{"id":"fa0c7ca4-40af-41d1-a916-9cf8fce5c3fe","arxiv_id":"2611.16329","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Under local quantum measurements and finite-time local controls, the work extractable from a pure microscopic thermal equilibrium state equals that from the corresponding canonical ensemble in the thermodynamic limit, expressible via quantum-classical mutual information and Holevo information.","lead":"This paper develops a framework for information thermodynamics in closed quantum many-body systems, showing that under local operations the work extractable from a pure thermal state equals that from a canonical ensemble. This provides a rigorous foundation for understanding Maxwell's demon in many-body systems and has implications for quantum simulators and quantum batteries.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"The proof that the outcome-averaged operator is O(L^s)-local relies on the assumption that each feedback term depends only on measurement outcomes from regions within distance ℓ=O(L^s); the GHZ counterexample shows that when this condition is violated the central claim fails, but the paper…","rationale":"The reader's weakest_assumption correctly identifies the MITE condition as the key assumption. My stress-test focuses on a different but related assumption: the locality of feedback dependence on measurement outcomes (Eq. 6). The paper acknowledges this condition and demonstrates its necessity via the GHZ counterexample. However, the physical justification for this condition in generic many-body systems is not deeply discussed. The proof is rigorous under the stated assumptions, and the paper is transparent about the limitations. The GHZ example shows that the condition is not just technical but physically significant. The concrete test I propose would clarify whether the condition is necessary or merely sufficient. I do not find a fatal flaw; the paper's claims are well-supported. The verdict should remain ACCEPT. I agree with the reader's assessment that the MITE condition is the weakest assumption, but I add that the feedback locality condition is comparably load-bearing. The paper's own discussion of the GHZ example shows they are aware of this, so no change to the verdict is needed.","tokens_in":29462,"tokens_out":946,"duration_ms":9265,"concrete_test":"Construct a concrete many-body model (e.g., a 1D Ising chain with transverse field) where the initial state is MITE for s=0 (ordinary MITE), the measurement is local (d_M=1), but the feedback Hamiltonian depends on measurement outcomes from a distance ℓ that scales as L^α with 0<α<1. Compute the extractable work numerically for small system sizes (L up to, say, 12) and compare with the canonical ensemble prediction. If the deviation from Eq. (2) scales as L^{α} rather than being subextensive, then the condition ℓ=O(L^s) with s<1 is necessary and the theorem's applicability is limited to strictly local feedback.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central theorem (Eq. 2) is proved under the assumption that each local feedback term depends only on measurement outcomes from regions within distance ℓ=O(L^s) (Eq. 6). The proof in the Supplemental Material uses this condition to bound the support of the outcome-averaged operator G_X^{(l)}(t) (Eq. S34) and to apply the MITE property. However, the paper's own GHZ example (Eq. 16) demonstrates that when feedback depends on distant outcomes (ℓ ~ L/3), the equivalence breaks down and extensive additional work can be extracted. This is presented as a feature, not a bug. The concern is not that the condition is violated in that example, but that the physical justification for the condition in generic many-body systems is unclear. The condition requires that classical communication of measurement outcomes is spatially local, which is a strong assumption about the feedback controller's architecture. In many experimental implementations (e.g., quantum computers with classical control hardware), the controller can access all measurement outcomes globally, making ℓ ~ L physically plausible. The paper does not discuss how restrictive this locality condition is or provide a physical argument for why it should hold in typical scenarios. Additionally, the proof assumes that the measurement depth d_M satisfies (D-1)d_M = O(L^s) (Eq. S36), which limits the spatial extent of measurement-induced correlations. For s>0, this allows measurements that are not strictly local but have a sublinear range; the proof's handling of the measurement-induced nonlocal state update in this regime relies on the Lieb-Robinson bound for the feedback Hamiltonian, but the measurement itself is not generated by a Hamiltonian and its nonlocal effects are controlled only by the depth condition. The argument that the outcome-averaged operator is O(L^s)-local (Eq. S34) is correct under the stated assumptions, but the physical interpretation of the depth condition for s>0 deserves more scrutiny.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper develops an information-thermodynamic framework for closed quantum many-body systems initially in microscopic thermal equilibrium (MITE), under constraints of local measurements, finite-time local controls, and feedback depending only on nearby measurement outcomes. The main result is that the extractable work from MITE equals that from the canonical ensemble in the thermodynamic limit, expressed via quantum-classical mutual information, Holevo information, and relative entropy. The paper also demonstrates that violating the locality condition (using distant measurement outcomes) allows extraction of additional work from long-range correlations, as illustrated with a GHZ state example.","tokens_in":29923,"tokens_out":1564,"duration_ms":15779,"significance":"The paper bridges information thermodynamics and many-body quantum physics, providing rigorous results under realistic constraints. The central theorem is proved using Lieb-Robinson bounds, and the information-thermodynamic identity is derived algebraically. The examples are explicit and illustrate the main claims. The work is significant for quantum simulators and quantum computers where pure many-body states are common.","major_comments":[{"comment":"The locality condition ℓ = O(L^s) for feedback dependence on measurement outcomes is central to the proof. However, the physical justification for this condition in generic many-body systems is not discussed. In many experimental implementations (e.g., quantum computers with classical control hardware), the controller can access all measurement outcomes globally, making ℓ ~ L physically plausible. The paper should discuss how restrictive this condition is and provide a physical argument for why it should hold in typical scenarios, or at least acknowledge that it limits applicability.","section":"§3, Eq. (6) and Supplemental Material §S1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The assumption (D-1)d_M = O(L^s) limits the spatial extent of measurement-induced correlations. For s>0, this allows measurements that are not strictly local but have sublinear extent. The paper should clarify whether this condition is physically motivated or purely technical. Additionally, the proof assumes that the measurement depth d_M is such that (D-1)d_M = O(L^s), but it is not clear how this relates to the locality of the measurement operators themselves. A brief discussion of the physical interpretation of this condition would strengthen the paper.","section":"§S1, Eq. (S36) and surrounding text","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The information-thermodynamic identity (Eq. 9) is derived for the canonical ensemble initial state, not for MITE. The paper states that this identity does not require any locality constraints. However, the main theorem (Eq. 2) combines this identity with the MITE-canonical equivalence (Eq. 7). The paper should explicitly note that the identity holds for the canonical ensemble protocol, and the MITE result inherits it via the equivalence. This is clear from the text but could be emphasized to avoid confusion.","section":"§4, Eq. (9) and End Matter","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The GHZ example demonstrates that violating the locality condition (using distant measurement outcomes) allows extraction of additional work. However, the paper does not discuss whether this additional work is extensive (O(L)) or subextensive. From Eq. (17), W = gL(3q - q^3)/2, which is extensive. The paper should explicitly state that the additional work is extensive, as this is a key point. Also, the example uses a noninteracting Hamiltonian; it would be helpful to comment on whether similar effects are expected in interacting systems.","section":"§5, Eq. (17) and surrounding text","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The abstract mentions 'microscopic thermal equilibrium (MITE)' but does not define the acronym. Consider defining it explicitly.","section":"Abstract","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The phrase 'the demon's original setting' (line after Ref. [17]) is slightly informal. Consider rephrasing.","section":"§1, Introduction","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The definition of work W = Tr[Hρ_i] - Tr[Hρ_f] assumes that the Hamiltonian is the same at the beginning and end of the protocol. This is stated as a cyclic protocol, but it would be helpful to explicitly note that H_i = H_f = H.","section":"§2, Eq. (1)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The notation for measurement operators is somewhat dense. A brief explanation of the indices (n, Y, k_n,Y) would improve readability.","section":"§3, Eq. (4)-(5)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The definitions of I_QC and χ use the von Neumann entropy of branch states ρ'^{ens}_k. It might be helpful to note that these branch states are normalized (as done in the text).","section":"§4, Eq. (10)-(11)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The measurement operator m_k(η) is defined with a factor 1/√(2(1+η^2)). It would be helpful to verify that the completeness relation ∑_k m_k^† m_k = I holds. A brief check or reference would be useful.","section":"§5, Eq. (12)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The expression for W^{loc-opt}_{ens} has a max(0, ...) function. The paper should clarify that this is the optimal work under single-qubit feedback, not necessarily the global optimum.","section":"§5, Eq. (13)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The GHZ state is written as a product of triplets. It would be helpful to explicitly state that this state is MITE for β=0 (infinite temperature) and that it has long-range entanglement.","section":"§6, Eq. (16)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The feedback unitary u_+ = σ_1 is used. The paper should note that this is a bit-flip operation, which is physically realizable.","section":"End Matter, Eq. (24)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The constant C is defined as c_LR J c_0/(κv). It would be helpful to explicitly state that this constant is independent of L, l, and t.","section":"Supplemental Material, Eq. (S16)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The support of G^{[l]}_X(t) is bounded by X_{R_M} with R_M = l + ℓ + (D-1)d_M. The paper should note that this is a key step in the proof.","section":"Supplemental Material, Eq. (S34)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The table summarizes operational time vs extractable work. It would be helpful to include a reference to the relevant equations or sections for each case.","section":"Table I","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"minor_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"The actual new content: information thermodynamics for closed many-body systems starting from possibly-pure MITE states, under local measurements and finite-time local controls. The main theorem says that when each feedback term depends only on outcomes from a nearby region, the extractable work satisfies βW ≃ βW_ens = I_QC − χ − D(ρ_ens^f ∥ ρ_ens^i). The second equality is algebra. The first is the content, and the proof in the supplemental material looks sound to me: the Lieb-Robinson truncation of U_k^† h_X U_k, the layer-by-layer argument that the outcome-averaged operator is supported on X_{R_M} with R_M = l + ℓ + (D−1)d_M, and the MITE replacement at the end. Errors are tracked per unit volume, the s=0 case is handled by taking l large but L-independent, and the conditions t, t′ = O(L^s), (D−1)d_M = O(L^s), ℓ = O(L^s) are stated plainly in the theorem. No circularity, no fitted parameters. The identity is verified explicitly in the qubit example, and the GHZ triplet calculation genuinely beats the canonical bound: W > W_opt > W_ens for 0 < η < 1, confirming Eq. (7) fails when feedback reads distant outcomes. Credit where due: the authors mark the boundary of their own result rather than obscuring it.\n\nSoft spots, in proportion. First, the ℓ locality condition on feedback is load-bearing and its physical justification is one sentence ('classical communication is spatially local'). The stress-test note is right that many real architectures — superconducting qubits, trapped ions with a classical control bus — give the controller global access, making ℓ ~ L plausible. But that makes this a scope-and-motivation issue, not a hole in the proof. I would push a referee to ask for a concrete argument about which experimental settings satisfy the condition, beyond asserting it. Second, a minor textual tension: the main text at [b43] says 'L-independent distance' while Eq. (6) and the theorem use ℓ = O(L^s); for s > 0 these differ and should be harmonized. Third, in the s > 0 regime the measurement is not Hamiltonian-generated, and the nonlocal state update is controlled only by the depth condition; the averaging argument handles it, but the paper could spend a sentence explaining why the depth bound suffices — currently the reader reconstructs it.\n\nWho it is for: quantum thermodynamics and information thermodynamics researchers, and simulators where a pure state is the realistic resource. This is a solid, honest paper whose assumptions are neither hidden nor overclaimed. Deserves proper refereeing; I would accept with revisions mostly directed at motivating the locality conditions and tidying the s > 0 statements.","headline":"Clean theorem, honest scope: work from MITE matches the canonical ensemble under outcome-local feedback, and the GHZ example marks the boundary; the soft spot is motivation for the locality conditions, not the proof.","tokens_in":30409,"tokens_out":3009,"duration_ms":21265},"pith_extraction":{"msc":[],"pacs":["05.30.-d","05.70.-a"],"model":"deepseek/deepseek-v4-flash","headline":"A demon on a pure state that looks thermal locally extracts the same work as from the Gibbs ensemble when all measurements, controls, and feedback are local.","keywords":["information thermodynamics","Maxwell's demon","microscopic thermal equilibrium","quantum feedback control","work extraction","Lieb-Robinson bound","quantum-classical mutual information","Holevo information"],"falsifier":"On a programmable quantum device, prepare the $\\beta=0$ MITE state of GHZ triplets, apply the noisy single-qubit $\\sigma_3$ measurement of strength $0<\\eta<1$, and flip each triplet's spins with a $\\sigma_1$ rotation when the majority outcome is $+$: the paper predicts work $W = gL(3q-q^3)/2$ with $q=2\\eta/(1+\\eta^2)$, which exceeds the canonical-optimal value $gLq$; observing $W \\le gLq$ for any $\\eta\\in(0,1)$ would falsify the claim that distant-outcome feedback harvests hidden correlations. In the local-feedback regime, the theorem predicts $|W - W_{\\rm ens}|/V \\to 0$ in the thermodynamic limit together with a sharp onset of zero work for $\\beta > \\beta_c = g^{-1}\\operatorname{arctanh}(2\\eta/(1+\\eta^2))$, either of which is directly measurable.","tokens_in":31007,"feed_emoji":"⚛️","tokens_out":18678,"duration_ms":107551,"temperature":0.7,"pith_summary":"Maxwell's demon extracts work by measuring a system and acting on the outcome, and information thermodynamics accounts for the demon's apparent violation of the second law in terms of the information gained. This paper asks how that accounting works in a large closed quantum system that begins in a pure state which is thermal for every local observable but is not a true Gibbs ensemble. Its main theorem says that if the measurements are local, the unitary controls are generated by local Hamiltonians over times that grow sublinearly with system size, and each feedback term depends only on measurement outcomes obtained nearby, then the work extractable from such a microscopically thermal equilibrium (MITE) state equals, up to subextensive corrections, the work extractable from the corresponding canonical ensemble, with the common value $\\beta W \\simeq \\beta W_{\\rm ens} = I_{\\rm QC} - \\chi - D(\\rho_{\\rm ens}^{f} \\| \\rho_{\\rm ens}^{i})$. This matters because pure thermal-looking states are far easier to prepare on quantum simulators than genuine Gibbs states, and because the identity shows precisely where the second law keeps its force when a demon is present. The paper also maps the boundary of the theorem: feedback based on distant measurement outcomes can harvest long-range correlations hidden in the MITE state and extract extensively more work than the canonical bound allows.","feed_headline":"Local Maxwell's demon can't beat the Gibbs work bound","feed_subtitle":"Pure states that look thermal locally yield the same work as Gibbs states — unless feedback uses distant outcomes.","key_machinery":"Two objects carry the argument. The first is MITE, the class of (possibly pure) states whose local-observable expectation values are indistinguishable from the canonical ensemble; it is the assumption that lets a pure state stand in for a heat bath. The second is the Lieb-Robinson bound, applied to the outcome-averaged Heisenberg operator $\\sum_k U_1^\\dagger M_k^\\dagger U_k^\\dagger(t) h_X U_k(t) M_k U_1$: a single measurement trajectory updates the state nonlocally, but averaging over outcomes cancels the trajectory-specific nonlocal parts, and Lieb-Robinson locality shows the averaged operator is exponentially close to one supported on an $O(L^s)$ region as long as the feedback light cone, the measurement causal cone, and the pre-measurement light cone all remain sublinear. Once that operator is local, the MITE property transfers the final-state energy from the pure state to the Gibbs state, completing the equivalence $W \\simeq W_{\\rm ens}$. The third piece is the information-thermodynamic identity $\\beta W_{\\rm ens} = I_{\\rm QC} - \\chi - D(\\rho_{\\rm ens}^f \\| \\rho_{\\rm ens}^i)$, which is exact bookkeeping: the measurement information $I_{\\rm QC}$ is partly spent distinguishing outcomes that remain visible in the final state (the Holevo term $\\chi$) and partly dissipated into the nonequilibrium free energy of the final state (the relative-entropy term).","core_discovery":"Microscopic thermal equilibrium (MITE) is the paper's central object: a possibly pure state whose expectation values for every observable supported on a region of diameter $O(L^s)$ agree with the canonical ensemble $\\rho_{\\rm ens}^{i} = e^{-\\beta H}/Z$ at inverse temperature $\\beta$, for a fixed $0 \\le s < 1$. The main theorem states that for any such initial state, the work $W$ extracted by a single round of local measurement followed by feedback satisfies $\\beta W \\simeq \\beta W_{\\rm ens} = I_{\\rm QC} - \\chi - D(\\rho_{\\rm ens}^f \\| \\rho_{\\rm ens}^i)$, where $W_{\\rm ens}$ is the work the same protocol extracts from the canonical initial state, $I_{\\rm QC}$ is the quantum-classical mutual information (the entropy reduction from learning the outcome), $\\chi$ is the Holevo information (the outcome information that survives in the final state), and $D$ is the quantum relative entropy (the nonequilibrium free energy of the final state). The theorem requires low-depth local measurements, pre-measurement and feedback unitaries generated by local Hamiltonians with durations $O(L^s)$, and feedback terms that depend only on outcomes measured within distance $O(L^s)$ of their support; the proof uses Lieb-Robinson bounds to show that the outcome-averaged Heisenberg operator for any local energy term is exponentially close to an $O(L^s)$-local operator, so the MITE property transfers the final energy from the pure state to the Gibbs state. The identity $\\beta W_{\\rm ens} = I_{\\rm QC} - \\chi - D(\\rho_{\\rm ens}^f \\| \\rho_{\\rm ens}^i)$ itself needs no locality assumptions. The paper closes with the counterexample that defines the boundary: for the $\\beta=0$ MITE state $|\\Psi_0\\rangle = \\bigotimes_j (|{\\uparrow\\uparrow\\uparrow}\\rangle + |{\\downarrow\\downarrow\\downarrow}\\rangle)/\\sqrt{2}$ of GHZ triplets, feedback using distant outcomes extracts $W = gL(3q-q^3)/2$, exceeding both the same-protocol canonical work $W_{\\rm ens} = gLq^2$ and the optimally controlled canonical work $W_{\\rm opt}^{\\rm ens} = gLq$ with $q = 2\\eta/(1+\\eta^2)$.","pith_inferences":["If the theorem is right, quantum simulators gain a practical shortcut: any pure state that is locally indistinguishable from the Gibbs ensemble, such as a quench product or a random microcanonical-typical state, can stand in for the hard-to-prepare canonical state in demon and battery experiments, because only its local expectation values enter the work bound.","The GHZ example suggests a quantitative program the paper leaves open: the excess work $W - W_{\\rm opt}^{\\rm ens}$ should be expressible through multipartite entanglement measures of the MITE state, effectively turning work extraction into an entanglement witness with a thermodynamic scale.","The paper counts only the energy change of the system and excludes the cost of the controller; if broadcasting measurement outcomes globally carries an energy cost, the apparent gain from distant-outcome feedback might be exactly cancelled, a question a full accounting that includes communication costs could settle.","The proof's core device, localizing the outcome-averaged Heisenberg operator, is a general technique: the same argument could carry other thermodynamic functions of MITE states, such as the full counting statistics of work or heat under repeated and continuous monitoring, over to the canonical ensemble."],"forward_implications":["In closed many-body systems the second law survives the demon: for an initially thermal pure state under local measurements and local finite-time feedback, the extractable work satisfies $W \\simeq W_{\\rm ens} \\le W_{\\rm opt}^{\\rm ens}$, so no local protocol beats the best protocol applied to the corresponding canonical ensemble.","Measured information is not directly convertible to work in a closed system: the identity $\\beta W \\simeq I_{\\rm QC} - \\chi - D(\\rho_{\\rm ens}^f \\| \\rho_{\\rm ens}^i)$ shows the yield is reduced by the Holevo information of the outcome-conditional final states and by the relative-entropy cost of leaving the system out of equilibrium.","The exactly solved $V$-qubit example predicts a nonanalytic onset of optimal local work at the critical inverse temperature $\\beta_c = g^{-1}\\operatorname{arctanh}(2\\eta/(1+\\eta^2))$, a control-landscape transition that persists in the thermodynamic limit and is identical for MITE and canonical initial states.","The bound is tight: at $\\beta=0$, locally optimal product feedback already renders every conditional branch passive, so the local protocol reaches the globally optimal canonical work $W_{\\rm opt}^{\\rm ens}$.","Distant-outcome feedback breaks the equivalence: with GHZ-triplet MITE states, constant-time feedback that uses distant measurement outcomes extracts extensively more work than the canonical ensemble allows, demonstrating that hidden long-range correlations are a genuine thermodynamic resource."],"supporting_citations":[{"why":"Supplies the single-round measurement-then-feedback protocol, the definition of extracted work as the closed system's energy decrease, and the quantum-classical mutual information used in the main identity.","marker":"[5]"},{"why":"Defines microscopic thermal equilibrium (MITE), the class of pure states whose local expectation values match the canonical ensemble, which is the paper's starting assumption.","marker":"[41]"},{"why":"The unitary-only predecessor establishing the second law for local finite-time operations, which this work extends to measurement and feedback.","marker":"[42]"},{"why":"Canonical typicality, cited to justify the extended MITE notion with $0<s<1$ for typical states drawn from the microcanonical ensemble.","marker":"[50]"},{"why":"The Lieb-Robinson bound that controls operator spreading and yields the exponential error estimates at the core of the proof.","marker":"[53]"},{"why":"Commutator-form Lieb-Robinson bounds used in the truncation-error estimates for the feedback, measurement, and pre-measurement processes.","marker":"[54]"},{"why":"Supplies the noisy single-qubit measurement parametrized by $\\eta$ used in the exactly solved $V$-qubit example and the GHZ-triplet counterexample.","marker":"[57]"},{"why":"Provides the commutator form of the Lieb-Robinson bound for Hamiltonian dynamics that the supplemental proof invokes directly.","marker":"[62]"}],"fun_headline_variants":[],"cache_read_input_tokens":0,"weakest_assumption_plain":"The load-bearing premise is that the initial pure state is in microscopic thermal equilibrium, meaning every observable supported on a region of diameter $O(L^s)$ has the same expectation value as in the canonical ensemble at inverse temperature $\\beta$ — a property that, when it fails or hides correlations no local observable can see (as in the paper's $\\beta=0$ GHZ state), lets distant-outcome feedback extract more work than the thermal bound allows.","fun_headline_variants_meta":{"raw":{"variants":[]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":9.9e-05,"raw_usage":{"cost":9.86622e-05,"is_byok":false,"cost_details":{"upstream_inference_cost":9.86622e-05,"upstream_inference_prompt_cost":9.82982e-05,"upstream_inference_completions_cost":3.64e-07},"total_tokens":1287,"prompt_tokens":1285,"completion_tokens":2,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":256,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":0}},"tokens_in":1285,"tokens_out":2,"cost_source":"openrouter_usage","duration_ms":4057,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":0,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-13T06:40:55.893816+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":"On a programmable quantum device, prepare the $\\beta=0$ MITE state of GHZ triplets, apply the noisy single-qubit $\\sigma_3$ measurement of strength $0<\\eta<1$, and flip each triplet's spins with a $\\sigma_1$ rotation when the majority outcome is $+$: the paper predicts work $W = gL(3q-q^3)/2$ with $q=2\\eta/(1+\\eta^2)$, which exceeds the canonical-optimal value $gLq$; observing $W \\le gLq$ for any $\\eta\\in(0,1)$ would falsify the claim that distant-outcome feedback harvests hidden correlations. In the local-feedback regime, the theorem predicts $|W - W_{\\rm ens}|/V \\to 0$ in the thermodynamic limit together with a sharp onset of zero work for $\\beta > \\beta_c = g^{-1}\\operatorname{arctanh}(2\\eta/(1+\\eta^2))$, either of which is directly measurable.","supporting_citations":[{"cited_title":"Second law of thermodynamics with discrete quantum feedback control, Phys. Rev. Lett. 100, 080403","cited_arxiv_id":null,"evidence_quote":"Supplies the single-round measurement-then-feedback protocol, the definition of extracted work as the closed system's energy decrease, and the quantum-classical mutual information used in the main identity."},{"cited_title":"Thermal equilibrium of a macroscopic quantum system in a pure state, Phys. Rev. Lett.115, 100402","cited_arxiv_id":null,"evidence_quote":"Defines microscopic thermal equilibrium (MITE), the class of pure states whose local expectation values match the canonical ensemble, which is the paper's starting assumption."},{"cited_title":"Sec- ond law of thermodynamics in closed quantum many- body systems, PRX Quantum7, 043023","cited_arxiv_id":null,"evidence_quote":"The unitary-only predecessor establishing the second law for local finite-time operations, which this work extends to measurement and feedback."},{"cited_title":"and N. Zangh ` ı, Canonical typicality, Phys. Rev. Lett.96, 050403","cited_arxiv_id":null,"evidence_quote":"Canonical typicality, cited to justify the extended MITE notion with $0<s<1$ for typical states drawn from the microcanonical ensemble."},{"cited_title":"The finite group velocity of quantum spin systems, Communications in Mathemat- ical Physics28, 251","cited_arxiv_id":null,"evidence_quote":"The Lieb-Robinson bound that controls operator spreading and yields the exponential error estimates at the core of the proof."},{"cited_title":"Lieb- robinson bounds and the generation of correlations and topological quantum order, Phys. Rev. Lett.97, 050401 7","cited_arxiv_id":null,"evidence_quote":"Commutator-form Lieb-Robinson bounds used in the truncation-error estimates for the feedback, measurement, and pre-measurement processes."},{"cited_title":"Measurement-induced spectral transition, Phys. Rev. Lett.134, 010410","cited_arxiv_id":null,"evidence_quote":"Supplies the noisy single-qubit measurement parametrized by $\\eta$ used in the exactly solved $V$-qubit example and the GHZ-triplet counterexample."},{"cited_title":"Quasilocality and efficient simulation of markovian quantum dynamics, Phys. Rev. Lett.108, 230504","cited_arxiv_id":null,"evidence_quote":"Provides the commutator form of the Lieb-Robinson bound for Hamiltonian dynamics that the supplemental proof invokes directly."}],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}