Quantum Computing Foundations — A Briefing Note with Sources

Quantum computing fundamentals briefing — error correction, hardware architectures, computational advantage, and where the field stands — key papers, expert commentary, and lab progress from January 2023 to April 2026

academic
frontier
blogs

Synthesised 2026-04-19

Narrative

The 2023–2026 academic literature on quantum computing is defined by three concurrent storylines: the genuine breakthrough in surface-code error correction, a fierce architectural race across hardware modalities, and an honest reckoning with the gap between demonstrated and useful quantum advantage. On error correction, two Nature papers from 2024 are the defining texts: Google's Willow chip paper (Nature, Dec 2024) demonstrating the first below-threshold surface code memories with a logical error suppression factor of Λ=2.14 per unit distance increase, and IBM's Bravyi et al. paper (Nature, March 2024) introducing qLDPC 'bivariate bicycle' codes that achieve comparable thresholds at roughly 10x lower qubit overhead than the surface code. Both represent genuine milestones; neither is a solved problem. Harvard/QuEra's Bluvstein et al. (Nature, Dec 2023) delivered the first programmable 48-logical-qubit processor on neutral atoms, and follow-on work by Reichardt et al. (arXiv, Nov 2024) demonstrated fault-tolerant computation with 24 logical qubits on a 256-atom ytterbium processor. A 2026 Nature paper extended this to 448 atoms. Microsoft's February 2025 Majorana 1 announcement — claiming the world's first topological qubit using InAs-Al topoconductors — generated the most scientifically contested discussion of the period: multiple Nature news articles and Scott Aaronson's public FAQ documented that independent physicists questioned whether the measurement protocol demonstrated genuine topological protection, and an arXiv critique (Legg, 2502.19560) challenged the validation method directly. On quantum advantage, a Google Quantum AI arXiv perspective (The Grand Challenge, Nov 2025) candidly acknowledged that advantage claims from IBM (127-qubit Eagle) and D-Wave were subsequently simulated classically, and that the field's hardest unsolved challenge is identifying concrete problem instances — not just sampling tasks — where quantum computers genuinely outperform the best classical methods. A landmark panel discussion published on arXiv (Aaronson, Childs, Farhi, Harrow; June 2025) achieved rare expert consensus that hardware has progressed to near-threshold fidelity, but algorithmic discovery has stalled since the 1990s, and no architecture winner has emerged.

Across hardware modalities, the academic literature reveals a nuanced picture rather than a dominant winner. Superconducting systems (Google, IBM) lead in gate speed and manufacturing scale but face fundamental coherence-time limitations and require millikelvin cryogenics. Neutral atoms (QuEra, Pasqal, Atom Computing) have emerged as the early leaders in logical qubit demonstrations precisely because their reconfigurable all-to-all connectivity is architecturally suited to QEC codes that would require costly SWAP gates on fixed-topology chips; IEEE Spectrum (Feb 2026) reported that both Microsoft/Atom Computing and QuEra are targeting Level-2 logical qubit machines by 2026–2027. Trapped ions offer the highest single-qubit fidelity (Quantinuum achieving quantum volume over 2 million) but face gate-speed and scaling bottlenecks. Topological qubits remain pre-product. The Global Risk Institute's 2024 survey of 47 quantum experts placed only a 34% probability on cryptographically-relevant quantum computers by 2034 — up from 17% in 2022, reflecting accelerating but still deeply uncertain progress.

Sources

ID	Title	Outlet	Date	Significance
a1	Quantum error correction below the surface code threshold	Nature	2024-12	Google's landmark Willow-chip paper demonstrating the first below-threshold surface code memories on superconducting processors, with logical error rates exponentially suppressed as code distance grows — widely cited as the definitive proof-of-concept for scalable QEC.
a2	Quantum error correction below the surface code threshold (arXiv preprint)	arXiv	2024-08	Full technical arXiv version of Google's Willow QEC paper, providing detailed methodology including the Sparse Blossom real-time decoder achieving 63 µs latency at distance-5 over one million correction cycles.
a3	High-threshold and low-overhead fault-tolerant quantum memory	Nature	2024-03	IBM's Bravyi et al. introduce the 'gross code' (bivariate bicycle qLDPC family) achieving 0.7–0.8% error threshold at 10x lower physical qubit overhead than the surface code, reshaping the roadmap for fault-tolerant quantum computing.
a4	High-threshold and low-overhead fault-tolerant quantum memory (arXiv)	arXiv	2023-08	Preprint of IBM's qLDPC gross-code paper by Bravyi, Cross, Gambetta et al., which established that quantum LDPC codes can match surface-code error thresholds with far fewer physical qubits, triggering a wave of experimental follow-up.
a5	Logical quantum processor based on reconfigurable atom arrays	Nature	2023-12	Bluvstein et al. (Harvard/MIT/QuEra) demonstrate the first programmable logical quantum processor with up to 48 logical qubits on 280 physical qubits in a neutral-atom array, heralding the era of early error-corrected computation.
a6	Logical quantum processor based on reconfigurable atom arrays (arXiv)	arXiv	2023-12	Full arXiv version of the Harvard/QuEra logical processor paper demonstrating 228 logical two-qubit gates on 48 logical qubits, showing logical encoding substantially improves algorithmic performance over physical qubit fidelities.
a7	Fault-tolerant quantum computation with a neutral atom processor	arXiv	2024-11	Reichardt et al. (Microsoft/Caltech) demonstrate fault-tolerant computation on a 256-qubit neutral Ytterbium atom processor, showing 24 logical qubits with erasure conversion and better-than-physical error rates on the Bernstein-Vazirani algorithm.
a8	A fault-tolerant neutral-atom architecture for universal quantum computation	Nature	2026-01	Uses reconfigurable arrays of up to 448 neutral atoms to implement and combine key elements of a universal, fault-tolerant quantum processing architecture, including 2.14x below-threshold surface code performance with machine learning decoding.
a9	Learning high-accuracy error decoding for quantum processors (AlphaQubit)	Nature	2024-11	Google DeepMind's AlphaQubit paper presenting a transformer-based neural network decoder that outperforms state-of-the-art decoders on real Sycamore hardware data for distance-3 and distance-5 surface codes, opening neural-network decoding for real hardware.
a10	Demonstrating quantum error mitigation on logical qubits	Nature Communications	2026-02	Proposes and experimentally demonstrates zero-noise extrapolation applied to error correction circuits on superconducting processors, advancing early fault-tolerant quantum computing by combining error mitigation with error correction.
a11	Experimental demonstration of logical magic state distillation	arXiv	2024-12	QuEra/Harvard team demonstrates logical magic state distillation on a neutral-atom processor — a critical missing ingredient for universal fault-tolerant computation beyond Clifford operations.
a12	Microsoft unveils Majorana 1 — the world's first quantum processor powered by topological qubits	Microsoft Azure Quantum Blog	2025-02	Official announcement of the Majorana 1 chip, claiming the first hardware-protected topological qubit using an InAs-Al topoconductor, accompanied by a Nature paper on parity measurement and an arXiv device roadmap.
a13	Roadmap to fault tolerant quantum computation using topological qubit arrays	arXiv	2025-02	Microsoft's four-generation device roadmap (Aasen et al.) for building fault-tolerant quantum computers using Majorana-based tetron qubits, spanning from single-qubit benchmarking through lattice surgery on two logical qubits.
a14	Microsoft quantum-computing claim still lacks evidence: physicists are dubious	Nature	2025-03	Nature news article capturing the scientific community's skeptical reception of Microsoft's Majorana 1 claim, noting that attendees left a key presentation with questions unanswered about whether topological protection was truly demonstrated.
a15	Microsoft quantum computing 'breakthrough' faces fresh challenge	Nature	2025-03	Reports a physicist's (H.F. Legg, arXiv:2502.19560) challenge to the measurement protocol underpinning Microsoft's topological qubit claim, representing the most substantive independent critical analysis of the Majorana 1 announcement.
a16	Microsoft claims quantum-computing breakthrough — but some physicists are sceptical	Nature	2025-02	First Nature news response to the Majorana 1 announcement, explaining Microsoft's topological approach and documenting the initial wave of external skepticism about whether the demonstration constitutes a genuine qubit.
a17	The Grand Challenge of Quantum Applications	arXiv	2025-11	Google Quantum AI perspective paper proposing a five-stage framework from quantum advantage discovery to deployment, candidly noting that advantage claims from IBM and D-Wave were later classically simulated and that identifying concrete advantage instances remains under-resourced.
a18	Future of Quantum Computing	arXiv	2025-06	Panel summary authored by Barry Sanders with Scott Aaronson, Andrew Childs, Eddie Farhi, and Aram Harrow, providing a frank expert debate on hardware progress, algorithmic stagnation, and the field's honest disagreements about timelines and what counts as useful advantage.
a19	The vast world of quantum advantage	arXiv	2025-08	Comprehensive arXiv survey mapping the landscape of proven and contested quantum advantages across computation, learning/sensing, and cryptography, noting the relentless competition from improving classical tensor network and ML-based simulation methods.
a20	Quantum advantage for learning shallow neural networks with natural data distributions	Nature Communications	2025-12	Lewis et al. prove an exponential quantum advantage for learning periodic neurons over non-uniform distributions in the quantum statistical query model, one of the most rigorous recent demonstrations of quantum ML advantage beyond synthetic uniform cases.
a21	Quantum Deep Learning Still Needs a Quantum Leap	arXiv	2025-11	Critical quantitative assessment arguing that quantum deep learning faces three structural barriers — qubit overhead, missing QRAM infrastructure, and narrow applicability of existing speedups — supported by hardware trend forecasts through the 2020s.
a22	Quantum computing with atomic qubit arrays: confronting the cost of connectivity	arXiv	2025-11	Lecture-based review from a 2024 Varenna school (updated through summer 2025) providing a rigorous architectural analysis of neutral-atom quantum computing, focusing on connectivity costs, error correction protocols, and comparison with competing modalities.
a23	Quantum computing: foundations, algorithms, and emerging applications	Frontiers in Quantum Science and Technology	2025-12	Comprehensive 2025 review synthesizing foundational theory, hardware architectures, and application readiness, critically noting that end-to-end resource analyses are frequently incomplete and benchmarking remains at an early, inconsistent stage.
a24	FAQ on Microsoft's topological qubit thing (Scott Aaronson's Shtetl-Optimized blog)	Shtetl-Optimized (Scott Aaronson)	2025-02	Widely-read independent expert commentary by complexity theorist Scott Aaronson on the Majorana 1 announcement, providing a technically literate critical perspective that circulated widely in the research community as a reality-check on Microsoft's claims.
a25	IBM lays out clear path to fault-tolerant quantum computing (IBM Quantum roadmap 2025)	IBM Quantum	2025	IBM's updated fault-tolerant roadmap detailing the Loon (2025), Kookaburra (2026), Cockatoo (2027), and Starling (2028) processor sequence for implementing qLDPC logical processing units, providing a concrete engineering timeline for independent assessment.