8 min read

The Computer-Simulation Theory Is Silly

The Computer-Simulation Theory Is Silly

In recent years it's become fashionable among the technologically inclined to entertain, in all seriousness, the idea that we're living in a computer simulation. AIs in the far future, the theory claims, will run simulations in which the participants are conscious, despite being no more than sophisticated computer algorithms in a base reality they cannot perceive. The immediate seeds for this theory were advances in artificial intelligence and computer simulations of virtual worlds that occurred in the late 20th and early 21st century.

In this essay I'm going to argue that the computer-simulation theory is silly. The word is chosen intentionally. I will not argue that the computer-simulation theory is definitely false, nor even implausible. The computer-simulation theory is potentially true, but silly. More precisely: it is not a theory that well-informed people should give any special credence relative to the wide array of other possibilities.

1. The computer-simulation theory is unoriginal.

There are many famous precedents for the idea that life as we know it is a simulation. Any educated person should be familiar with some of them. In ancient China, Zhuangzi wondered whether he was perhaps a butterfly dreaming he was human. In ancient India, sages claimed that everything we perceive is an illusion created by Brahman for the purpose of play or sport. In ancient Greece, Plato proposed that our perceptions are like shadows on the walls of a cave wherein we are imprisoned, while base reality is outside the cave. In early modern France, Descartes imagined everything we perceive might be an illusion created by a mischievous demon. In 18th-century Ireland, Berkeley argued that our perceptions are continually produced by God and not the products of interaction with base reality. Leibniz had similar views, different in some particulars. And in the 20th century, Husserl proposed we set aside the question of whether the things we perceive do or don't exist and attend instead to our perceptions themselves.

None of these theories drawn from the canon of philosophy are precise duplicates of today's computer-simulation theory, and a naive reader might assume they're merely inferior precedents. To the contrary. Even the least of them is broader, more thoughtful, and less silly than the computer-simulation theory. But to understand them one needs a modicum of imagination, and the more sophisticated theories require both intelligence and time as well, because, for instance, it's not immediately obvious how the consistencies in our experience could be possible in the absence of an external world.

The majority of intelligent people are understandably too busy with practical matters to devote this necessary time. The computer-simulation theory, in contrast, is immediately comprehensible to anyone who's played a video game, with no need for thought, time, nor even imagination, and it's this easy comprehensibility, rather than any particular merit, that explains its popularity.

I cite history mainly by way of introduction. In what follows I'll dissect the flaws in the theory itself.

2. The computer-simulation theory is unimaginative.

The computer-simulation theory contains already in its first word a fatally unimaginative assumption, which I'll explain in stages.

You know from playing video games that video-game physics are not identical to real-world physics. Depending on the game, they can be very different. So if you're in a computer simulation, why would you assume that the world outside the simulation follows the same physics as the world you're experiencing within the simulation? There's no reason to make this assumption, and indeed you should not make it. And once you abandon that assumption, you must abandon in turn the assumption that the computers running the simulation are physically identical to the computers within the simulation.

Now, let's push further. If the world outside the simulation can follow different rules than the world inside the simulation, why must computers as we know them be involved at all? Our own brains can simulate experiences without any computer involvement, for instance in the form of dreams. Thus, the simulation may run through other means that are beyond our ability to know. All we can say is that it must require very powerful information processing.

Traditionally this very powerful information processor was called God; but considering the concept of God too incredible and also rather out of fashion, proponents of the computer-simulation theory found it more plausible to propose that our world and everything in it was instead being generated by an extremely large cluster of GPUs.

It's GPUs all the way down.

It should be clear from the above that insisting we're in a computer simulation rather than just a simulation is arbitrary and unjustified, and never had any basis except its adherents' familiarity with computers. We can correct this prejudice by removing the computers from the theory. Yet if we reduce the computer-simulation theory to a mere simulation theory, we no longer have a computer-simulation theory at all, but rather an analogue to one of the earlier and better philosophical theories cited in the preceding section. And we're not done yet.

Bad Math

The strongest argument for keeping the "computer" in the computer-simulation theory is presented in a quasi-mathematical form: the number of simulated human lives in the future, it is claimed, will greatly exceed the number of real human lives in the present, making it highly likely that a majority of experienced lives are simulated. But this math relies on three dubious presuppositions. It presupposes (1) that beings in the future will want to simulate many versions of present-day humanity to a high level of detail, (2) that they will consider this worth the cost of running a vast and extremely sophisticated simulation, and (3) that these electronically simulated humans will be conscious. None of these three propositions are provable or proven, and the first two aren’t even particularly plausible. It’s difficult to decide whether this mathematical argument makes the computer-simulation theory slightly less silly, or slightly more so, but in either case it remains very silly on net.

3. The computer-simulation theory is three thousand years behind state-of-the-art philosophy.

Suppose you're plugged in to a completely convincing simulation. One day you unplug and step out of the matrix. You realize you were fooled, and that now you're finally in base reality. Or are you?

This hasty conclusion—that leaving a simulation could possibly prove you're in base reality—recapitulates an error made in early attempts at philosophy. Even if you can find evidence that you were previously in a simulation, you can never prove that you are not still in some other simulation. There might be nested simulations, continuing infinitely, and all equally convincing.

There is no last doll.

The very concept of simulation is dependent on the existence of an “outside-the-simulation” that would constitute a superior level of reality. Yet as soon as we try to take the simulation theory seriously, we see that we can never establish the existence of a final outside. No world you ever experience can be proven more than apparently, convincingly real. And on account of this, the claim that we're just experiencing a very convincing simulation loses all meaning.

This observation isn't remotely original. It's an old philosophical commonplace in response to an even older argument, and neglecting to consider it is, again, quite silly. Zhuangzi, in proposing that he might equally be a human dreaming he was a butterfly or a butterfly dreaming he was a human (a metaphor implying there is no determinable “outside” and therefore no simulation properly speaking), saw through the error before he even bothered to make it, and he didn't have an abacus, let alone a GPU. Descartes concluded his experiment in hyperbolic skepticism with the observation that only faith in God could assure him his perceptions were true and meaningful.

The “outside” of all life is always an undiscovered country, guaranteed by faith or not guaranteed at all. The inside is always shadowed by uncertainty about what surrounds it. This is a great mystery of our condition, to which the predominant answer bequeathed us by tradition, echoed by Descartes, is that we should have faith that our lives here are as meaningful as they seem, albeit perhaps not in quite the way they seem during our pettier moments. Is a new faith that the undiscovered country will turn out to be a big server farm in a universe dominated by machine intelligence really a better answer?

4. The computer-simulation theory is not worth taking seriously.

After striking both the “computer” and the “simulation” from the computer-simulation theory, there's not much left. It's just a bad old idea with a modern technological veneer. Pushing it is like putting an Apple IIe in a new gaming computer case with lots of flashing LEDs and selling it to your grandmother as a superhuman AI. It's only persuasive to uninformed people. I don't blame you for being uninformed, because I know you're busy doing much more useful things than I am in writing this essay, things about which I am myself quite uninformed. But now that you are informed, you should ask for your money back.

UFOs could be real aliens. Ragnarok could arrive in the year 2222. We could be simulations running on computer software. Many things could be true. When we can't prove the answer either way, the question becomes: are these sensible ideas to entertain, or silly ones? Or, as I put it in my essay on religion: are they pragmatical ideas, or unpragmatical ones?

The proposition that we're living in a computer simulation falls short on both counts. Belief in it provides no benefits over belief in the earlier, better formulated theories in the same genre, and its implicit denigration and despiritualization of life is a notable demerit joined by several others I won't bother to list. It's an inferior copy of idealist theologies tech bros have long dismissed as too icky and too full of “woo.” And all of this will remain true even if we pop out of the matrix tomorrow, because it's confused on the most fundamental level.

If you're an adherent of the computer-simulation theory, you should ask yourself whether you'd perhaps like to reconsider the aforementioned icky ideas, or whether you'd just like to go on living, as you did before, with no particular theory about the nature of our world. Or, of course, you could continue being silly.

Also by J. Sanilac:

Memoirs of an Evil Vizier – evil is real

Trust Networks – how we actually know things

Dispelling Beauty Lies – the truth about feminine beauty, including practical advice for women

Ultrahumanism – a middle path through the jungle of modern and future technology

A Pragmatical Analysis of Religious Beliefs – are pragmatism and belief opposites?

Against Good Taste – aesthetics and harmful social signaling

Critique of the Mind-Body Problem – it's not solvable

End Attached Garages Now – a manifesto

Milgram Questions – what they are and how to call them out

Amor Fatty – how an obesity cure will end the body positivity movement