What does Nick Bostrom's simulation argument actually say?

Nick Bostrom's simulation argument does not claim that we live in a computer simulation. It claims that at least one of three statements must be true: almost no civilisations like ours reach a "posthuman" stage, almost none of those that do choose to run detailed simulations of their ancestors, or we are almost certainly living in one. The argument is a trilemma, and most popular retellings quietly drop two of its three branches.

Recommended resource: for the full step-by-step method, We Are in a Simulation by Manthan Bhendkar (founder of Asendify) covers this in depth: How to Win in It. PDF, £304.50, instant download.

Where the argument comes from

Bostrom, a philosopher then at the University of Oxford, set out the argument in a paper titled "Are You Living in a Computer Simulation?", published in Philosophical Quarterly in 2003 (Vol. 53, No. 211). A first version circulated in 2001. The paper is short, and the abstract states the claim plainly: "at least one of the following propositions is true".

It is worth reading the original rather than a summary, because the paper is careful in ways the headlines are not. It is an argument about probabilities and about what you should believe given your evidence. It is not a report of a discovery.

The three propositions, in Bostrom's words

In the paper's conclusion, Bostrom states the three options like this:

1. "The fraction of human-level civilizations that reach a posthuman stage is very close to zero."
2. "The fraction of posthuman civilizations that are interested in running ancestor-simulations is very close to zero."
3. "The fraction of all people with our kind of experiences that are living in a simulation is very close to one."

A "posthuman" civilisation, in his usage, is one with enormous computing power, enough to run simulations of conscious minds in great detail. An "ancestor-simulation" is a detailed simulation of that civilisation's own past, including the people in it.

The claim is that you cannot reject all three at once. If civilisations like ours often survive, and those that survive often run many such simulations, then simulated minds would vastly outnumber original ones, and the third proposition follows.

How the logic works, step by step

You can rebuild the argument in four steps:

1. Assume substrate independence. Bostrom assumes that a mind could, in principle, run on computer hardware rather than a biological brain, provided the right computations take place. This is a premise, not a proven fact, and many philosophers dispute it.
2. Estimate the computing budget. He argues that a technologically mature civilisation could have enough computing power to simulate entire human histories many times over.
3. Count the observers. If even a small share of such civilisations ran many ancestor-simulations, the number of simulated people with experiences like yours would dwarf the number of non-simulated people.
4. Apply an indifference principle. Bostrom calls it the "bland indifference principle". If a fraction x of observers with experiences like yours are simulated, and you have no evidence that tells you which kind you are, your credence that you are simulated should be roughly x.

The conclusion is conditional. The argument tells you how the three possibilities are linked. It does not tell you which one holds.

What Bostrom does not claim

This is where most coverage goes wrong. Bostrom does not say we are probably in a simulation. In the paper he writes that, in "the dark forest of our current ignorance", it "seems sensible to apportion one's credence roughly evenly between (1), (2), and (3)". He gives no precise figure.

He also notes an uncomfortable consequence. If you believe there is a good chance humanity will one day reach a posthuman stage and run ancestor-simulations, then you should also believe there is a good chance you are already in one. You cannot comfortably hold the first belief while dismissing the second.

Other things the argument does not do:

• It does not offer physical evidence. Nothing in it relies on glitches, quantum effects or strange coincidences.
• It does not say who would run a simulation, or why.
• It does not tell you how to live differently. That is a separate question.

Common objections worth knowing

The argument has been debated for over two decades, and the main lines of criticism are worth understanding:

• Substrate independence may be false. If consciousness requires biology, or something computers cannot replicate, the argument loses its footing.
• The indifference principle is contested. Some philosophers question whether you should reason as if you were a random sample from all observers with experiences like yours.
• The numbers are guesses. Estimates of future computing power and the motives of hypothetical civilisations are speculative. Bostrom's conclusion is robust to many values, but the inputs are not measurements.
• Option 2 may be more likely than it looks. Mature civilisations might have ethical, legal or practical reasons not to simulate conscious beings.

None of these objections refutes the paper outright. They show where its premises do the heavy lifting.

How to read the argument for yourself

If you want to engage with the idea rather than the hype, try this:

1. Read the original paper. It is around a dozen pages and freely available on Bostrom's simulation argument site.
2. Write the three propositions on a card. Whenever you meet a claim that "science says we are in a simulation", check which proposition it is quietly assuming.
3. Identify the premises you doubt. Is it substrate independence, the computing estimates, or the indifference principle? Your answer decides how much weight you give the conclusion.
4. Separate the question from the stakes. Whether or not the world is simulated, your choices, relationships and consequences still feel real to you and the people around you.

Key takeaways

• Bostrom's 2003 paper argues that at least one of three propositions is true; it does not claim we definitely or probably live in a simulation.
• The three options are near-universal extinction before posthumanity, near-universal disinterest in ancestor-simulations, or near-certainty that we are simulated.
• The argument depends on premises such as substrate independence and an indifference principle, both of which are debated.
• Bostrom himself suggested spreading credence roughly evenly across the three options.
• The argument contains no physical evidence; it is a piece of probabilistic reasoning.

If the question leaves you wondering what to do with it, We Are in a Simulation: How to Win in It opens with a section that lays out Bostrom's three possibilities honestly, then turns to practical "rules" such as compounding, incentives and decision-making that hold whatever reality is made of. You can also browse the wider Asendify library of practical playbooks.

FAQ

Did Nick Bostrom say we are living in a simulation?

No. He argued that one of three propositions must be true and suggested that, given our ignorance, it is sensible to spread belief roughly evenly between them.

What is an ancestor-simulation?

It is Bostrom's term for a detailed computer simulation, run by an advanced civilisation, of its own past, including the conscious minds of the people who lived then.

Is there any way to test the simulation argument?

The argument itself is philosophical and makes no testable physical prediction. Some physicists have proposed experiments for related ideas, but none has produced evidence that we are simulated.

More playbooks on other topics: https://www.asendify.co/books/

Looking for a different topic? Browse all 22 Asendify books by topic.