What Is the Simulation Hypothesis?

Background

The idea that perceived reality might not be what it appears predates computing by centuries. René Descartes' 17th-century thought experiment asked how anyone could rule out an "evil demon" feeding false sensory experiences directly into their mind, and philosopher Hilary Putnam's 1981 "brain in a vat" scenario updated the same problem for a scientific age: a brain kept alive in a vat, wired to a machine simulating an entire life's worth of sensation, would have no internal way to tell its experience apart from an unsimulated one.

The modern, computational version of the argument was formalised by Swedish-born philosopher Nick Bostrom, then at the University of Oxford, in his 2003 paper "Are You Living in a Computer Simulation?", published in the Philosophical Quarterly. Bostrom did not argue that we are definitely simulated. Instead, he constructed a trilemma: given some modest assumptions about future computing power and about how a sufficiently advanced civilisation might use it to run detailed simulations of its own ancestors, at least one of three propositions must be true. Either almost every civilisation at our technological stage goes extinct before reaching the capability to run such simulations, or almost every civilisation that reaches that capability chooses not to run them, or the number of simulated minds vastly exceeds the number of non-simulated ones, in which case a randomly selected observer, including the reader, is almost certainly one of the simulated ones. Bostrom takes no explicit position on which branch is true, and has said in later interviews that he considers the odds across the three roughly comparable.

The argument found a much wider public audience than most philosophy papers achieve, discussed by physicists, technologists, and podcasters over the two decades since publication, in part because it connects an old philosophical puzzle to concrete, tractable ideas from computer science, most notably the exponential growth in computing power that made "ancestor simulations" feel like a plausible future capability rather than pure fantasy.

Main Theories

Bostrom's simulation argument

The case for taking the hypothesis seriously rests on the trilemma's internal logic rather than on any direct evidence. If advanced civilisations that survive long enough eventually gain the ability to run high-fidelity simulations of their own evolutionary history, for research, entertainment, or ancestor-simulation purposes, and if even a modest fraction of them choose to run many such simulations, the population of simulated conscious observers across all of reality would dwarf the population of original, non-simulated observers. Applying a basic principle of reasoning under uncertainty, that a randomly selected observer is more likely to be a member of the larger group, an observer with no special reason to think otherwise should assign meaningful probability to being one of the simulated majority rather than a member of the much smaller original population.

The argument's force depends entirely on its assumptions holding: that consciousness or subjective experience can, even in principle, be substrate-independent (arising from computation regardless of what physical material runs it), that some future civilisations will have both the capability and the interest to run vast numbers of such simulations, and that a randomly selected observer's reasoning about their own likely status is a valid form of inference. Bostrom's own contribution was showing that these assumptions, if granted, lead to a genuine trilemma rather than a refutable claim, which is a different and more modest achievement than proving simulation is likely.

The skeptical critique

The most common objection targets the argument's testability rather than its internal logic: even if the trilemma is valid reasoning, it produces no prediction that could distinguish a simulated universe from a non-simulated one, since a sufficiently capable simulation could, by construction, hide any trace of itself from observers inside it. Philosophers of science generally regard this as placing the hypothesis outside the normal boundary of a testable scientific claim, closer to a metaphysical proposition than a physical theory, however interesting the reasoning behind it.

This is precisely the concern philosophers of science raise when applying the demarcation problem more broadly: Karl Popper's falsifiability criterion, the most widely used practical test for what counts as a scientific claim, is one the simulation hypothesis fails by construction, which is why it is generally classified as philosophy rather than physics, whatever its scientific-sounding vocabulary.

A second line of criticism is more directly physical: some physicists argue that simulating a universe at the full fidelity required to reproduce quantum-mechanical behaviour across cosmological volumes would demand computational resources that may be physically impossible to assemble even for an arbitrarily advanced civilisation, since certain quantum many-body calculations appear to scale in complexity faster than any conceivable computer could keep pace with. This objection does not disprove the hypothesis outright, since a simulation might not need to render full quantum fidelity everywhere at once, but it weakens the assumption that "sufficiently advanced computing power" is simply a matter of waiting long enough.

A third, more philosophical critique questions the reasoning itself: the argument's step from "simulated observers would greatly outnumber real ones" to "you should therefore think you're probably simulated" relies on a specific and contested principle about how to reason under this kind of uncertainty, and philosophers who reject that principle, or who apply it differently, do not reach the same trilemma at all.

Common Misconceptions

The most widespread error is the one furthest from the source: that Bostrom argued we are probably living in a simulation. His paper argues that at least one of three propositions must hold and takes no position on which, and he has since said he regards the odds across the branches as roughly comparable. The popular compression drops two branches and keeps the most striking one, which turns a conditional argument into a claim it never made.

A related misconception treats the hypothesis as a scientific theory because it is built from scientific materials. Computing power, quantum mechanics, and probability all appear in the argument, but no observation is specified that a simulated universe would produce and a real one would not, which is what a scientific theory has to supply. Borrowing physics vocabulary is not the same as making a physical prediction.

Finally, the hypothesis is often presented as an inevitable consequence of improving computers: if processing power keeps growing, ancestor simulations must eventually exist. That does not follow. The argument requires substrate-independent consciousness, a civilisation that survives long enough to build such simulations, and a decision to run many of them — three assumptions no trend in hardware establishes. The physical objection cuts the other way too: some physicists argue the quantum many-body calculations involved may scale beyond what any computer could perform, however long anyone waits.

Current Consensus

There is no scientific consensus on the simulation hypothesis, because it is not, by the standard definition, a scientific claim capable of being confirmed or refuted by evidence. Among philosophers, Bostrom's trilemma is widely regarded as a genuinely interesting and logically serious argument, one that has generated substantial academic discussion, without most specialists concluding that any of its three branches is actually true. Among physicists, engagement is more mixed: some treat proposed indirect tests, such as searching for signatures resembling a computational grid in high-energy cosmic-ray data, as worth attempting, though none has found supporting evidence, while others regard the entire enterprise as outside physics' proper domain, since it makes no falsifiable prediction.

What is generally agreed, across both fields, is that the hypothesis cannot currently be confirmed, cannot currently be ruled out, and does not follow automatically from ordinary advances in computing power, contrary to how it is sometimes presented in popular discussion.

Why the Question Endures

The simulation hypothesis endures because it reframes one of philosophy's oldest problems, how can you be certain your experience of reality is real, in the vocabulary of an era that trusts computation as its most powerful explanatory tool. Descartes needed a deceiving demon; a culture that grew up alongside video games and virtual reality reaches instead for a simulation, a substitution that makes the same old doubt feel newly plausible rather than merely philosophical.

It also endures because, unlike a pure thought experiment, Bostrom's version is built from a testable-feeling scaffolding: real trends in computing power, a real argument structure, real academic engagement, even while the conclusion itself resists any test. That combination, rigorous-looking reasoning paired with permanent unfalsifiability, gives the idea unusual staying power in public discussion, where it functions less as a settled or even testable claim and more as a standing invitation to notice how much of what feels certain about reality is actually assumed rather than verified. The same shape of unfalsifiable-but-compelling reasoning is part of why people find conspiracy theories persuasive more generally, even though the simulation hypothesis itself is a mainstream philosophical argument rather than a conspiracy claim.

It sits alongside physical science's other great open questions, the Fermi paradox, dark matter, and dark energy, as a case where confident-sounding reasoning coexists with a complete absence of confirming detail. What separates it from those three is falsifiability in principle: a dark matter particle could, at least in theory, be directly detected tomorrow, a genuine biosignature could resolve part of the Fermi paradox, and dark energy's evolving-versus-constant nature is already being tested against new survey data, but no equivalent observation could ever confirm or rule out that reality itself is simulated, which is exactly the property that keeps philosophers, rather than experimentalists, at the centre of the debate. The simulation hypothesis is covered alongside the rest of this site's scientific theories and frontiers.