A quantum computer is like a fragile snowflake resting on a warm palm. The slightest breath of heat or a gentle vibration, and it turns into a droplet. To keep the snowflake alive, you constantly adjust it, restoring its original shape. That's quantum error correction. But the warmth of the palm plays another role: it's precisely because of it that we know the snowflake existed, since the droplet retains a memory of the icy pattern. This phenomenon is known as quantum Darwinism — information about a quantum object "multiplies" in the environment and becomes accessible to a classical observer.

The authors of the new study took a specific correction scheme, Shor's nine-qubit code, and placed it in an environment of many auxiliary qubits. The result was a simple, solvable model: one logical qubit interacting with N environment qubits. The researchers honestly calculated three quantities: how well the original quantum state is restored after correction, how much information about it can be extracted from the environment, and how many redundant copies of that information appear there.

Then the authors proved a general theorem: as soon as the restoration accuracy exceeds a certain threshold, redundant classical records in the environment cannot arise — regardless of how the interaction is specifically arranged. Their model reaches this limit, showing that the boundary is not accidental but fundamental. Protecting quantum logic and the emergence of classical memory turned out to be two sides of the same coin, and you can't flip that coin to show both sides at once.