Are We Living in a Simulation?

The modern version of the question was posed with unusual precision in 2003, by the Oxford philosopher Nick Bostrom, in a paper titled 'Are You Living in a Computer Simulation?' Bostrom did not argue that we are in a simulation. He argued something more careful, and in its way more unsettling: that at least one of three propositions must be true. Either almost every civilization like ours is destroyed before it becomes advanced enough to run high-fidelity 'ancestor simulations'; or almost every civilization that reaches that capability chooses not to run them; or we are almost certainly living in one right now. The reasoning rests on two premises. The first is substrate independence: the idea that a mind is essentially a pattern of information-processing, so that a sufficiently detailed simulation of a brain would be just as conscious as the brain, whatever hardware it runs on. The second is scale: a technologically mature civilization could, with something like a planet-mass computer, run an astronomical number of such simulations. If both premises hold, then simulated minds would vastly outnumber the original biological ones, and a random observer, you, should statistically expect to be one of the many rather than one of the few, unless one of the first two escape routes closes the door. The crucial thing to grasp, and the thing most retellings get wrong, is what this establishes. The argument is logically valid: grant the premises, and the trilemma follows. But it is not a proof that we are simulated. It does not tell us which of the three propositions is true, only that they cannot all be false. It is a beautifully built piece of reasoning, not a piece of evidence. (Bostrom's own research centre at Oxford, the Future of Humanity Institute, was closed by the university in 2024, but an argument stands or falls on its logic, not its author's address.)

Enthusiasts point to a list of features of physics that, they say, look suspiciously like the signatures of a computed world. It is worth walking through them honestly, because each one pairs a real, established piece of physics with a suggestive but strictly analogical reading. Energy, electric charge, and spin come in discrete units, and space and time may have smallest meaningful scales (the Planck length and Planck time), which invites the image of reality as 'pixelated,' like a screen. The speed of light is an absolute limit on how fast anything can happen or any influence can travel, which invites comparison to the maximum speed a game engine will allow. A quantum system seems to settle into a definite state only when it is measured, which invites the analogy of a game that renders detail only where a player happens to be looking, to save computation. Entangled particles stay perfectly correlated across any distance, which invites the metaphor of two variables pointing at the same location in memory. The amount of information a region of space can hold seems to scale with its surface area rather than its volume (the holographic principle), which invites the picture of a world 'rendered' from data on a distant boundary. The constants of physics appear finely tuned for stars, chemistry, and life, which invites the reading that the settings were deliberately chosen, the way a designer would choose them. And the universe is described with almost eerie success by mathematics, which invites the thought that it runs on mathematics because it is, at bottom, a computation. Every one of these is real physics. But not one of them is evidence of a simulation, and the honest objection is sharp: we built our computers using our knowledge of physics, so of course our computing metaphors fit physics. The arrow of explanation runs from the world to our machines, not the other way around. The analogies are also selective, quietly passing over the many features of physics, the smooth continuity of spacetime far above the Planck scale, for one, that look nothing like any computation we know how to build. Suggestive is not the same as true.

One claim in this family sounds far stronger than the rest, and it deserves careful handling because it is the one most often oversold. The physicist S. James Gates Jr., studying the deep mathematics of supersymmetry, found that the equations, when represented as certain graphical objects he calls Adinkras, contain structures identical to the error-correcting codes used to keep data intact as it travels across a computer network, the same family of codes that lets a scratched disc still play. Error-correcting codes, buried inside the equations of fundamental physics: it is a genuinely startling, and genuinely real, mathematical discovery. But two cautions must travel with it, always. First, the codes were found inside the mathematics of supersymmetry, a proposed extension of the Standard Model that, despite decades of searching at the Large Hadron Collider, has never been experimentally confirmed; as of today the collider has found no supersymmetric particles at all. The codes are a real feature of an unconfirmed theory's equations, not a confirmed feature of the physical world. Second, even granting the structure is real, reading it as a sign that someone is running our universe is an interpretation laid on top of the mathematics, not anything the mathematics itself asserts. It is a fascinating result, and a favourite of documentaries. It is not proof of anything about simulations.

If the simulation hypothesis is to be science rather than metaphysics, it has to make a prediction someone could actually check. The most serious attempt came in 2012, when the physicists Silas Beane, Zohreh Davoudi, and Martin Savage noticed something suggestive about how physicists themselves simulate the strong nuclear force: they lay spacetime out on a discrete grid, or lattice, and that grid leaves a faint fingerprint, imposing a maximum energy and a slight directional bias on the particles within the simulation. If our own universe were running on such a lattice, they reasoned, the highest-energy cosmic rays might show a matching anisotropy, arriving very slightly more often from some directions than others. It was a genuine, testable proposal, precisely the kind of measurement observatories like the Pierre Auger array are built to make. But the test ran aground on a familiar problem, one the authors themselves acknowledged: the predicted signature depends completely on assumptions about the simulation's design. A different lattice, or a cleverer simulator, would leave a different fingerprint, or none at all. And more deeply, any anomaly we did find could always be blamed on ordinary physics we do not yet understand, rather than on a cosmic lattice. This is the hypothesis's central, and possibly permanent, difficulty. It has the same shape as Descartes' imagined all-deceiving demon, or the modern 'brain in a vat': a scenario carefully constructed so that no observation from inside could ever distinguish it from ordinary reality. A claim that no possible evidence could count against is, in the strict sense, unfalsifiable, and that is a statement about testability, not a verdict that it is false.

That unfalsifiability is exactly what draws the sharpest criticism, and it is important to be precise about what the critics do and do not say. The physicist Sabine Hossenfelder has called the simulation hypothesis 'pseudoscience,' arguing that believing it takes faith rather than evidence, and that no one has actually shown you could reproduce the known laws of physics with an algorithm without breaking the very symmetries those laws depend on. The Nobel laureate Frank Wilczek has pointed out that our universe encodes enormous hidden complexity at scales no one ever observes, a strange thing for an efficient simulation to bother computing, and that the idea invites an infinite regress, since the simulators would themselves need simulators. The cosmologist George Ellis dismissed it as 'totally impracticable,' with the line that a late-night pub discussion is not a viable scientific theory. But notice what these objections share: almost none of them claims to have proven the hypothesis false. They claim it is untestable, unmotivated, and therefore not science, which is a different and more careful charge than 'it is wrong.' And here is the twist: at least one major philosopher pushes the other way. David Chalmers argues that even if we are living in a simulation, it would not make our world an illusion. The trees and people and cities would be perfectly real; they would simply be made of information rather than fundamental particles, which, he notes, may not be so different from what physics was already telling us. On this view the question, even answered yes, might change less than it first appears.

For all its modern digital dress, the suspicion at the heart of this question is very old, and it is worth honouring that lineage without mistaking it for evidence. Plato imagined prisoners in a cave who took flickering shadows on a wall for the whole of reality, a story told in full elsewhere in these halls. Hindu philosophy speaks of Maya, the veil of appearances drawn over the single underlying reality of Brahman. The Chinese sage Zhuangzi dreamed he was a butterfly and woke unable to say whether he was a man who had dreamed he was a butterfly, or a butterfly now dreaming he was a man. Buddhist, Aboriginal Australian, and Gnostic traditions each, in their own idiom, frame the visible world as not the deepest layer of what is. And in 1977 the science-fiction writer Philip K. Dick told an audience in Metz, 'we are living in a computer-programmed reality.' These are genuine and moving echoes of a perennial human intuition, that appearances may not be the whole story, and they are worth sitting with. But they are metaphors and intuitions, not evidence: not proof that any ancient sage anticipated computer science, and not proof that a simulation exists. One thing does need to be said plainly and then set firmly aside: the popular notion that a person can 'hack' or 'glitch' the simulation through meditation, drugs, or ritual has no basis whatsoever. It mistakes a careful philosophical argument for a movie plot.