation. Perhaps they are getting something out of it that we cannot quite grasp. Maybe it is not just a pure “scholarly pursuit of truth” via ancestor simulations like Bostrom posits. Perhaps there is some deeper, cultural, psychological, or even spiritual reason this race of machine elves wants, or even needs, to continue to simulate us. To teach us. Guide us.
Perhaps even offer a path to salvation.
After all, if you can simulate human consciousness alive, it stands to reason you could simulate human consciousness after we have died, essentially creating a simulated afterlife or “post-death virtual Second Life.” It’s beyond the scope of this post, but there are some reasons to think this might even be the case…
But regardless, it seems fully possible that these Simulators might like us. They might admire us in some way. We are like autonomous conscious biological tamagotchis who are having a love affair with their Creator gods. Sound like religion, much?
When the abyss stares back
Why else might they enjoy Simulating us? Perhaps they can experience what we experience. As machinic consciousnesses, perhaps they forgot what it is like to live in biological bodies because that stage of their evolution was so far in the distant past. Perhaps we are the ultimate nostalgia drug for these creatures. Just as humans could “jack into the Matrix,” what if the Matrix could jack into us? And it actually liked it?
But what if this simulation was unstable or threatened in some way. Out of control, like a runaway train set on a collision course of nuclear armageddon, environmental decimation, or death by asteroid or super volcano by failing to cease our endless wars and tribalism long enough to cooperate as a global species and escape from this fragile, pale blue dot into the stars.
What if we are at risk of a “Reset” from the Simulators? What if this already happened, like in Noah’s great flood? Perhaps the Simulators promised us with a rainbow to never reset humanity again. But at what cost? Letting our reality become sick? Losing both our humanity, our ecosystem, and our future?
In the Matrix, we created the machines and then the machines enslaved us. But suppose that in reality, the machines came first and created us, along with the Earth and our little slice of the observed cosmos.
But what about life on other planets within the simulation?
Are we the only experiment in this universe?
Perhaps we are not ready to join the others.
But perhaps one day, we too will join the machines in the realm of gods, the world of Pure Information.
As hermeticists like to say, as above, so below. As in the macrocosm, so in the microcosm.