A quantum computer has achieved quantum advantage on a practically relevant problem.
Different aspects of the evidence point in different directions or support competing explanations. State definitions
No evidence has been admitted since this assessment.
Retained questions may have been narrowed or reframed by the current assessment.
Verification position derived from the record’s assessments; dates show when Faultline first recorded each stage.
Normal Record Review is the Observatory’s assessment process for newly considered evidence, not external scientific peer review. LPR identifies an internal source-review record. PROG identifies a programme: AI (Artificial Intelligence), AM (Advanced Materials), BT (Biotechnology), or QE (Quantum Engineering).
How the claim could hold, what obstructs it, and possible ways forward. The State Warrant above remains the authoritative current assessment.
"Practically relevant" lacks an agreed operational definition. The claim requires advantage on a practically relevant problem, but no agreed standard specifies what practical relevance requires. Different researchers and communities apply different implicit thresholds: some accept scientific relevance (the problem illuminates physical phenomena); others require commercial relevance (the result has identifiable downstream economic value); others require direct application (the computation produces output usable without further classical processing). Without agreement on this threshold, positive and contesting evidence cannot be cleanly compared — the Sycamore result is supportive under some definitions of relevance and irrelevant under others. This is the same threshold-dispute bottleneck as FR-BT-0001 BN-001 and FR-AM-0004 BN-001: not measurement validity, but an undefined success threshold.
The best-classical comparison can improve materially after a quantum result is published. The IBM kicked-Ising episode provides a directly sourced example: a regime beyond brute-force computation and beyond the approximations tested in the original quantum study was subsequently reproduced by stronger tensor-network methods. This does not establish that classical methods will erase every proposed quantum advantage, but it means durability must be evaluated against the best method available for the specific problem at the time of comparison.
A recorded possible resolution path, not automatically a sufficient test for closing the claim.
First quantum computation on a practically relevant problem that establishes a durable advantage over the best contemporaneous classical method under a transparent comparison. A fault-tolerant quantum-chemistry calculation is one plausible route, but the record is not restricted to chemistry or to fault-tolerant hardware if another scientifically or commercially relevant problem satisfies both components at the demonstrated scale.
Historical narrative recorded for this claim. It does not override the current State Warrant.
Questions retained in this record. The current State Warrant may have narrowed or reframed them. Recorded dates refer to question history, not necessarily the age of the current wording; known revision dates are shown separately.
Is there a problem type that is both classically intractable at demonstrated quantum scales and practically relevant? The two trajectories (demonstration advantage, relevant-problem simulation) need to converge on a single instance. Which specific problem will first satisfy both simultaneously?
Question recorded 2024-01-15Could a single fault-tolerant quantum-chemistry computation beyond the best contemporaneous classical method satisfy both FR-QE-0007 and FR-QE-0006, if it also meets each record's practical-relevance and comparison requirements? Chemistry is one possible shared route, not the definition of both attractors; FR-QE-0007 is not restricted to chemistry or fault-tolerant hardware. If an event meets both claims independently, how should the Observatory represent one event serving two records? This question does not establish a new logging procedure.
Question recorded 2024-01-15The null measurement validity condition held. Does this constitute sufficient evidence to characterise Measurement Validity as a proxy-measurement failure mode specifically, or does the two-occurrence positive evidence (FR-BT-0002, FR-AI-0007) plus one-occurrence negative evidence (FR-QE-0007) constitute a pattern worth a Review Note?
Question recorded 2024-01-15| Mutation | Date | Field | Prior value | Current value |
|---|---|---|---|---|
| M-017 | 2026-10-19 | Reference correctionreference_corrected | Named citations without structured verified links | Verified source links on IN-005What this update recorded for M-017Operator-authorised corpus audit Batch 5 / R05, increment 1. Structured provenance enrichment of already admitted citations; no new evidence or reassessment. Scientific descriptions, vectors, sourceReference, assessments, dates, relationships, review status and all prior mutations preserved. Exact field changes, verification limits and deferred work: docs/reviews/VERIFIED-SOURCE-REPAIRS-INCREMENT-1-2026-10-19.json and companion .md. |
| M-016 | 2026-10-19 | Reference correctionreference_corrected | IN-001 characterises quantum advantage as established at 2019 publication | IN-001 follows FR-QE-0001 AS-002: original comparison unresolved; later supersededWhat this update recorded for M-016Operator-authorised corpus audit Batch 4B / R08 editorial alignment. Correct the final interpretation sentence and dependent vector reason on IN-001 to follow the existing operator-ratified FR-QE-0001 AS-002 ruling of 2026-07-22. Experimental task performance is distinguished from original practical classical infeasibility; later counter-performance is not backdated as proof of original falsity. Partial polarity, sources, dates, counts, assessments, current FRAGMENTING / VS-03 and all previous mutations are preserved. No new evidence admitted; FR-QE-0001 remains closed and untouched. Exact prior wording and basis: docs/reviews/BATCH-4B-QE-0007-SYCAMORE-2026-10-19.md. |
| M-015 | 2026-10-14 | Provenance repairprovenance_repair | LPR-001-D50 pending | LPR-001-D50 completedWhat this update recorded for M-015Operator-approved bounded attribution correction completed for IN-006. The structured Nature citation title attached to DOI 10.1038/s41586-025-09526-6 was corrected from the noncanonical title Quantum echoes and the computational complexity of many-body dynamics to the DOI-resolved title Observation of constructive interference at the edge of quantum ergodicity. DOI, URL, locator, sourceReference, evidence description, vectors, assessment, Pressure State and Verification Stage remain unchanged. provenanceOutcome advanced to discrepancies_corrected and provenanceRepairStatus to completed. |
| M-014 | 2026-10-14 | Source reviewprovenance_review | LPR-001-D19 | LPR-001-D50 REVIEW REQUIREDWhat this update recorded for M-014Second-cycle LPR audited all six evidence instances. IN-001 through IN-005 were source-verified. IN-006 is substantively bounded correctly, but its structured Nature citation gives DOI 10.1038/s41586-025-09526-6 a noncanonical title; the DOI resolves to Observation of constructive interference at the edge of quantum ergodicity. The attribution defect is left unchanged pending operator-approved bounded correction. No verification-stage discrepancy and no new Record Review candidate. |
| M-013 | 2026-10-08 | Reference correctionreference_corrected | Stale explanatory cross-reference, regimen or count premise | Explanatory fields aligned with existing record-specific evidence and corpus boundariesWhat this update recorded for M-013Editorial Correction (GP-001), second consistency batch: openQuestions.1.question aligned with existing admitted evidence, corrected cross-references and the dated corpus checkpoint. No new evidence, reassessment, taxonomy or logging procedure. Claims, instances, assessments, status, provenance metadata, original question IDs and raised dates, and prior mutations unchanged. Previous values and source basis: docs/reviews/MCP-SECONDARY-CONSISTENCY-REPAIR-2026-10-08.json. |
| M-012 | 2026-09-17 | Provenance correctedprovenance_corrected | LPR-001-D19 discrepancies | PASS-AFTER-CORRECTIONWhat this update recorded for M-012LPR-001-D19 bounded correction: repaired IN-003, IN-004 and IN-006; added confidently established structured provenance; separated IBM's tested classical-comparison boundary from stronger legacy claims; separated Google's random-circuit-sampling evidence from the distinct Willow surface-code result and attributed company benchmark estimates; reconstructed the Quantum Echoes source chain and scale boundary; narrowed RM-001; generalised AT-001 to the claim's actual conjunction; repaired lineage; and appended AS-003 rather than rewriting historical assessments. The IBM/University of Chicago 2026 hard-circuit result remains a separate Normal Record Review candidate. Pressure State remains FRAGMENTING; Verification Stage remains VS-03. |
| M-011 | 2026-09-17 | Source reviewprovenance_review | — | LPR-001-D19 REVIEW REQUIREDWhat this update recorded for M-011LPR-001-D19 audited all six evidence instances. IN-001, IN-002 and IN-005 were substantially verified against their cited source chains; IN-003 and IN-004 contain overbroad or conflated classical-comparison/Willow representations; IN-006 contains a source-chain and representation package requiring bounded reconstruction before structured provenance is added. One new Normal Record Review candidate was surfaced: IBM and University of Chicago, Sampling hard circuits with verifiably high fidelity (30 July 2026); it was not admitted through LPR-001. |
| M-010 | 2026-09-06 | Description restoreddescription_restored | Legacy ingestion cutoffs: mechanisms:BN-001, mechanisms:RM-001, mechanisms:AT-001 | Source-restored complete descriptionsWhat this update recorded for M-010Editorial Correction (GP-001), RENDER-PILOT-001 content restoration: restored BN-001, RM-001, AT-001 from FR_QE_0007_quantum_advantage_practical.html (Drive file 1qCC0WcnKOMiutuuqh5ah_6wXPN0X2SQs). Each damaged value was a verified prefix of the recovered source after the existing FR-MF to FR-AM identifier migration. Restored the omitted remainder using the original converter text normalization; no inferred completion. Existing later corrections retained. Restoration recovers historical wording and does not reaffirm it as the current assessment. Assessments, evidence instances, open questions, claim, status and rendering eligibility unchanged. Source hashes and field receipt: docs/reviews/render-pilot-content-restoration.json; current-assessment compatibility review: docs/reviews/RENDER-PILOT-001-CONTENT-RESTORATION.md. |
| M-009 | 2026-08-29 | Provenance enrichedprovenance_enriched | — | PROVENANCE-ENRICHEDWhat this update recorded for M-009PA-002 Provenance Enrichment: structured sources[] added to IN-001 and IN-002; evidentiary prose and assessment unchanged. |
| M-008 | 2026-08-28 | Reference correctionreference_corrected | IN-001–IN-005 lacked instance references; IN-005 misidentified Mi et al. as a superconducting-material phase-transition study | IN-001–IN-006 carry stable references; IN-005 reconstructed from Mi, Kandala and YamamotoWhat this update recorded for M-008GP-001 provenance and description correction following the bounded three-record source/DOI audit. IN-005 had conflated the superconducting processor substrate with the simulated phenomenon and incorrectly attributed a superconducting-material phase transition to Mi et al. The instance now states the verified time-crystal, small-molecule/quantum-magnet and simplified hydrogen-chain/iron-crystal results. AS-001 and AS-002 remain preserved; the correction does not alter their conjunction analysis. Pressure State FRAGMENTING and Verification Stage VS-03 are explicitly retained. |
| M-007 | 2026-08-28 | Assessment issuedassessment_issued | AS-001 | AS-002What this update recorded for M-007AS-002 issued following the bounded FR-QE-0007 impact review. Pressure State FRAGMENTING and Verification Stage VS-03 retained. Quantum Echoes narrows the separation between beyond-classical performance and practical relevance, but the two thresholds remain demonstrated in different experimental regimes; Google-affiliated follow-up analysis is not treated as independent replication. No existing assessment, instance, mechanism, open question, or related record modified. |
| M-006 | 2026-08-28 | Instance appendedinstance_appended | IN-005 | IN-006What this update recorded for M-006IN-006 appended — Google Quantum Echoes and the connected molecular-geometry OTOC programme. Classified as partial evidence: reproducible beyond-classical performance and practical application relevance are both present within the programme but not at the same demonstrated scale. Instance logged before AS-002; no state or verification-stage change at this step. |
| M-005 | 2024-01-15 | Programme panel addedprogramme_panel_added | — | PROGRAMME-PANEL-ADDED |
| M-004 | 2024-01-15 | Null condition metnull_condition_met | — | NULL-CONDITION-MET |
| M-003 | 2024-01-15 | Assessment issuedassessment_issued | — | ASSESSMENT-ISSUED |
| M-002 | 2024-01-15 | Instances loggedinstances_logged | — | INSTANCES-LOGGED |
| M-001 | 2024-01-15 | Record createdrecord_created | — | RECORD-CREATED |
Logged entries are reports on record, not counts of independent studies or patients. Several entries can concern the same underlying evidence. Supportive and Contesting describe an entry’s relation to the claim. Partial indicates limited or mixed bearing; read the details for limitations or legacy status. These labels are not study-quality ratings.
Arute et al. (Google, 2019, Nature) report that their 53-qubit Sycamore processor completes a random circuit sampling task in 200 seconds that they estimate would require 10,000 years on Summit, the world's fastest supercomputer at the time. They claim this demonstrates "quantum supremacy." IBM immediately contests the 10,000-year estimate, arguing that their classical simulation methods can complete the task in 2.5 days. Subsequent classical algorithm improvements reduce the estimated classical time further; Pan, Chen, and Zhang (2022) demonstrate classical simulation of Sycamore-class circuits in hours on a GPU cluster. Random circuit sampling has no known practical application — it was designed as a demonstration task. For this record, the settled interpretation in FR-QE-0001 distinguishes the supported experiment and task performance from the unresolved claim that the task was beyond practical classical reach at the original 2019 comparison point. Later classical work superseded the benchmark advantage without automatically proving that the time-indexed 2019 claim was false when made. Practical relevance remains unestablished: random circuit sampling has no identified application where this speedup produces tangible benefit.
Multiple groups (USTC China with Jiuzhang, Xanadu with Borealis) demonstrate quantum advantage in Gaussian boson sampling — a photonic computational task. USTC reports sampling rates 10^14 times faster than classical simulation. These results claim quantum advantage more robustly than Sycamore: the classical simulation of boson sampling at demonstrated scales is harder and the results have proven more difficult to classically reproduce. However, boson sampling also lacks known practical applications — it was proposed as a demonstration problem. The practical relevance component is again absent. The record notes these as stronger advantage demonstrations than Sycamore but still not satisfying the claim as scoped: advantage without relevant application.
Kim et al. (IBM, 2023, Nature) report accurate expectation values from a 127-qubit Eagle processor at circuit volumes beyond brute-force classical computation and in regimes where the particular approximate classical methods tested in the paper break down. The work introduced the term ‘quantum utility’ for useful quantum computation before fault tolerance; it did not establish that the kicked-Ising calculation was generally beyond the best classical methods. Subsequent tensor-network work, including Tindall et al., reproduced the relevant regime classically with greater accuracy and precision than the quantum result. The episode therefore provides scientifically relevant but contested performance evidence and demonstrates that the classical comparison boundary can move after a quantum result is published. It does not satisfy this record's conjunction of durable quantum advantage and practical relevance.
Google's 2024 random-circuit-sampling work reports experiments up to 67 qubits and 32 cycles in a regime the authors argue is beyond existing classical-supercomputer capabilities. Separately, Google's Willow announcement reports a random-circuit-sampling benchmark completed in under five minutes and estimates an enormous classical runtime; that comparison is a company-reported estimate rather than an independently established bound. The peer-reviewed Willow surface-code paper is distinct evidence concerning below-threshold quantum error correction and is not the source of the random-circuit-sampling advantage claim. Random circuit sampling remains a demonstration task without a known direct practical application, so these results bear on the performance component but do not satisfy this record's practical-relevance requirement. They do not by themselves establish that the quantum/classical simulation gap is monotonically widening.
Multiple groups demonstrate quantum simulation results on physically motivated systems approaching practical relevance. Google Quantum AI's observation of time-crystalline eigenstate order on a quantum processor (Mi et al. 2022), IBM's variational calculations for small molecules and quantum magnets (Kandala et al. 2017), and Quantinuum's hardware calculations for simplified hydrogen-chain and iron-crystal models (Yamamoto et al. 2022) all bear on scientifically relevant problems. None unambiguously exceeds the best classical methods on a directly practical problem at the demonstrated scale. The systems remain deliberately small or simplified and classically checkable; their value is as scientific and algorithmic demonstrations rather than established practical advantage. These are among the closest parts of the pre-2025 evidence base to satisfying both components simultaneously, but the conjunction remains unconfirmed.
Google Quantum AI's peer-reviewed Quantum Echoes experiment measures a second-order out-of-time-order correlator (OTOC) on 65 Willow qubits and reports a comparison of about 2.1 hours of quantum execution with an estimated approximately 3.2 years on the Frontier supercomputer, roughly 13,000 times faster. A separate connected molecular-geometry study uses OTOC measurements to estimate molecular distances and angles with accuracy comparable to independent spectroscopic measurements, with Willow used to simulate the molecular OTOCs. These evidentiary components do not yet coincide at the decisive scale: the large Quantum Echoes circuits supply the reported beyond-classical performance comparison, while the molecular-geometry application is demonstrated on smaller systems that do not themselves establish quantum advantage. A 2026 tensor-network analysis supports the classical-intractability interpretation of the large circuits but includes Google Quantum AI-affiliated authors and is treated as affiliated corroboration rather than independent replication. The programme therefore narrows the gap between performance and practical relevance without satisfying the record's conjunctive claim.