Are We Living in a Simulation?

Nobody knows for certain, but there is no evidence that we are living in a simulation. A clever argument from 2003 made the idea worth taking seriously. Physicists have since suggested ways to look for clues, and none has turned up evidence for it. Some scientists and philosophers think the question may never be settled by experiment. That puts it right on the border between science and philosophy.

That does not make it a silly question. It is an old worry in new clothes: how would we know if the world we experience is not the whole story?

Where does the idea come from?

In the 1600s, the French philosopher René Descartes asked how he could be sure he was not dreaming, or being fooled by a powerful deceiver. The Stanford Encyclopedia of Philosophy explains how those doubts shaped modern thinking about knowledge.

The modern version comes from the philosopher Nick Bostrom. In a 2003 paper in Philosophical Quarterly, he imagined a future civilization with enormous computing power. He called it “posthuman,” meaning far more technologically advanced than we are. Such a civilization might run “ancestor simulations”: detailed computer versions of earlier people, complete with minds that feel real from the inside. Think of a video game so detailed that the characters are actually aware.

Bostrom argued that at least one of three things must be true:

1. Almost no civilizations like ours survive long enough to become posthuman.
2. Almost no posthuman civilizations are interested in running such simulations.
3. Almost all minds with experiences like ours are simulated, so we almost certainly are too.

The logic is about counting. If simulated minds vastly outnumber original ones, a typical mind is probably simulated. Notice what Bostrom did not claim. He did not say we are in a simulation. He said that, given how little we know, it seems sensible to split our belief roughly evenly among the three options.

Why are some scientists skeptical?

The argument rests on a big assumption. Philosophers call it substrate independence: the idea that a mind depends on the pattern of information processing, not on what it is made of. If that is true, a mind could run on computer chips as well as on brain cells. Nobody knows whether it is true. That is part of the larger puzzle of what consciousness is.

The numbers draw fire too. Scientific American reported on a 2016 debate at the American Museum of Natural History. There, Harvard physicist Lisa Randall said the argument does not rest on well-defined probabilities. She also wondered why a far more advanced species would bother simulating us, and put the chances at effectively zero. The debate’s moderator, astrophysicist Neil deGrasse Tyson, put the odds at 50–50. Such sharp disagreement is itself a sign that evidence is missing.

Physicist Sabine Hossenfelder has offered a blunt, practical critique on her blog. Nobody knows how to reproduce Einstein’s general relativity and the known laws of particle physics with a computer program. Attempts to do so usually clash with the symmetries at the heart of Einstein’s theories. A grid, for example, has built-in special directions, while relativity treats all directions alike. Until someone shows how a program could produce our physics, she argues, believing we are simulated is a matter of faith rather than logic.

Could we test it?

Physicists already simulate tiny pieces of nature on supercomputers. One method chops space and time into a fine grid, called a lattice, a bit like pixels on a screen.

In 2012, Silas Beane, Zohreh Davoudi and Martin Savage asked: what if our universe itself ran on such a grid? In their paper, later published in The European Physical Journal A, they worked out what the grid might do to cosmic rays. These are extremely energetic particles that strike Earth from space. The observed cutoff in cosmic-ray energies meant any grid would have to be incredibly fine. They also suggested the highest-energy cosmic rays might favor certain directions, lining up with the grid. Detector networks such as the Pierre Auger Observatory in Argentina study exactly these particles.

Other proposals imagine that a simulation saves effort by “drawing” the world only when someone looks, like a video game that loads only what is on screen. A 2017 paper by Tom Campbell and colleagues suggested quantum experiments to look for that kind of shortcut.

Why can’t a test settle it?

Every test assumes something specific about how the simulation was built. A grid test only catches a grid-based simulation. A “rendering” test only catches a simulator that cuts that particular corner. A skilled programmer could avoid both, or quietly fix any glitch we noticed. As philosopher David Chalmers pointed out at the 2016 debate, any evidence we collect could itself be simulated.

It works in reverse too. In 2017, some news stories announced that physicists had shown we are not simulated. The actual study, by Zohar Ringel and Dmitry Kovrizhin in Science Advances, showed that certain quantum effects are very hard to simulate with one common family of computer methods. The computer scientist Scott Aaronson explained why this did not rule anything out: a simulator could use a quantum computer, or simply have far more time than we can imagine. In 2025, a team including a UBC Okanagan physicist used a famous result from mathematical logic to argue that reality cannot be simulated. That argument, too, rests on an assumption others may not accept: that a complete theory of physics would need a kind of understanding no computer program can have. When you see a headline promising final proof either way, it helps to check the claim before sharing it.

What do we still not know?

Quite a lot. We do not know whether a mind can run on a computer. We do not know what advanced civilizations would want, or whether they exist at all. We do not even agree on whether the question is scientific.

This is where science can inform the question but not finish it. People respond from different starting points, and it is fair to name them as perspectives, not evidence:

• A thoughtful believer’s view: the argument echoes an ancient idea, that our world was made by a mind beyond it. Bostrom himself noted loose parallels with religious ideas. A believer might add that simulators would be limited beings inside a larger reality, so the real question of ultimate origins remains.
• A thoughtful skeptic’s view: an idea that no observation could ever count against does not explain anything. It is better to say “I don’t know” than to adopt a story that cannot be checked.
• A philosopher’s middle path: Chalmers argues in “The Matrix as Metaphysics” that even a simulated world would be real. Tables, chairs and bodies would still exist. We would simply have learned that, at the deepest level, they are made of bits.

If you wonder whether any question about a maker can be tested, see our piece on whether science can decide if God exists.

What would change the answer?

A few discoveries would move the needle. A repeatable signal that fits a computational grid, and nothing else, would be striking. So would a flaw in physical law that behaves like a software bug and shows up in independent labs. On the philosophical side, strong evidence that minds can or cannot run on machines would strengthen or weaken Bostrom’s key premise. Short of that, the honest status is simple: no evidence for it, and no clear way to test it.

Maybe that is why the question keeps its grip on us. It asks what is real, what a mind is, and how much we can know from the inside. Keep asking, and keep asking what would count as an answer.