Imagine you're packing a shipping container. You know in theory that a certain volume of goods can survive a certain amount of damage in transit — that's the capacity limit. Random packing arrangements can achieve this limit, but good luck telling a warehouse worker how to actually pack that way or how to unpack it quickly when it arrives dented. For decades in classical coding theory, mathematicians knew random codes could hit the list decoding capacity limit, but nobody could write down an explicit, structured code that did it until Guruswami and Rudra cracked it in 2006. This paper does the quantum version — and adds the crucial constraint that the code must be LDPC (low-density parity-check), meaning the structure is sparse enough to be practical. The committed claim: this is the first explicit construction of quantum LDPC codes that achieve list decoding capacity — meaning the decoding radius approaches the quantum Singleton bound with constant list sizes — together with near-linear time list decoding algorithms. That's three properties simultaneously: explicitness, LDPC structure, and capacity-achieving list decoding. Any two of these three existed before; all three together is new. The construction leans on expander graphs via quantum AEL (Alon-Edmonds-Luby) distance amplification, a technique recently ported to the quantum setting by Bergamaschi, Golowich, and Gunn at STOC 2024. The classical AEL trick takes a code with modest distance and amplifies it using expander graphs — think of it as a signal booster for error correction capability. The quantum version is trickier because quantum codes must simultaneously correct X-type and Z-type errors (the CSS structure), and the amplification must preserve both. The LDPC property falls out of the expander structure: expander graphs are inherently sparse, so the resulting parity-check matrices have bounded row and column weights. The decoding side generalizes recent classical results on expander-based weak-regularity list decoding by Srivastava-Tulsiani (FOCS 2025) and Jeronimo-Singh (2025). The key insight is that the expander structure provides enough regularity to run a combinatorial peeling/pruning algorithm that enumerates all codewords within the list decoding radius, and it does so in time near-linear in the block length. Near-linear time matters enormously: quantum error correction runs under brutal latency constraints, and a decoding algorithm that scales quadratically or worse in block length is a non-starter for large codes. The ladder here is nuanced. The quantum Singleton bound is the theoretical ceiling — it's the quantum analogue of the classical Singleton bound that governs how much redundancy you need for a given error rate. Prior quantum list decoding results either used random codes (non-explicit), lacked the LDPC property, or didn't approach capacity. Panteleev and Kalachev's 2022 breakthrough gave good quantum LDPC codes (constant rate, constant relative distance) but didn't address list decoding capacity. This paper closes the gap by combining LDPC structure with capacity-achieving list decoding. No prior explicit construction had all three properties. Integrity-wise, this is a mathematical construction paper — the validation is proof, not experiment. The claims are about existence and algorithmic complexity, not empirical performance on hardware. The relevant question is whether the proofs are correct and the parameters are tight, not whether the codes run on a quantum computer today. The paper is accepted at SODA 2027, which means it has passed peer review in a top-tier theory venue. The building blocks (quantum AEL from STOC 2024, classical list decoding from FOCS 2025) are recent but from strong groups. The milestone that matters is implementation at scale. Theoretical capacity-achieving codes need to translate into codes with practical parameters — block lengths, qubit overhead, and syndrome measurement circuits that work on real hardware. Classical LDPC codes took roughly 15 years to go from theoretical capacity-achieving constructions to the codes running in 5G base stations today. The quantum version faces additional hardware constraints (coherence times, gate fidelities). The next concrete step is explicit parameter regimes: for a target logical error rate of 10⁻¹² at code distance d, what block length and qubit overhead does this construction require, and how does it compare to surface codes and other candidates? The obvious experiment not run: no concrete parameter instantiation or numerical comparison against surface codes or other quantum LDPC families (like Panteleev-Kalachev or Dinur et al. c³ codes) at specific block lengths. This is a theory paper, so this omission is standard rather than suspicious — the authors are establishing the existence result, not engineering a deployment. But the field will need those numbers before this construction competes with surface codes in practice. Most likely explanation: (c) they're saving it for the next paper or leaving it for the systems community.
