In 1982, physicist Richard Feynman proposed a simple idea: to understand a quantum system, you need to build another quantum system that behaves the same way. Thus was born the quantum simulator, a device that does not calculate equations but reproduces them in its own behavior. Just as a mechanical planetarium turns gears to show the motion of planets, a simulator manipulates its atoms to reveal what happens inside a superconductor.
But there is a catch. You can take apart a planetarium and check every gear. A simulator built from quantum particles, however, obeys the same strange rules it aims to explain. How do you know it is not lying? The authors of the article answer: simulation is not just a demonstration. It is a bridge between model and reality, and that bridge must be stress-tested.
To bring order to the concepts, the authors introduce a simple criterion. A device is considered classical if its state can be described without quantum "entanglements" between particles, like a list of probabilities rather than a unified whole. By this criterion, even a Bose-Einstein condensate, where thousands of atoms merge into a single wave, can be a classical simulator if it operates in the condensate regime. But use its excitations, and it becomes quantum. It all depends on which part of its behavior the device lends to the model.
Trust in a simulator is not an objective property but the result of a falsification process: tests, comparisons, and gradual accumulation of confidence. Just as you trust a bridge after thousands of trucks have driven over it, you trust a simulator after its predictions have matched reality in thousands of cases.
In the future, when simulators begin to outperform classical computers in tasks where exact solutions are unavailable, the problem of verification will become central. Efforts will shift from demonstrating control to developing methods for establishing trust: finding regimes where results are sensitive to errors, and understanding which quantities can be reliably simulated at all. The bridge between model and reality will have to be stress-tested anew each time.
The term "quantum simulator" is often attributed to Richard Feynman, who in 1982 suggested that quantum systems are best modeled by quantum devices. However, the idea of using one physical system to study another dates back to ancient mechanical models of celestial bodies.