The Simulation Argument, Explained Without the Hype

Every few months, a headline announces that "scientists say we might be living in a simulation." The idea usually gets credited to a 2003 paper by the Oxford philosopher Nick Bostrom, titled "Are You Living in a Computer Simulation?" But if you read the paper itself, you'll find that its claim is both more modest and more unsettling than the headlines suggest.

Bostrom does not argue that we are in a simulation. He argues that at least one of three possibilities must be true. That structure — a trilemma — is what makes the argument so sticky. You can reject any one of the options, but you can't reject all three at once.

The three options

Here is the argument in plain language. At least one of the following is true:

• Almost no civilizations like ours reach a "posthuman" stage — a level of technology at which they could run detailed simulations of minds like ours. Maybe they destroy themselves, or maybe it's simply impossible.
• Almost no posthuman civilizations are interested in running such simulations. Perhaps they find it unethical, boring, or pointless.
• We are almost certainly living in a simulation.

Why does the third option follow if the first two are false? Because of numbers. If even a small fraction of advanced civilizations run "ancestor simulations," and each runs many of them, then simulated minds with experiences like ours would vastly outnumber the original, biological ones. If you have no special evidence about which kind of mind you are, then simple reasoning about proportions says you should bet you're one of the many, not one of the few.

The quiet assumption

The argument rests on a premise Bostrom states openly: substrate independence. This is the idea that a mind doesn't depend on being made of neurons specifically. What matters is the pattern of information processing. If you could reproduce that pattern in silicon — or in any suitable computing system — you'd get a genuine mind with genuine experiences.

Many philosophers and cognitive scientists find substrate independence plausible. Many others don't. If it's false — if consciousness requires something about biology that computation can't replicate — then simulated beings might behave like us without experiencing anything at all, and the argument loses much of its force. We'll come back to this question in another essay, because it's one of the deepest open problems we have.

What the argument is not

A few things are worth separating from the argument itself:

It isn't a scientific theory. There's no experiment it predicts, at least not directly. It's a piece of philosophical reasoning about probabilities, built on assumptions you're free to question.

It isn't a claim about "glitches." Online, people sometimes point to déjà vu or strange coincidences as evidence of a simulation. Bostrom's argument doesn't rely on anything like that. A well-made simulation, if one existed, might be indistinguishable from the inside.

It isn't brand new. Doubts about whether our experience reflects reality go back thousands of years — to Plato, to Zhuangzi, to Descartes. What Bostrom added was a specific modern mechanism (computers) and a specific line of reasoning (counting observers).

Common objections

Thoughtful people push back in several ways.

The computing cost is too high. Simulating every atom in a universe might be impossible. But Bostrom notes that a simulation would only need to render what its inhabitants observe, in enough detail to fool them — the way a video game doesn't draw the far side of a mountain until you walk there. Whether that shortcut is actually feasible is an open question.

The probability reasoning is shaky. The step from "most minds like mine are simulated" to "I'm probably simulated" uses a kind of reasoning about observers that philosophers still argue over. Some think it's sound; others think it smuggles in hidden assumptions.

We can't know what posthumans want. True — and that's exactly why the second option remains on the table. The argument doesn't tell us which branch of the trilemma is correct. It only says we can't comfortably believe all three are false.

Why it matters anyway

You might wonder what difference any of this makes. If the simulation is perfect, life goes on exactly the same either way. Your coffee is still hot. Your rent is still due.

But the argument does something valuable: it forces us to examine assumptions we usually never notice. What is a mind? What counts as "real"? Is a world less meaningful if it's built from information rather than from whatever lies beneath? Philosopher David Chalmers, in his 2022 book Reality+, argues that even a simulated world would be genuinely real for those living in it — its tables and trees would be real digital objects, not illusions.

You don't have to accept that conclusion, or Bostrom's trilemma, to find the exercise worthwhile. Questions like these are a kind of mental stretching. They loosen the grip of the obvious, and that is where curiosity begins.

So the next time a headline tells you scientists have proven we're in a simulation, you'll know the real story: no one has proven anything. A philosopher made an argument — and two decades later, we're still arguing about it.