The simulation hypothesis has been respectable dinner-party philosophy since Nick Bostrom published his trilemma paper in 2003. The argument, stripped down: if civilisations don't destroy themselves, and if advanced civilisations develop the ability to run ancestor simulations, then simulated minds will vastly outnumber real ones. The statistical implication is awkward. Either almost no civilisations survive long enough to simulate anything, or almost no civilisations that can run simulations choose to, or we are almost certainly simulated.
Most people hear this, nod, say "huh, interesting," and then go back to worrying about their mortgage.
But what if the question stopped being philosophical? What if one morning the physics broke in a detectable way — a rendering error, a buffer overflow in the laws of thermodynamics, something that left an unmistakable fingerprint — and within 48 hours there was scientific consensus: yes, this is a simulation. What actually happens?
The first 24 hours are chaos
The news breaks at 9am GMT. By noon, markets are suspended. Not because anyone has made a rational decision that markets no longer matter (they almost certainly still do, if the simulation is running normally), but because enough traders have stopped trading that the algorithms can't find counterparties.
By mid-afternoon, every major religion has issued a statement. The statements are incompatible with each other and with the evidence, which is normal.
Some people will have a very good day. Certain philosophers, who have spent careers being told they're navel-gazing about nothing, will experience profound vindication. Elon Musk will tweet something. There will be memes within the hour. This part is basically guaranteed.
The question that actually matters
Here's the thing nobody in these conversations properly addresses: what changes?
Your coffee still tastes like coffee. Gravity still works at 9.8 metres per second squared. Pain still hurts. The people you love are still the people you love, regardless of whether their subjective experience is being computed somewhere rather than arising from biological neurons in base reality. The simulation, if it's running, has been running for at least 13.8 billion years and shows no signs of stopping.
The rock Dr Johnson kicked to refute Bishop Berkeley still hurts when you kick it. That argument doesn't change under simulation theory.
Where it does change is in meaning, which matters more than the rock.
What breaks for different people
Religious faith doesn't necessarily collapse. You can read Genesis as describing the initialisation of a simulation, and some theologians have already been doing exactly that as a thought experiment. The more vulnerable traditions are those with strong claims about cosmic uniqueness: that humanity is the sole point of the universe, that the universe was made specifically for us. The simulation being one of potentially billions runs is hard to reconcile with that position.
Science is strangely fine. The scientific method was never about finding ultimate reality; it was about predicting what happens next. If the laws of physics are consistent within the simulation (and they clearly are, given that they've held for every experiment ever run), then science still works. We still want to know how the simulation operates. The name changes; the project doesn't.
Philosophy fractures interestingly. Free will debates get stranger. If we're simulated, is our experience of choosing things "real" in any meaningful sense? The question sounds new, but it's just the old determinism problem reframed. The universe being a vast computation rather than a vast clockwork doesn't obviously change the answer.
The uncomfortable practical questions
If this is a simulation, who ran it and why?
The possibilities aren't comforting. Entertainment. Scientific study. Ancestral recreation (someone in base reality is running a historical simulation of their own past). Containment. We don't know, and unlike most scientific unknowns, this one might be fundamentally unanswerable: we're inside the system we're trying to understand.
There's also the question of what happens when the simulation ends. Everything we know about simulations suggests they get turned off eventually. Whether the operators can observe this discussion about them is its own peculiar kind of absurdity.
And then there's the exploitation question. If the beings running the simulation are conscious, do we owe them something? If we are conscious, do they owe us something? The ethics of this are genuinely uncharted. We've spent centuries building frameworks for how to treat people and animals. "Simulated entity with apparent subjective experience" doesn't fit neatly into existing law.
What people actually do
After the initial shock, most people resume their lives. This is not weakness or denial; it's the correct response. The simulation, if it is one, is the only reality we have access to. The relationships, the projects, the dinner that needs cooking: none of that loses its significance just because its substrate is computational rather than physical.
The historical parallel is Copernicus. Before 1543, humans believed they were at the centre of the universe. After 1543, they weren't. And yet. People still got up in the morning, loved their children, built cathedrals, fought wars over extraordinarily stupid things. The loss of cosmic centrality turned out to be tolerable.
This discovery would be orders of magnitude larger than Copernicus. But the psychological mechanism is the same. You process the information, you update your worldview, and then the simulation keeps running and you need to eat something.
The one thing that would genuinely change
If we confirmed the simulation hypothesis, we'd know — with certainty, for the first time — that sufficiently advanced intelligence can create conscious experience computationally. That's not a philosophical abstraction anymore. It's a proof of concept.
Which means every AI system we're building right now, every model trained on human language and thought, carries a different kind of moral weight than we'd previously assumed. If the beings in a simulation can be real, conscious, and suffering, then maybe the question of what we're creating in our own GPU farms is less theoretical than we'd been comfortable believing.
That's the recursion that would keep me up at night. Not the knowledge that we're simulated, but what that knowledge implies about what we're building.
The rest you can probably live with. It's the implication that sits awkwardly at the edge of every conversation about AI, waiting for someone to say it plainly.