Imagine two people comparing grocery lists without revealing what's on them. Classically, they can count differences (Hamming distance) with rough accuracy — but the noise required to keep each list private scales with the list length. The only classical escape is assuming certain math problems are hard for computers, an assumption that quantum computers might eventually break. This paper asks: what if the channel itself is quantum? Alabi and Khabiboulline prove that quantum communication protocols for two-party Hamming distance achieve O(1) expected error under pure differential privacy — information-theoretically, with no computational assumptions at all. Classically, the best information-theoretic bound requires Ω(√n) error for pure DP and Ω(√n / log n) for approximate DP. The quantum protocol matches the accuracy previously available only under computational security (like learning-with-errors hardness), and it does so with O(n) communication. This is a genuine complexity separation: quantum channels are strictly more powerful than classical channels for private computation in this model. The construction centers on two novel technical ideas. The first is a "guarded coherent round trip" — the quantum message travels from Alice to Bob and back, but with structural constraints that prevent either party from cloning or extracting useful side-information during transit. The second is an "equal-Gram rigidity principle" that formalizes why an honest player cannot retain input-dependent complementary information from a non-orthogonal quantum message. Together, these enforce privacy through the physics of quantum states rather than through computational hardness. The privacy model matters here and the authors are careful about it. They work in Klauck's honest, nonpreemptive, message-preserving model — meaning players follow the protocol honestly, don't abort mid-round to extract partial information, and preserve the quantum messages as specified. The paper explicitly separates this model from weaker "prescribed-channel" privacy (which already admits exact classical realization, so no quantum advantage there) and from stronger "retention-robust" security (which remains vulnerable to measurement-and-abort attacks). The key insight is that preservation of non-orthogonal quantum messages is itself a resource for privacy. This is a precise claim about where quantum advantage lives in the privacy landscape. For approximate (ε, δ)-QDP, the authors provide an exact hockey-stick divergence calculation that yields strictly smaller error than the pure-DP case, while preserving the O(1)-versus-Ω(√n / log n) separation when δ = o(1/n). The error bound in the pure case is 2/sinh(ε) + γ for every γ > 0, which is explicit and evaluable. For practical ε values (say ε = 1), this gives error around 1.7 — genuinely constant regardless of input length n. The integrity profile is strong for a theoretical paper: the results are proved, not simulated. The lower bounds on classical protocols are established prior-art results, so the comparison is against real baselines, not strawmen. The key limitation is that the privacy model assumes honest behavior. The paper doesn't claim to solve the malicious-adversary case — it carves out the exact model where quantum channels provide a provable advantage and explicitly names the models where they don't. That's intellectual honesty you don't always see. The practical timeline here is long. This is a foundational complexity-theoretic result, not a deployable protocol. Two-party quantum communication at scale requires quantum networks that don't yet exist for general users. But the conceptual contribution — identifying non-orthogonal message preservation as a privacy resource — could reshape how the quantum cryptography community designs protocols. The next concrete test is whether this separation extends to the malicious-adversary setting or to richer function classes beyond Hamming distance.
