Research · Summary
Back to sweepResearch sweep · deep · 2023 – present
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
Quantum Computing Fundamentals: Error Correction, Hardware Architectures, and the Path to Advantage
Overview
Quantum computing crossed a genuine inflection point between 2023 and April 2026. The defining shift was not an increase in raw qubit counts but the first experimental demonstrations that error correction can work as theory predicts: adding more physical qubits can suppress logical error rates rather than compound them. Google's Willow chip achieved this "below-threshold" milestone in December 2024, demonstrating a surface code memory where logical errors fell exponentially with increasing code distance. IBM published a parallel breakthrough showing that quantum low-density parity-check codes can achieve comparable error thresholds at roughly one-tenth the qubit overhead. Neutral-atom platforms, largely dismissed as laboratory curiosities three years earlier, delivered the field's first programmable logical processors with tens of logical qubits. These results collectively moved the conversation from whether fault-tolerant quantum computing is physically possible to how quickly it can be engineered at useful scale.
Sources: Nature (2024) (↗); Nature (2024) (↗); Nature (2023) (↗)
The period also forced an honest reckoning with hype. Multiple quantum advantage claims from 2019 through 2023 were subsequently matched or exceeded by classical algorithms, a pattern that Google's own researchers acknowledged in a candid November 2025 perspective paper. Microsoft's February 2025 announcement of the world's first topological qubit generated the most scientifically contested moment of the period: independent physicists at Cornell, NYU, and Caltech challenged whether the data demonstrated genuine topological protection, and Nature's editorial coverage noted that the peer-review file explicitly stated the paper contained no evidence for Majorana zero modes. The Majorana controversy crystallized a broader tension: frontier labs face intense pressure to announce breakthroughs, while the physics community demands reproducible evidence that often takes years to establish.
Sources: arXiv (2025) (↗); Nature (2025) (↗); Shtetl-Optimized (Scott Aaronson) (2025) (↗)
The current state is best characterized as early fault-tolerant: hardware fidelities have reached or approached the thresholds required for error correction to function, but the physical-to-logical qubit ratios demonstrated in experiments remain orders of magnitude below what would be needed to run cryptographically or commercially relevant algorithms. The Global Risk Institute's 2024 survey of 47 quantum experts placed only a 34% probability on cryptographically relevant quantum computers existing by 2034, up from 17% in 2022. This reflects genuine acceleration, but also deep uncertainty about the engineering path from dozens of logical qubits to thousands. The field's internal debate, captured in a landmark 2025 panel paper by Aaronson, Childs, Farhi, and Harrow, achieved rare consensus that hardware has progressed dramatically while algorithmic discovery has stalled since the 1990s.
Sources: arXiv (2025) (↗); Shtetl-Optimized (Scott Aaronson's personal blog) (2024) (↗)
Key Findings
1. The below-threshold milestone is real but narrow. Google's Willow chip demonstrated the first surface code memory operating below the fault-tolerance threshold, with a logical error suppression factor of Λ=2.14 per unit increase in code distance. This means each step up in code distance roughly halves the logical error rate, the fundamental requirement for scalable error correction. However, this was demonstrated for memory operations, not the full gate set required for universal computation. The result establishes that the physics works; the engineering challenge of maintaining this performance under active computation remains open.
Sources: Google Quantum AI Blog (2024) (↗); Nature (2024) (↗)
2. qLDPC codes offer a credible path to 10x overhead reduction. IBM's March 2024 Nature paper introduced the bivariate bicycle code, a quantum LDPC construction that encodes 12 logical qubits into 288 physical qubits while achieving comparable error thresholds to surface codes. This represents roughly an order-of-magnitude improvement in encoding efficiency. The catch is architectural: qLDPC codes require long-range qubit connectivity that current superconducting chip layouts do not naturally support. IBM's roadmap addresses this with purpose-built Loon and Kookaburra processors designed to enable the required connectivity patterns.
Sources: Nature (via PMC) (2024) (↗); IBM Quantum Blog (2025) (↗)
3. Neutral atoms emerged as the leading logical-qubit platform. Harvard and QuEra's December 2023 Nature paper demonstrated a 48-logical-qubit processor on reconfigurable atom arrays, followed by fault-tolerant computation with 24 logical qubits on a 256-atom ytterbium processor in late 2024. A January 2026 Nature paper extended this to a fault-tolerant universal architecture. The key architectural advantage is that neutral atoms can be physically rearranged to implement arbitrary connectivity, making them naturally suited to error correction codes that would require costly SWAP operations on fixed-topology chips. IEEE Spectrum reported that both Microsoft/Atom Computing and QuEra are targeting Level-2 logical qubit machines by 2026-2027.
Sources: Nature (2023) (↗); arXiv (2024) (↗); Nature (2026) (↗)
4. Topological qubits remain unproven. Microsoft's February 2025 Majorana 1 announcement claimed the world's first topological qubit using indium arsenide-aluminum topoconductors. The claim was immediately contested. Nature's news coverage noted that the accompanying paper's peer-review file stated it contained no evidence for Majorana zero modes. An arXiv critique challenged the validation methodology directly. At the APS March Meeting 2025, physicists cited Microsoft's 2018 retraction of an earlier Majorana paper as grounds for requiring higher evidentiary standards. Scott Aaronson's detailed FAQ characterized the situation as worth careful watching, noting that the press release spoke differently from the paper.
Sources: Nature (2025) (↗); APS Physics (2025) (↗); Shtetl-Optimized (Scott Aaronson's personal blog) (2025) (↗)
5. Neural network decoders improved error correction by 6-30%. Google DeepMind's AlphaQubit, published in Nature in November 2024, used a transformer-based architecture to decode surface code errors more accurately than prior methods. On Google's Sycamore and Willow processors, AlphaQubit reduced logical error rates by 6-30% compared to previous decoders. However, the system was too computationally slow for real-time decoding on superconducting hardware as of publication. The result demonstrates that machine learning can contribute to error correction but also highlights the tight timing constraints superconducting systems impose.
Sources: Nature (2024) (↗); Google DeepMind Blog (2024) (↗)
6. Quantum advantage claims remain contested. Google's October 2025 Quantum Echoes algorithm demonstrated a computation running 13,000 times faster than the Frontier supercomputer. However, Google's own November 2025 perspective paper acknowledged that prior advantage claims, including IBM's 127-qubit Eagle results and various D-Wave demonstrations, were subsequently simulated classically. The honest assessment from leading researchers is that quantum advantage has been demonstrated for artificial sampling tasks but not yet for any computation with practical commercial or scientific value. The field's hardest unsolved challenge is identifying concrete problem instances where quantum computers genuinely outperform the best classical methods.
Sources: Google Quantum AI Blog (2025) (↗); arXiv (2025) (↗)
7. Algorithm discovery has stalled since the 1990s. The 2025 panel paper by Aaronson, Childs, Farhi, and Harrow documented a striking asymmetry: hardware has improved dramatically over the past five years, but no new algorithm since Shor (1994) and Grover (1996) has been proven to offer asymptotic advantage for practically important problems. Variational quantum eigensolvers and other NISQ-era approaches have not demonstrated scalable advantage. The panel concluded that no architecture winner has emerged and that algorithmic discovery is now the binding constraint.
Sources: arXiv (2025) (↗)
8. Commercial projections remain aggressive relative to demonstrated capability. McKinsey's 2025 Quantum Technology Monitor projected $72 billion in quantum computing revenue by 2035 and noted that startup investment in 2024 was 50% higher than 2023. These projections assume fault-tolerant machines at useful scale within the decade. Independent commentators, including O'Reilly Radar and the More Is Different Substack, have issued explicit warnings against Moore's-law extrapolations of qubit quality metrics, noting that error correction overhead and architectural constraints make linear projections unreliable.
Sources: McKinsey & Company (Quantum Technology Monitor) (2025) (↗); More Is Different (Substack) (2025) (↗); O'Reilly Radar (2024) (↗)
Evidence & Data
The most important quantitative results from the 2023-2026 period establish the current state of error correction and qubit scaling.
Google's Willow chip operated a distance-7 surface code with 105 physical qubits, achieving a logical error suppression factor Λ=2.14 per unit distance increase. At distance 7, this implies roughly 100 physical qubits encoding a single logical qubit with error rates below the threshold required for fault tolerance. Extrapolating from this, a distance-17 code (the minimum estimated for cryptographically relevant computations) would require several thousand physical qubits per logical qubit using surface codes alone.
Sources: Nature (2024) (↗)
IBM's bivariate bicycle code demonstrated a [[144,12,12]] encoding: 144 physical qubits encoding 12 logical qubits at distance 12. The gross code variant used 288 total physical qubits, achieving roughly 24 physical qubits per logical qubit, a substantial improvement over surface code overhead. IBM's roadmap targets 200 logical qubits executing 100 million operations by 2029 with the Quantum Starling system.
Sources: Nature (via PMC) (2024) (↗); IBM Quantum Blog (2025) (↗)
Neutral-atom systems demonstrated dramatically larger qubit arrays. A September 2025 Nature paper reported continuous operation of a coherent 3,000-qubit system. A companion paper demonstrated a tweezer array with 6,100 highly coherent atomic qubits with coherence times reaching 12.6 seconds. QuEra's Algorithmic Fault Tolerance approach claimed 10-100x overhead reduction compared to standard surface codes, though this remains to be validated at scale.
Sources: Nature (2025) (↗); Nature (2025) (↗); Medium (nehalmr) (2026) (↗)
Quantinuum's trapped-ion system achieved a quantum volume exceeding 2 million, the highest single-system fidelity metric reported. Trapped ions offer the highest individual gate fidelities but face scaling and gate-speed constraints that limit their path to large logical qubit counts.
The Global Risk Institute's 2024 expert survey provides the most credible probability estimate for cryptographic relevance: 34% chance of a quantum computer capable of breaking RSA-2048 by 2034, up from 17% in their 2022 survey. This doubling reflects genuine acceleration but also implies a majority expert view that such capability remains unlikely within the decade.
Sources: Shtetl-Optimized (Scott Aaronson's personal blog) (2024) (↗)
Riverlane's analysis documented a jump from 36 to 120 peer-reviewed quantum error correction papers in a single year (2024-2025), reflecting the field's intensified focus on fault tolerance.
Sources: Riverlane Blog (2026) (↗)
Signals & Tensions
The surface code vs. qLDPC code transition is underway but contested. IBM's pivot to bivariate bicycle codes represents a strategic bet that surface codes cannot scale efficiently. Riverlane predicts widespread adoption of qLDPC codes by 2026. However, qLDPC codes require long-range connectivity that superconducting architectures do not naturally provide. This creates a potential architectural mismatch: the codes that offer the best overhead may favor platforms (neutral atoms, trapped ions) with inherent all-to-all connectivity over the superconducting systems that currently dominate.
Sources: Riverlane Blog (2026) (↗); arXiv (2025) (↗)
Neutral atoms are underreported relative to their results. The academic literature and practitioner blogs document neutral-atom systems achieving the largest logical qubit counts and longest coherence times. Yet public attention disproportionately focuses on Google and IBM announcements. This may reflect the absence of a dominant neutral-atom company with Google or IBM's marketing resources, or a structural lag in recognizing platforms that originated in atomic physics rather than solid-state engineering.
Sources: Nature (2026) (↗); Nature (2025) (↗)
The Aaronson-Kalai debate remains unresolved. Scott Aaronson's December 2025 post described fault-tolerant Shor's algorithm as a "live possibility" before the next US election, representing his most optimistic public statement in years. Gil Kalai maintains that correlated noise will prevent any scalable demonstration, a position he defended in detail through March 2026. This is not a marginal disagreement: Kalai argues from within physics and mathematics that scalable quantum computing is impossible, while Aaronson considers recent error correction results strong evidence that the theoretical objections have been answered experimentally. The blogosphere is the primary venue where this debate is conducted with technical rigor.
Sources: Shtetl-Optimized (Scott Aaronson's personal blog) (2025) (↗); Combinatorics and More (Gil Kalai's blog) (2026) (↗)
Microsoft's credibility problem is specific and documented. The 2018 retraction of a Nature paper claiming Majorana zero mode signatures created lasting reputational damage. Independent commentators, including Quantum Zeitgeist (which graded Microsoft's quantum program B- before the Majorana 1 announcement) and an anonymous Substack critic who called the Majorana framing "a form of misinformation," have applied heightened scrutiny to Microsoft claims. This skepticism is not generic anti-hype sentiment but a specific response to a specific institutional failure.
Sources: Substack (independent author) (2025) (↗); The Quantum Insider (2025) (↗)
Quantum machine learning advantage claims face sustained criticism. A November 2025 arXiv paper titled "Quantum Deep Learning Still Needs a Quantum Leap" documented that proposed quantum ML algorithms have not demonstrated advantage over classical methods on practical tasks. A companion Nature Communications paper showed quantum advantage for learning shallow neural networks on specific data distributions, but this remains far from general-purpose ML capability. The gap between quantum ML hype and demonstrated results is one of the field's clearest examples of overclaiming.
Sources: arXiv (2025) (↗); Nature Communications (2025) (↗)
Open Questions
What is the true physical-to-logical qubit ratio required for useful computation? Current demonstrations range from roughly 24:1 (IBM qLDPC) to roughly 100:1 (Google surface code at distance 7). Estimates for running Shor's algorithm on RSA-2048 require thousands of logical qubits with code distances of 17 or higher, implying millions of physical qubits under pessimistic assumptions. The actual ratio depends on achievable physical error rates, code choices, and decoder performance, all of which remain active research areas.
Will qLDPC codes be practically implementable on superconducting hardware? IBM's roadmap assumes that purpose-built processor architectures can provide the long-range connectivity qLDPC codes require. Whether this is achievable at scale without unacceptable overhead from routing and crosstalk is unknown. The alternative is that qLDPC codes may favor platforms with native all-to-all connectivity, potentially shifting the hardware landscape.
Has Microsoft demonstrated topological protection? The February 2025 announcement and subsequent APS March Meeting discussions left this question genuinely open. If topological qubits work as theorized, they would provide hardware-level error suppression that could dramatically reduce software overhead. If the Majorana 1 results do not hold up, topological qubits may remain a theoretical possibility without practical realization for another decade or more.
Sources: Nature (2025) (↗); APS Physics (2025) (↗)
What problem instances will demonstrate useful quantum advantage? The field's leading researchers acknowledge that no concrete problem instance has been identified where quantum computers outperform the best classical methods on a practically meaningful task. Google's Grand Challenge perspective explicitly frames this as the hardest unsolved problem in the field. Until such instances are identified and demonstrated, commercial projections rest on theoretical extrapolation rather than evidence.
Sources: arXiv (2025) (↗)
Is the algorithm gap closable? The Aaronson-Childs-Farhi-Harrow panel documented that no new algorithm with proven asymptotic advantage for important problems has emerged since 1996. Whether this reflects a fundamental limitation or a temporary research drought is unknown. Hardware improvements without algorithmic breakthroughs may produce powerful machines with limited practical applications.
What is the realistic timeline to cryptographic relevance? Expert estimates cluster around a 30-40% probability by 2034, implying majority uncertainty. Post-quantum cryptography standardization is proceeding on the assumption that the threat may materialize within the decade, but the actual timeline depends on engineering progress that cannot be reliably projected from current results.
Can neutral-atom systems maintain coherence and gate fidelity at 10,000+ qubit scale? Current demonstrations at 3,000-6,000 atoms are impressive but an order of magnitude below what would be needed for large-scale fault-tolerant computation. Scaling effects, loading losses, and gate-speed limitations at larger scales are not yet characterized.
Sources: Nature (2025) (↗); Nature (2025) (↗)
![[sources-quantum-computing-fundamentals-briefing-error-corr]]
Sources
Summary: ↑ Back to summary
Academic & arXiv
| 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. |
Frontier Lab & Model News
| ID | Title | Outlet | Date | Significance |
|---|---|---|---|---|
| t1 | Meet Willow, our state-of-the-art quantum chip | Google Quantum AI Blog | 2024-12 | Official announcement of Willow's below-threshold quantum error correction and benchmark claim of completing a computation in under five minutes that would take a supercomputer 10 septillion years, a milestone celebrated and critiqued across the field. |
| t2 | Quantum error correction below the surface code threshold (Willow paper, Nature) | Nature | 2024-12 | Peer-reviewed Nature publication confirming Willow's first 'below-threshold' quantum error correction, with the error rate at the logical qubit level decreasing as more physical qubits are added—a three-decade milestone. |
| t3 | AlphaQubit: Google's research on quantum error correction | Google DeepMind Blog | 2024-11 | Announces AlphaQubit, a transformer-based AI decoder jointly developed by Google DeepMind and Google Quantum AI that makes 6% fewer errors than tensor-network methods and 30% fewer than correlated matching, published in Nature. |
| t4 | Our Quantum Echoes algorithm is a big step toward real-world applications for quantum computing | Google Quantum AI Blog | 2025-10 | Claims the first-ever verifiable quantum advantage on hardware: the Quantum Echoes (OTOC) algorithm ran 13,000× faster on Willow than the fastest classical supercomputer, with results published in Nature and independently verified against NMR data. |
| t5 | Google Claims Quantum Advantage with Willow Chip | HPCwire | 2025-10 | Detailed independent technical coverage of Google's Quantum Echoes Nature paper, including context on the distinction between verifiable quantum advantage and earlier supremacy claims, with caution about classical counterattacks. |
| t6 | Microsoft's Majorana 1 chip carves new path for quantum computing | Microsoft Source | 2025-02 | Official Microsoft announcement of Majorana 1, described as the world's first quantum chip with a Topological Core, placing eight topological qubits on a chip designed to scale to one million, with an accompanying Nature paper. |
| t7 | Microsoft unveils Majorana 1, the world's first quantum processor powered by topological qubits | Microsoft Azure Quantum Blog | 2025-02 | Technical blog post explaining the Majorana 1 roadmap, from single-qubit devices to quantum error correction arrays, detailing the role of Majorana Zero Modes, interferometric parity measurements, and hardware-protected topological qubits. |
| t8 | Microsoft's Claim of a Topological Qubit Faces Tough Questions | APS Physics | 2025-03 | Peer-reviewed expert commentary reporting that condensed-matter physicists at Cornell and NYU questioned Microsoft's topological qubit data at APS March Meeting, with Nature's editorial team noting the paper does not yet represent evidence for topological modes. |
| t9 | Microsoft's Topological Qubit Claim Faces Quantum Community Scrutiny | The Quantum Insider | 2025-02 | Captures the scientific community debate around Majorana 1, including Scott Aaronson's cautious endorsement and Microsoft's 2018 retraction history, providing balanced context on the contested nature of the announcement. |
| t10 | Topological quantum processor marks breakthrough in computing | UC Santa Barbara (The Current) | 2025-02 | Academic institution perspective from Microsoft Station Q's home institution, explaining the physics of Majorana zero modes in topological qubits and contextualizing the eight-qubit Majorana 1 as a proof-of-concept rather than a commercial device. |
| t11 | IBM lays out clear path to fault-tolerant quantum computing | IBM Quantum Blog | 2025-06 | IBM details its end-to-end modular framework for fault-tolerant quantum computing based on bivariate bicycle (qLDPC) codes, with the gross code encoding 12 logical qubits in 288 physical qubits—10× more efficient than surface codes. |
| t12 | High-threshold and low-overhead fault-tolerant quantum memory (IBM bivariate bicycle Nature paper) | Nature (via PMC) | 2024-03 | The landmark IBM Nature paper (Bravyi, Cross, Gambetta et al.) introducing the bivariate bicycle qLDPC code, which became the foundation of IBM's entire fault-tolerant roadmap and sparked industry-wide adoption of high-rate LDPC codes. |
| t13 | IBM Delivers New Quantum Processors, Software, and Algorithm Breakthroughs on Path to Advantage and Fault Tolerance | IBM Newsroom | 2025-11 | IBM QDC 2025 announcement revealing Quantum Loon (all key processor components for fault-tolerant computing demonstrated) and Nighthawk (120-qubit, 5,000-gate processor), with real-time qLDPC decoding under 480 nanoseconds achieved one year ahead of schedule. |
| t14 | IBM Offers Roadmap Toward Large-Scale, Fault-Tolerant Quantum Computer at New IBM Quantum Data Center | The Quantum Insider | 2025-06 | Comprehensive coverage of IBM's Starling roadmap targeting a 200-logical-qubit, 100-million-operation fault-tolerant system by 2029, with intermediate steps Loon (2025), Kookaburra (2026), and Cockatoo (2027) clearly defined. |
| t15 | Engineering Fault Tolerance: IBM's Modular, Scalable Full-Stack Quantum Roadmap | The Quantum Insider | 2025-06 | Technical analysis of IBM's pivot from surface codes to qLDPC codes, explaining how bivariate bicycle codes reduce physical qubit overhead by up to 90% and enable the non-local connectivity critical to scalable fault-tolerant architectures. |
| t16 | Scaling for quantum advantage and beyond (IBM QDC 2025 blog) | IBM Quantum Blog | 2025-11 | IBM's official QDC 2025 state-of-the-union detailing Nighthawk's 120-qubit, 5,000-gate milestone, Heron's lowest-ever median two-qubit gate errors, and the company's dual-track strategy of near-term advantage and long-term fault tolerance. |
| t17 | A fault-tolerant neutral-atom architecture for universal quantum computation | Nature | 2025-11 | Harvard/MIT team demonstrates key elements of a universal fault-tolerant architecture using up to 448 reconfigurable neutral atoms, a peer-reviewed advance establishing neutral atoms as a credible path to fault-tolerant quantum computation. |
| t18 | Continuous operation of a coherent 3,000-qubit system | Nature | 2025-09 | Nature paper demonstrating continuous operation of more than 3,000 physical qubits in a neutral-atom array with coherent storage, enabling deep-circuit quantum evolution and real-time syndrome extraction without atom losses halting computation. |
| t19 | A tweezer array with 6,100 highly coherent atomic qubits | Nature | 2025-09 | Demonstrates an optical tweezer array of 6,100 neutral-atom qubits across 12,000 sites, indicating that quantum computing with 6,000+ qubits is a near-term prospect and providing a path toward QEC with hundreds of logical qubits. |
| t20 | A manufacturable platform for photonic quantum computing | Nature | 2025-02 | PsiQuantum-led Nature paper introducing silicon-photonics-based modules with photonic qubit state-preparation fidelity of 99.98%, providing the first demonstrated manufacturable platform architecture for photonic quantum computing. |
| t21 | Scaling and networking a modular photonic quantum computer | Nature | 2025-01 | Proof-of-principle study using 35 photonic chips (the 'Aurora' machine) demonstrating a complete photonic quantum computer architecture capable of universal and fault-tolerant operation once component performance thresholds are reached. |
| t22 | Fault-tolerant quantum computation with polylogarithmic time and constant space overheads | Nature Physics | 2025-11 | Theoretical proof that QLDPC codes combined with concatenated Steane codes achieve constant space overhead and polylogarithmic time overhead, resolving a longstanding gap in fault-tolerant quantum computation theory. |
| t23 | Making fault-tolerant quantum computers a reality | McKinsey & Company | 2025-12 | McKinsey's structured industry assessment of quantum error correction challenges across qubit modalities, with data showing quantum start-up investment was 50% higher in 2024 than 2023 and projecting the quantum market near $100 billion by 2035. |
| t24 | The Year of Quantum: From concept to reality in 2025 | McKinsey & Company (Quantum Technology Monitor) | 2025-06 | McKinsey's annual Quantum Technology Monitor estimating quantum technology revenue growing to $72 billion in computing alone by 2035, documenting $1.8 billion in government funding in 2024 and cataloguing error-correction progress across labs and start-ups. |
| t25 | Neutral atom quantum computing hardware: performance and end-user perspective | EPJ Quantum Technology (Springer Nature) | 2023-08 | Peer-reviewed industrial end-user review of neutral-atom quantum hardware covering physical qubit architecture, gate fidelities, connectivity, and fault-tolerant prospects, providing honest trade-off analysis against other modalities. |
Blogs & Independent Thinkers
| ID | Title | Outlet | Date | Significance |
|---|---|---|---|---|
| b1 | Quantum Computing: Between Hope and Hype | Shtetl-Optimized (Scott Aaronson's personal blog) | 2024-09 | Aaronson's most comprehensive 2024 state-of-the-field address, covering error correction progress, Google Willow, Microsoft's Quantinuum collaboration, and the honest gap between theoretical speedup and practical utility. |
| b2 | More on whether useful quantum computing is 'imminent' | Shtetl-Optimized (Scott Aaronson's personal blog) | 2025-12 | Aaronson's December 2025 reassessment, calling it a 'live possibility' that a fault-tolerant computer running Shor's algorithm could appear before the next US presidential election — a major update from his prior skeptical posture. |
| b3 | FAQ on Microsoft's topological qubit thing | Shtetl-Optimized (Scott Aaronson's personal blog) | 2025-02 | Aaronson's point-by-point FAQ on Microsoft's Majorana 1 claim, noting Nature's own reviewers found no direct evidence for Majorana zero modes and calling the commercial utility 'not yet' in unequivocal terms. |
| b4 | And yet quantum computing continues to progress | Shtetl-Optimized (Scott Aaronson's personal blog) | 2024-04 | Detailed tracking of Quantinuum and ion-trap progress in early 2024, with Aaronson comparing his 2024 skepticism of fault-tolerant QC to mid-1960s skepticism of moon landings — acknowledging unexpected obstacles remain possible. |
| b5 | Book Review: 'Quantum Supremacy' by Michio Kaku (tl;dr DO NOT BUY) | Shtetl-Optimized (Scott Aaronson's personal blog) | 2023-05 | Canonical debunking of popular quantum hype, illustrating the gap between media-friendly narratives and the rigorous consensus on what current NISQ devices can and cannot do. |
| b6 | Scott Aaronson's View of my View About Quantum Computing | Combinatorics and More (Gil Kalai's blog) | 2026-03 | Captures the definitive Aaronson-vs-Kalai debate in March 2026: Aaronson says Kalai's skeptic path 'has been getting narrower and narrower' as experimental milestones accumulate, offering the most current snapshot of the field's honest internal debate. |
| b7 | Quantum Computing Skepticism, Part 2: My View and Responses to Skeptical Claims | Combinatorics and More (Gil Kalai's blog) | 2025-02 | Kalai's 2025 restatement of his correlated-noise argument against scalable QC, updated with commentary on Microsoft's Majorana 1 claims and featuring counterpoints from Preskill, Aaronson, Harrow and Barak. |
| b8 | Quantum Computers: A Brief Assessment of Progress in the Past Decade | Combinatorics and More (Gil Kalai's blog) | 2025-12 | An ongoing reference post updated through December 2025, tracking where Kalai's skeptical predictions stand against accumulating evidence including Aaronson's revised timeline for Shor's algorithm on fault-tolerant hardware. |
| b9 | Five Perspectives on Quantum Supremacy | Combinatorics and More (Gil Kalai's blog) | 2024-08 | Kalai curates five expert views (Aaronson, Frolov, Emerson, Sondhi and himself) on whether and how quantum supremacy has or can be achieved, providing a multi-voice audit of the field's most contested claim. |
| b10 | Quantum computing: hype vs reality | More Is Different (Substack) | 2025-01 | Independent first-principles analysis from a physics-literate blogger, arguing that quantum volume scaling trends (doubling every ~4 months) are unlikely to sustain Moore's-law-style extrapolation, and that quantum optimization advantage claims remain 'far from settled'. |
| b11 | The YEAR in Quantum Computing – Wrapping up 2024! | Quantum Pirates (Substack) | 2024-12 | Practitioner-level year-in-review covering QEC milestones, Microsoft-Atom Computing 24-logical-qubit announcement, and the industry consensus that MegaQuOp machines — not NISQ — will define the next era. |
| b12 | Quantum 2025 Wrapped – The YEAR in Quantum Computing, December 29th, 2025 | Quantum Pirates (Substack) | 2025-12 | Comprehensive narrative review of 2025's quantum year, identifying the hybrid GPU-QPU stack (NVQLink, IBM-Fugaku) and Microsoft's contested Majorana 1 as the year's two defining story arcs, with early commercial wins from HSBC and D-Wave cited. |
| b13 | Quantum Predictions for 2025 | Quantum Zeitgeist (Substack) | 2024-12 | Independent prediction post calling the death of NISQ and the rise of logical qubits, with candid assessment that Microsoft's topological hardware 'hasn't yielded any working machines' and grades Microsoft a 'B–' relative to Google. |
| b14 | The Quantum Matrix | Quantum Zeitgeist (Substack) | 2026-04 | April 2026 landscape map arguing that the gap between lab demonstrations and commercial deployments is 'closing faster than almost anyone predicted', positioning capital flows and hardware shipping as the key acceleration signal. |
| b15 | Microsoft Quantum Majorana 1 Topological Qubit Approach | Substack (independent author) | 2025-02 | Sharply critical independent Substack analysis calling Microsoft's PR on Majorana 1 'a kind of illegal lobbying' and arguing topological qubits are a 'moonshot at best', citing Jensen Huang's 15–30-year timeline as more credible than Satya Nadella's framing. |
| b16 | 2025 Year-End News Digest: Quantum Error Correction (QEC) | Dr. Bob Sutor – Quantum and AI (Substack) | 2025-12 | Month-by-month 2025 QEC milestone log by IBM's former quantum chief, covering every major lab result across Google, IBM, QuEra, Quantinuum, Nord Quantique, and others — the most thorough single digest of the year's QEC activity. |
| b17 | Quantum – Dr. Bob Sutor Weekly Commentary (March 3 2025) | Dr. Bob Sutor – Quantum and AI (Substack) | 2025-03 | Practitioner commentary covering IonQ leadership change, AWS/Caltech cat-qubit QPU announcement, and the theme of cutting through hype — representative of Sutor's role as an authoritative weekly monitor of industry events. |
| b18 | Quantum Computing 2025: From Verifiable Advantage to Fault-Tolerant Architectures | Medium (nehalmr) | 2026-01 | Technical Medium post documenting 2025's seven key milestones including QuEra's Algorithmic Fault Tolerance (10–100× overhead reduction) and neutral-atom coherence times of 12.6 seconds at 6,100-atom scale — one of the most technically dense independent summaries found. |
| b19 | [Quantum Sundays | 47⟩ — From Qubit Counts to Logical Reality: Mapping the Modern Quantum Computing Landscape](https://medium.com/@adnanmasood/quantum-sundays-47-from-qubit-counts-to-logical-reality-mapping-the-modern-quantum-computing-9edff59a7383) | Medium (Adnan Masood, PhD) | 2026-01 |
| b20 | The Quantum Leap: How 2024–2025 Became the Turning Point for Quantum Computing | Medium (Arun Bansal) | 2025-10 | Accessible synthesis of the 2024–2025 inflection point, covering IBM's modular Starling roadmap, Jensen Huang's CES 2025 market-moving comments, and the investment pattern shift toward hardware over software — useful for the 'current state' research angle. |
| b21 | The Shocking Breakthroughs in Quantum Error Correction of 2025 | Medium (Shailendra Kumar) | 2025-09 | Accessible explainer on 2025's QEC advances including Google's AlphaQubit AI decoder and Microsoft's 4D geometric coding, bridging lab results to a non-specialist technical audience. |
| b22 | Quantum Error Correction: Our 2025 Trends and 2026 Predictions | Riverlane Blog | 2026-01 | Industry practitioner analysis from a leading QEC-specialist firm, documenting that 120 peer-reviewed QEC papers appeared January–October 2025 (up from 36 in 2024), identifying the QuOps metric as the new standard for progress measurement, and predicting mass adoption of qLDPC codes in 2026. |
| b23 | 2024's Quantum Error Correction Highlights (aka the 12 Days of QEChristmas) | Riverlane Blog | 2024-12 | Narrative recap of 2024's twelve most significant QEC events, including Quantinuum's 800× logical vs physical error rate reduction and IBM's bivariate bicycle codes — written by practitioners who build QEC decoders and therefore credibly distinguish hype from substance. |
| b24 | Quantum Error Correction Update 2024 | O'Reilly Radar | 2024 | Authoritative technical-practitioner survey of the 2024 QEC landscape, noting NISQ's commercial failure, the 'death' of hybrid NISQ-classical approaches, and the emergence of hardware-assisted logical qubits from Alice & Bob and Nord Quantique as the field's next phase. |
| b25 | Microsoft's Topological Qubit Claim Faces Quantum Community Scrutiny | The Quantum Insider | 2025-02 | Structured community-reaction piece aggregating expert responses to Microsoft's Majorana 1 announcement, including Aaronson's FAQ, Hensinger's critique of the press release vs. the Nature paper, and the key fact that the Nature editorial team itself found 'no evidence for Majorana zero modes'. |