{"id":"089338bd-bb4e-4c57-8a2b-7b64fbe8a074","arxiv_id":"2608.11066","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A boundary-preserving semantic-compilation theorem converts known one-way, streaming, and contextuality separations into architecture-independent coordination-width lower bounds for AI state-tracking solvers, with three worked applications including an n-qubit versus Omega(n^2)-bit stabilizer…","lead":"A new transfer theorem shows that classical lower bounds on memory and communication for one-way and streaming tasks become lower bounds on any AI state-tracking solver that reads a transcript and later answers queries. The paper's three applications suggest quantum latent states could use far less coordination than classical recurrent models for specific tasks, though the paper notes these are theoretical asymptotic separations, not practical near-term advantages.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2's transfer argument is sound; the load-bearing risk is the stabilizer application's imported Lemma 1, whose printed count is internally inconsistent and whose finite-witness reduction depends on an unverified same-author preprint.","rationale":"The reader's weakest-assumption (boundary-state completeness) is real but is largely definitional in this paper: Definition 10 and Remark 5 define Z_t to contain all stream-dependent, operationally distinguishable information and explicitly charge test-time weight updates, adapters, and fast weights. Within the stated finite-state model, a fixed parameter table cannot distinguish two histories that map to the same future-accessible state, because the table receives only that state and the current query. So I do not regard boundary-state completeness as the decisive weak point. The decisive risk is instead the imported stabilizer witness: Lemma 1 is the only source of the Omega(n^2) lower bound for the quantum-native application, it is typeset with an internally inconsistent product formula, and its key reduction is taken from an unpublished same-author preprint. If that import is wrong, the central transfer theorem remains correct but the headline quantum-native separation is unsupported. The classical-looking applications (hidden matching, Max-DiCut, Max-kSAT) are explicitly imported and their cited sources are public, so they are less risky, though the Max-kSAT source is also very recent. Because the reader already issued a CONDITIONAL verdict, and my read identifies the same need for verification without showing a definite flaw, the appropriate disposition is UNCHANGED.","tokens_in":16944,"tokens_out":25262,"duration_ms":266539,"concrete_test":"Independently re-derive Lemma 1 from Karanjai-Wallman-Bartlett (Ref. [10]) and the finite-causal-witness lemma of Ref. [3]: (a) recompute K_n using the corrected product 2^n * product_{j=1}^n (2^j+1) and verify the log equals (1/2)n^2 + (3/2 - log_2 3)n + O(1); (b) verify that every stabilizer-partitioning measurement used in the witness W_n can be presented through the semantic grammar with O(log n) parser workspace per block, rather than needing to buffer a Theta(n^2)-bit Clifford or Pauli description. If either check fails, Corollary 7's Omega(n^2) coordination lower bound is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The boundary-preserving semantic-compilation theorem itself (Theorem 2) is a clean composition argument: composing a semantic solver with the compiler gives a source-model solver whose state is the semantic state plus compiler workspace and event counter, so I do not find a substantive gap in that central transfer. The genuinely load-bearing risk is in the quantum-native headline application, Corollary 7, which inherits its entire lower bound from Lemma 1. Lemma 1 as printed states K_n >= [2n * product_{j=1}^n (2j+1)] / [5*3^{n-2}], which is not the number of pure n-qubit stabilizer states and is inconsistent with the Corollary 6 expansion giving (1/2)n^2 + (3/2 - log_2 3)n + O(1); the intended formula is clearly 2^n * product_{j=1}^n (2^j+1). So the displayed lemma is wrong as written. More substantively, Lemma 1's upgrade from the Karanjai-Wallman-Bartlett overlap bound to arbitrary exact finite-state causal online realizations is imported from Ref. [3], an unpublished same-author preprint. That upgrade is exactly the step that makes the bound architecture-independent. If the finite-witness construction in Ref. [3] is incorrect, or if its witness requires query contexts whose natural encoding (Clifford circuits or Pauli lists) has Theta(n^2) bits rather than O(log n) workspace, then the a_n = O(log n) overhead in Theorem 2 is not satisfied and Corollary 7's Omega(n^2) separation would not follow from the transfer theorem. The classical transfer theorem can survive, but the headline stabilizer separation would be unproved.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a boundary-relative resource model (B, M, D) for inference-time AI state tracking, defines finite classical and quantum latent-state generators, and proves a semantic-compilation theorem (Theorem 2) that transfers lower and upper bounds for one-way, streaming, and adaptive causal tasks to bounds on the peak coordination width of any finite-state classical solver for a semantically compiled AI interface, up to parser overhead and error. Applications are given: matched-entity QA inherits the hidden-matching separation O(log N) qubits vs Ω(√N) classical bits; continual requirements auditing inherits a Max-kSAT streaming separation (O(log^5 n log(1/δ)) qubits vs Ω(√n) classical width for approximation ratio 0.7172); and a stabilizer dialogue is claimed to have an n-qubit quantum solver while every exact finite-state classical causal solver needs Ω(n²) coordination bits. The paper is explicit that the source protocols, streaming algorithms, and stabilizer witness are imported and that no runtime or empirical advantage is claimed.","tokens_in":17272,"tokens_out":10700,"duration_ms":109046,"significance":"The central transfer principle is a genuine conceptual contribution: Theorem 2 gives a clean composition argument that converts communication, streaming, and contextuality lower bounds into architecture-independent lower bounds on AI state trackers that are allowed recurrence, scratchpads, tools, and recomputation. The resource accounting is carefully designed so that classical repairs are charged rather than assumed away, and the paper explicitly names the imported content and states finite-size caveats. The one-way and online specializations are transparent. If the stabilizer witness is made fully self-contained and the count in Lemma 1 is corrected, Corollary 7 would be an interesting exact quantum-coordination separation. The significance is currently tempered by the reliance on an unpublished same-author preprint for the headline stabilizer result and by the fact that the classical-semantics applications transfer known results rather than new lower bounds.","major_comments":[{"comment":"The displayed bound in Lemma 1, K_n ≥ [2n ∏_{j=1}^n(2j+1)]/[5·3^{n-2}], is internally inconsistent with Corollary 6. The number of pure n-qubit stabilizer states is 2^n ∏_{j=1}^n(2^j+1); the printed product ∏(2j+1) has logarithm Θ(n log n), not Θ(n^2), so the Ω(n^2) conclusion does not follow from the displayed expression. The expansion log_2(2^j+1) = j + log_2(1+2^{-j}) and the n=3 check (8·3·5·9=1080) show the intended formula. Please correct the prefactor and product in both Lemma 1 and Corollary 6, and reconcile the sentence stating that the factor is 2^j+1 with the displayed formula.","section":"Section 11, Lemma 1 and Corollary 6"},{"comment":"The load-bearing step for Corollary 7 is the upgrade from the KWB overlap bound to 'every exact finite-state classical causal online realization.' The proof states that this finite-witness construction is exactly the lemma in Ref. [3], an unpublished same-author preprint, and no construction or proof is provided here. Since this step is what makes the stabilizer lower bound architecture-independent, and since Theorem 2's compiler-overhead clause requires a_n = O(log n), the manuscript should either include a self-contained proof of the finite witness or clearly mark Corollary 7 and the corresponding row of the audit table as conditional on the correctness of Ref. [3]. It should also verify that the adaptive witness's query contexts can be specified and semantically parsed with O(log n) workspace; if a context requires Θ(n^2) bits or workspace, the transfer theorem's overhead assumption fails.","section":"Section 11, Lemma 1"},{"comment":"The semantic stabilizer dialogue is specified only by saying that the transcript 'describes Clifford updates and previous Pauli measurement outcomes using an unambiguous finite grammar.' The exactness of the compiler and the O(log n) workspace claim cannot be checked without a specification of that grammar, or of the encoding of the imported witness W_n. Since Corollary 7 is the paper's quantum-native headline application, please provide the grammar or a precise reduction, and state explicitly how the witness's measurement contexts are rendered as query prompts without retaining stream-dependent parser state.","section":"Definition 15 and Corollary 7"}],"minor_comments":[{"comment":"Please use explicit superscript notation (2^n and ∏) throughout; the current rendering '2n Qn' is ambiguous and is the source of the Lemma 1 inconsistency.","section":"Section 11"},{"comment":"The proof says 'expanding log_2(2j+1) = j + log_2(1+2^{-j})'; after the formula is corrected, this should read log_2(2^j+1), and the displayed expansion should match the corrected product.","section":"Corollary 6"},{"comment":"The distinction between a 'restricted chart simulator' and a 'KWB-compatible stabilizer simulator' would benefit from an explicit formal definition, since the later finite witness is what converts the KWB count into a general causal bound rather than a chart-covering bound.","section":"Section 2.1"},{"comment":"The claim that fixed parameters cannot distinguish two histories that map to the same boundary state is a modeling assumption about the finite-state class; it should be stated as a formal hypothesis of Theorem 2 or Definition 10 rather than appearing only in a remark.","section":"Remark 5"},{"comment":"Ref. [3] is an unpublished preprint and is the source of the most load-bearing construction; please provide its current version and date, or better, include the finite-witness proof in an appendix.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The headline stabilizer application depends on Ref. [3], another paper by the same author, which is unpublished. This is not a logical circularity, but it makes the independent value of Corollary 7 hard to assess; the best fix is to include the witness construction in the present manuscript. There is also a scope question: the paper is mainly a transfer framework plus applications of known separations, and it is honest about that. I would not reject on scope grounds, but the stabilizer application needs to be fully verifiable before the paper can support its strongest claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The part of this paper worth taking seriously is Theorem 2, the boundary-preserving semantic-compilation theorem. It is a clean composition argument: compose any classical solver for the compiled semantic task with the online encoder/decoder, and you get a valid solver for the source task with at most a_n plus O(log T) overhead on the boundary state. That is exactly the kind of architecture-independent transfer that the field needs. I checked the one-way and online specializations (Corollaries 1 and 3); they follow from the theorem and from standard hidden-matching and streaming lower bounds. The B, M, D accounting is sensible, and the paper is unusually honest about what is imported and what is new. The matched-entity QA and continual requirements-auditing applications are correctly labeled as semantic re-wrappings of existing separations, and they work as calibrations of the framework.\n\nThe soft spot is real, and it is in the quantum-native stabilizer application. Lemma 1 as printed states K_n >= 2n * product_{j=1}^n (2j+1) / (5*3^{n-2}). That is not the number of pure n-qubit stabilizer states. The intended formula is clearly 2^n * product_{j=1}^n (2^j+1), and the n=3 check and the Corollary 6 expansion only make sense with the 2^j+1 factor. So the displayed lemma is wrong as written. A reader can guess the fix, but an internal inconsistency in the central lower bound is not a minor typo; it has to be corrected before the claim is usable.\n\nMore substantively, the upgrade from the Karanjai-Wallman-Bartlett overlap bound to arbitrary exact finite-state causal online realizations is imported from Ref. [3], an unpublished same-author preprint. That upgrade is precisely the step that makes the bound architecture-independent and feeds the a_n = O(log n) overhead in Theorem 2. If the finite-witness construction in Ref. [3] is wrong, or if its witness requires contexts too large to encode in O(log n) workspace, then Corollary 7's Omega(n^2) separation does not follow from the transfer theorem. The transfer theorem can survive that; the headline stabilizer claim cannot, as currently supported.\n\nOther concerns are minor. Remark 5's boundary-state completeness assumption is stated rather than proved, but it is reasonable for finite-information baselines. The paper's explicit disclaimers about exact simulation, ideal memory, and lack of finite-size relevance are to its credit.\n\nWho should read this? People working on quantum-AI theory, communication complexity, and state-tracking. The transfer theorem is a useful organizing tool even if the stabilizer application collapses. I would send it to a serious referee, with the instruction to require a fixed Lemma 1 and either independent verification of the Ref. [3] finite-witness claim or a reframing of Corollary 7 as conditional on it.","headline":"The boundary-preserving semantic-compilation theorem is a genuinely useful transfer result, but the headline stabilizer separation is not proved as written because Lemma 1 is internally inconsistent and its key upgrade is imported from an unpublished same-author preprint.","tokens_in":17805,"tokens_out":1782,"would_cite":true,"duration_ms":115484,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","68Q17"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that quantum latent memory can require fewer coordination bits than any finite-state classical AI solver for certain state-tracking tasks, via a boundary-preserving semantic compiler.","keywords":["quantum coordination","state tracking","semantic compilation","communication complexity","streaming algorithms","contextuality","stabilizer states","latent memory"],"falsifier":"A concrete way to disprove the main separations: exhibit an exact finite-state classical causal online algorithm for the stabilizer dialogue whose peak boundary width stays below the stated formula on all inputs, or a one-pass classical streaming algorithm for Max-kSAT attaining ratio 0.7172 with o(sqrt(n)) retained bits; either would refute the corresponding theorem as stated.","tokens_in":16723,"feed_emoji":"⚛️","tokens_out":7060,"duration_ms":60595,"temperature":0.7,"pith_summary":"This paper tries to establish that some AI state-tracking tasks—where a solver must compress a growing history into a retained state and later answer a query—have a genuine quantum coordination advantage: a quantum latent state can be exponentially or quadratically smaller in bits than the classical boundary state any finite-state solver must keep. The central mechanism is a boundary-preserving semantic-compilation theorem (Theorem 2) that maps a finite one-way, streaming, or adaptive causal task into a natural 'semantic AI' interface while preserving event order and access to past input, so that classical space lower bounds and quantum-memory upper bounds transfer verbatim up to explicit compiler overhead. If correct, the theorem turns existing communication-complexity, streaming, and contextuality separations into architecture-independent lower bounds on any finite-precision recurrent AI solver—RNN, state-space model, recurrent transformer, scratchpad, or tool-using agent—provided all retained information is counted. The paper applies this to hidden-matching QA, continual Max-kSAT requirements auditing, and a stabilizer latent-state dialogue, and it explicitly notes these are memory and coordination separations, not runtime or empirical advantages at present-day scales.","feed_headline":"Quantum memory beats classical state tracking on three tasks","feed_subtitle":"Existing one-way and streaming lower bounds transfer to every finite-state classical AI solver through a boundary-preserving compiler.","key_machinery":"The load-bearing object is the boundary-preserving semantic compiler: an online translation of a source task's events into text or symbolic blocks that preserves event order, adaptive choices, and source boundaries, never re-supplies a past event unless it is charged as persistent state, and uses at most $a_n$ bits of compiler workspace. It is paired with the boundary-relative coordination accounting ($B, M, D$) and the peak online coordination width $W$, defined as the maximum number of operationally distinguishable stream-dependent states a solver can carry across an online boundary. The compiler works by composing a semantic solver with the source solver, so that the semantic solver's retained state plus $O(\\log T)$ counter bits plus compiler bits forms a valid source-state representation; the finite causal stabilizer witness from Lemma 1 supplies the concrete classical state-count lower bound in the quantum-native application.","core_discovery":"The paper's central claim is that coordination cost—bits of communication $B$, bits of persistent instance-dependent memory $M$, and local work $D$—is the right quantity in which to compare classical and quantum AI state trackers, and that known classical lower bounds survive translation into this setting. Theorem 2 shows that if a source task has a classical solver needing at least $C_{\\mathrm{cl}}$ peak boundary bits, then any classical solver of the semantically compiled task with error at most $\\epsilon+\\eta$ has peak coordination width at least $C_{\\mathrm{cl}} - a_n - O(\\log T)$, where $a_n$ is compiler workspace; symmetrically, a quantum solver with $q$ qubits gives a compiled solver with $q$ qubits plus compiler overhead. Applied to three benchmarks, the paper claims: matched-entity synopsis QA needs $O(\\log N)$ qubits and $\\Omega(\\sqrt{N})$ classical boundary bits; continual requirements auditing can be approximated to ratio $0.7172$ with $O(\\log^5 n\\log(1/\\delta))$ qubits while every classical one-pass finite-information solver needs $\\Omega(\\sqrt{n})$ coordination width; and an exact stabilizer latent-state dialogue runs on $n$ qubits while every exact finite-state classical causal online realization needs $B+M \\ge \\frac{1}{2}n^2 + \\left(\\frac{3}{2}-\\log_2 3\\right)n + O(1)$.","pith_inferences":["Editorial inference: the theorem implies that any benchmark hoping to show a quantum state-tracking advantage must enforce a one-pass boundary and charge a retained full transcript as state; otherwise a classical model can satisfy the lower bounds by spending context length, as the paper's own full-context loophole remark concedes.","Editorial inference: because a classical hidden variable can always be repaired by a recurrent state (Proposition 2), the framework points toward noncommuting latent states—rather than ordinary ambiguity—as the only plausible source of quantum coordination advantage in language-model-style tasks.","Editorial inference: the same compiler template should transfer other one-way communication separations to QA-style benchmarks, provided the benchmark's relation avoids a random-access-code obstruction by not requiring retrieval of a prespecified stored bit."],"forward_implications":["Any one-pass streaming lower bound becomes a lower bound on the peak coordination width of every finite-information recurrent AI solver, with recurrence, scratchpads, tools, and recomputation all charged to $M$, $B$, or $D$.","Matched-entity consistency QA cannot be solved classically in the one-way boundary model with fewer than $\\Omega(\\sqrt{N})$ retained bits, while an $O(\\log N)$-qubit boundary state answers every query exactly.","Continual requirements auditing—a planning-style natural-language task—inherits a quantum solver that outputs a $0.7172$-approximation using $O(\\log^5 n\\log(1/\\delta))$ qubits, versus $\\Omega(\\sqrt{n})$ coordination width for any classical one-pass solver with the same ratio.","An exact stabilizer dialogue generated by $n$ qubits of latent memory forces every exact finite-state classical causal online realization to hold at least about $\\frac{1}{2}n^2 + O(n)$ boundary bits.","The separated quantities are memory and coordination, not runtime; the paper states that no finite-size crossover or practical quantum-memory advantage is claimed at ordinary scales."],"supporting_citations":[{"why":"Supplies the boundary-relative coordination accounting (B, M, D) and the finite causal stabilizer witness that Lemma 1 restates.","marker":"[3]"},{"why":"Supplies the stabilizer partitioning bound—every set of more than 5·3^{n−2} pure n-qubit stabilizer states admits a partitioning measurement—that yields the classical state-count lower bound.","marker":"[10]"},{"why":"Supplies the hidden-matching one-way quantum protocol and Omega(sqrt(N)) classical one-way lower bound that the matched-entity QA separation imports.","marker":"[19]"},{"why":"Supplies the one-pass quantum streaming algorithm for Max-DiCut and the classical streaming lower bound at approximation ratio above 4/9.","marker":"[23]"},{"why":"Supplies the one-pass quantum streaming algorithm for Max-kSAT with ratio 0.7172 and the classical streaming lower bound for ratios above sqrt(2)/2 used in requirements auditing.","marker":"[24]"}],"fun_headline_variants":["Quantum coordination advantage proven in three AI state-tracking tasks","Semantic compiler yields quantum memory advantage on three tasks","Boundary-preserving theorem gives quantum advantage in state tracking","Quantum coordination advantage on three state-tracking tasks","Quantum memory reduces coordination cost in three state-tracking tasks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The transfer holds only if the peak number of distinguishable future-accessible boundary states captures all instance-dependent information a classical solver can use; if fixed pretrained weights could act as an uncharged lookup table keyed to the instance, the streaming lower bounds would not transfer.","fun_headline_variants_meta":{"raw":{"variants":["Quantum coordination advantage proven in three AI state-tracking tasks","Semantic compiler yields quantum memory advantage on three tasks","Boundary-preserving theorem gives quantum advantage in state tracking","Quantum coordination advantage on three state-tracking tasks","Quantum memory reduces coordination cost in three state-tracking tasks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000943,"raw_usage":{"total_tokens":4156,"prompt_tokens":1197,"completion_tokens":2959,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":813,"completion_tokens_details":{"reasoning_tokens":2881}},"tokens_in":813,"tokens_out":2959,"duration_ms":20694,"temperature":1.0,"reasoning_tokens":2881,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:55:58.178015+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete way to disprove the main separations: exhibit an exact finite-state classical causal online algorithm for the stabilizer dialogue whose peak boundary width stays below the stated formula on all inputs, or a one-pass classical streaming algorithm for Max-kSAT attaining ratio 0.7172 with o(sqrt(n)) retained bits; either would refute the corresponding theorem as stated.","supporting_citations":[{"cited_title":"Genuine Global Kochen-Specker Contextuality as Classical Coordination Cost","cited_arxiv_id":"2606.23577","evidence_quote":"Supplies the boundary-relative coordination accounting (B, M, D) and the finite causal stabilizer witness that Lemma 1 restates."},{"cited_title":"Exponential Quantum Space Advantage for Approximating Max-$k$SAT in the Streaming Setting","cited_arxiv_id":"2606.05366","evidence_quote":"Supplies the one-pass quantum streaming algorithm for Max-kSAT with ratio 0.7172 and the classical streaming lower bound for ratios above sqrt(2)/2 used in requirements auditing."}],"review_version":1}