Curator's Take
This article demonstrates a practical route to bypass the costly magic‑state distillation bottleneck by using a cavity‑mediated, probabilistic injection of $T$‑gates with a 74 % success rate per attempt—far higher than many earlier proposals that relied on multiple rounds of purification. By leveraging readily available Rydberg atom–cavity interactions and Clifford‑only teleportation steps, the scheme could shrink both qubit overhead and circuit depth on near‑term hardware while still being extensible to logical, error‑corrected codes. If experimental groups can realize the required dressed‑state transitions, the approach may become a key building block for scalable universal quantum processors that need non‑Clifford resources without prohibitive resource costs.
— Mark Eatherly
Summary
Non-Clifford gates are a necessary resource for universal quantum computation, yet their fault-tolerant implementation typically relies on magic-state distillation, which incurs significant overhead in qubit count and circuit depth. In this work, we propose a probabilistic cavity-based magic-state injection protocol. Our scheme exploits controlled atom-cavity interactions and conditional measurements to probabilistically prepare an effective magic state encoded in the first two level Fock subspace of a single cavity mode, achieving a success probability of $0.74$ per attempt, independent of the target magic phase. The cavity-encoded magic state is subsequently injected into a computational atom via a teleportation-based protocol mediated by dressed-state transitions, requiring only Clifford operations and a single auxiliary atom for readout. We show that all required operations -- state preparation, two-qubit exchange gates, and projective measurement -- can be implemented with experimentally available techniques in Rydberg atom-cavity platforms. We further discuss how the scheme can in principle be adapted to operate at the logical level, where collective Rydberg interactions and optical nonlinearities provide a route toward cavity-mediated $T$-gate injection directly into code-encoded qubits.