A theoretical floor on how much reliable quantum information can survive a noisy channel just got raised. When both ends of a quantum line can talk back and forth, the proven minimum rate at which entanglement can be reliably shared across the qubit depolarising channel (a standard model of noise on a two-state quantum system) has notched upward across a wide range of channel conditions. The shift is theoretical, not engineering: a lower bound is the smallest rate a clever protocol is known to provably achieve, and the question of what the true maximum is remains open.
That distinction matters because the same number is the rate-loss constant that any future quantum repeater, the building block of a long-distance quantum network, will have to clear. The practical applications, long-distance quantum communication and distributed quantum computing, depend on this kind of bound, but the dependency is years-out, not a deployment signal. The method behind the new bound is a heuristic search over entanglement-distillation protocols, a trial-and-error scan for cleaner ways to pull usable entanglement out of noisy inputs, not a machine-learned model. Barber and Pirandola describe the result as progress on a long-standing open problem, which is the right frame: the floor moved, the ceiling is still unmapped.