{"id":"f5ad46cb-95fa-4c26-894c-a3997167d2bd","arxiv_id":"2501.15794","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Stabilizer operations cannot clone magic in any finite dimension, and the paper claims unconstrained broadcasting of magic is also limited, but that claim is not rigorously proven.","lead":"This paper proves new no-go theorems about copying or broadcasting 'magic', the quantum resource that makes states hard to simulate classically. It shows stabilizer operations cannot duplicate magic in any finite dimension, and it claims that even unconstrained operations face limits, though that part rests on an unjustified assumption.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proofs of Theorems 2 and 3 rely on the unsupported inference that an orthogonal input basis forces the machine states |µ_i⟩ to be orthogonal; unitarity only implies orthogonality of the full products, so the convex-mixture reduced states used in both proofs are not established.","rationale":"The reader identified precisely this gap, and my own re-derivation of Eq. (6) confirms it is load-bearing. There is a second, smaller gap in Theorem 2's final step (magic depending only on moduli versus phases), but it is moot while the convex-mixture premise is unsupported. Because the abstract claims fundamental limits for unrestricted operations, not just for protocols with orthogonal machine registers, the proofs do not support the headline result. The numerical section is suggestive but not central to the no-go claim. I therefore agree with rejection; a revision could either prove the missing inference, add machine orthogonality as an explicit assumption and restrict the claims accordingly, or supply an alternative proof that does not need machine orthogonality.","tokens_in":16324,"tokens_out":15039,"duration_ms":140126,"concrete_test":"Set H_S = H_A = H_M = ℂ². Choose |ψ̃_0⟩ = (|00⟩ + |11⟩)/√2 and |ψ̃_1⟩ = (|00⟩ − |11⟩)/√2 in H_S ⊗ H_A; these are orthogonal, but Tr_S(|ψ̃_0⟩⟨ψ̃_1|) = (|0⟩⟨0| − |1⟩⟨1|)/2 ≠ 0. Choose non-orthogonal machine states |µ_0⟩ = |0⟩ and |µ_1⟩ = cos θ |0⟩ + sin θ |1⟩ with 0 < θ < π/2. The global outputs |ψ̃_0⟩|µ_0⟩ and |ψ̃_1⟩|µ_1⟩ are orthogonal because ⟨ψ̃_0|ψ̃_1⟩ = 0, so the map from two orthogonal inputs to these outputs is an isometry and extends to a unitary. For |Φ⟩ = (|ψ_0⟩ + |ψ_1⟩)/√2, compute ρ_A = Tr_{S,M}[U(|Φ⟩⟨Φ| ⊗ |0⟩⟨0|_A ⊗ |µ_0⟩⟨µ_0|)U†]. The off-diagonal contribution is (1/2)⟨µ_1|µ_0⟩ Tr_S(|ψ̃_0⟩⟨ψ̃_1|) = (cos θ / 4)(|0⟩⟨0| − |1⟩⟨1|), which is nonzero, so ρ_A ≠ (1/2)(ρ̃_0^A + ρ̃_1^A). This is a concrete, valid unitary exhibiting non-orthogonal machine states and violating the convex-mixture step used in Theorem 2; the analogous two-term computation applies to Eq. (6).","verdict_should_be":"REJECT","load_bearing_attack":"In Section III, after Eq. (5), the paper asserts that ⟨ψ_i|ψ_j⟩=0 'also implies ⟨µ_i|µ_j⟩=0, since the final states |ψ̃_i⟩_SA and |ψ̃_j⟩_SA may not necessarily be orthogonal for i≠j.' This is a logical error. Unitarity of U gives ⟨ψ̃_i|ψ̃_j⟩_{SA} ⟨µ_i|µ_j⟩_M = 0 for every i≠j; each zero product only forces at least one factor to vanish. If the SA factors are orthogonal, the machine states can be arbitrary, including non-orthogonal. The parenthetical does not rule this out. The proof of Theorem 2 then uses machine orthogonality to write ρ̃_Φ^{S(A)} = Σ_i |α_i|² ρ̃_i^{S(A)}: tracing out M removes cross terms ⟨µ_j|µ_i⟩, and then tracing S or A leaves the diagonal mixture. Without machine orthogonality, cross terms of the form α_i α_j^* ⟨µ_j|µ_i⟩ Tr_S(|ψ̃_i⟩⟨ψ̃_j|) survive in ρ_A (and similarly in ρ_S), so the reduced state is not the convex combination used in the proof. Theorem 3's Eq. (6) makes the same unsupported leap: tracing out S and M yields extra terms unless machine states are orthogonal (or some other condition is proved). Since these convex decompositions are what allow the convexity bound R(ρ̃_ϕ^A) ≤ |α|² R(ρ̃_ψ^A) + |β|² R(ρ̃_ψ'^A) and the coefficient-dependence argument, the central no-go claims are not proven as written. The results may be true, but the unrestricted-operation case requires either a proof that machine orthogonality can be assumed without loss of generality or a different no-go argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies broadcasting and cloning of non-stabilizerness (\"magic\") in quantum states. Theorem 1 claims that stabilizer operations cannot clone nonvanishing magic for all states in any finite dimension, using additivity and monotonicity of the extended stabilizer Rényi entropy after a Stinespring dilation. Theorems 2 and 3 claim that even with unrestricted operations there is no universal magic-broadcasting transformation for qudits, and that a transformation designed to broadcast the magic of specific orthogonal qubit states cannot broadcast states with higher magic. The paper then analyzes the Wootters-Zurek and Bužek-Hillery cloning machines, derives conditions for partial or perfect magic broadcasting, and presents numerical results comparing the magic-generating power of magic-broadcasting and state-broadcasting unitaries.","tokens_in":16658,"tokens_out":6191,"duration_ms":61510,"significance":"If the unrestricted-operation no-go results were valid, they would constitute a substantial extension of no-broadcasting ideas to the resource theory of non-stabilizerness, with implications for magic distillation and fault-tolerant quantum computation. The paper also contains a useful self-contained derivation of the single-qubit robustness-of-magic witness in Appendix A, and Theorem 1 appears sound: its proof uses Clifford invariance, additivity, and monotonicity of the Rényi measure without relying on the questionable orthogonality inference. However, the proofs of Theorems 2 and 3 contain a load-bearing logical gap concerning orthogonality of machine states. Until this gap is repaired, the central unrestricted-operation claims are not established as written.","major_comments":[{"comment":"The inference that orthogonal reference input states imply orthogonal machine states is invalid. Unitarity of U gives ⟨ψ̃_i|ψ̃_j⟩_{SA} ⟨µ_i|µ_j⟩_M = 0 for i ≠ j, and each zero product only forces at least one factor to vanish. The parenthetical \"may not necessarily be orthogonal\" does not rule out the case where the SA factors are orthogonal and the machine states are non-orthogonal. This matters because the proof of Theorem 2 uses machine orthogonality to assert ρ̃_Φ^{S(A)} = Σ_i |α_i|² ρ̃_i^{S(A)}; without machine orthogonality, cross terms proportional to ⟨µ_j|µ_i⟩ survive after tracing, so the convex-decomposition step is not justified and the subsequent coefficient-dependence argument collapses.","section":"Section III, after Eq. (5)"},{"comment":"The expression ρ̃_ϕ^A = |α|² ρ̃_ψ^A + |β|² ρ̃_ψ'^A after tracing out S and M presupposes that the two product outputs |ψ̃⟩_SA |µ_1⟩_M and |ψ̃'⟩_SA |µ_2⟩_M are orthogonal in at least one factor, so that no cross term survives. The proof invokes machine-state orthogonality, which is not established. Consequently, the convexity bound R(ρ̃_ϕ^A) ≤ |α|² R(ρ̃_ψ^A) + |β|² R(ρ̃_ψ'^A) and the resulting conclusion that states with higher magic cannot be broadcast are not proven. A valid proof would need either a demonstration that machine states can be taken orthogonal without loss of generality or a different no-go argument.","section":"Section III, Eq. (6) and Theorem 3 proof"},{"comment":"The Observation that there exists no universal unitary broadcasting the magic of arbitrary non-stabilizer qubit states is justified only by the same convex-mixture assumption used in Theorem 3. Since that assumption is unsupported as described above, the Observation is likewise not established by the arguments given in the manuscript.","section":"Section III.A, Observation"}],"minor_comments":[{"comment":"The block of expressions after \"2 α = ...\" appears garbled: the quantities α and β are not clearly defined, and the displayed formula mixes undefined notation with what seems to be an attempted expression for a state or a parameter. Please rewrite this passage.","section":"Section IV.B.2, Eq. (11)"},{"comment":"Condition (2) reads |⟨ψ|ρ̃_S(A)|ψ⟩|² ≤ 1 - ε, but the text describes it as corresponding to \"perfect state broadcasting.\" For state broadcasting one would expect the fidelity to be close to 1, not bounded away from 1; please clarify whether the inequality direction is a typographical error and explain the intended numerical protocol.","section":"Section IV.A, conditions (1) and (2)"},{"comment":"There is a minor typo in Section III where \"the natural question to as is\" should be \"the natural question to ask is,\" and the phrase \"Schwartz inequality\" should be \"Cauchy-Schwarz inequality.\"","section":"Throughout"},{"comment":"References [54] and [95] are the same paper (Zhang, Feng, and Luo, Phys. Rev. A 110, 012462 (2024)); one duplicate citation should be removed.","section":"References"}],"recommendation":"reject","confidential_remarks":"The central unrestricted-operation claims rest on an invalid orthogonality inference, so the manuscript's main novelty is not established as written. I see no sign of misconduct; Theorem 1 and the numerical/cloning-machine analysis may be salvageable in a future revision, but the current version does not meet the bar for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper asks the right question—can unrestricted operations broadcast magic?—and proves one thing cleanly, but the two headline no-go theorems ride on a false inference, so the abstract is ahead of the math.\n\nWhat's genuinely good: Theorem 1 is a correct, self-contained proof that stabilizer operations cannot clone magic in any finite dimension, using additivity and monotonicity of the extended stabilizer Rényi entropy. The appendix gives a nice independent proof that single-qubit robustness of magic equals max{1, sum |m_j|}. The numerical comparison between magic broadcasting and state broadcasting unitaries is suggestive, and the observation that magic-broadcasting unitaries have higher magic-generating power is worth taking seriously. The Wootters–Zurek and Buzek–Hillery analyses provide concrete conditions and ratios, and the distinction between broadcasting states and broadcasting magic is well illustrated.\n\nThe soft spot is load-bearing. After Eq. (5), the paper claims that orthogonal reference states imply orthogonal machine states. Unitarity only gives ⟨ψ̃_i|ψ̃_j⟩ ⟨µ_i|µ_j⟩ = 0. The SA factors might be orthogonal for every pair; the machine states could be anything. The parenthetical that the final SA states 'may not necessarily be orthogonal' doesn't rule that out. Both Theorem 2 and Theorem 3 then write the reduced outputs as convex mixtures that require tracing out orthogonal machine states. Without machine orthogonality, cross terms survive and the convexity bounds don't go through. So the central no-go claims are not proven as written.\n\nThere's also an Observation on universal qubit broadcasting stated without proof, and the numerics lack code and error bars, but those are minor next to the gap.\n\nThis is a paper for people working on magic resource theory and no-broadcasting. The question is timely and the first theorem plus the numerics give it enough substance to justify referee time. But I'd want the authors to either prove that machine orthogonality can be assumed without loss of generality or rewrite Theorems 2 and 3 with a different argument. If that gap is closed, it becomes a solid paper.\n\nMy recommendation: send it to peer review, but the referee should focus on the machine-orthogonality step.\n\nYours,","headline":"A correct first theorem and a suggestive numerics section, but the two headline no-broadcasting results for unrestricted operations rest on a false machine-orthogonality inference.","tokens_in":17235,"tokens_out":3304,"would_cite":false,"duration_ms":28325,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Quantum 'magic' cannot be broadcast universally: even unrestricted operations fail to clone non-stabilizerness beyond a fixed reference level.","keywords":["non-stabilizerness","magic broadcasting","quantum no-cloning","stabilizer operations","robustness of magic","state-dependent cloning","quantum resource theory","qudit systems"],"falsifier":"Search numerically over two-qubit unitary parameters for any single-qubit pure state whose input robustness of magic is larger than the reference-state robustness but whose output auxiliary copy, after the designed broadcasting unitary, still has the same magic as the input; finding one would refute Theorem 3, and measuring the inner product of machine states for two orthogonal reference inputs would check whether the assumed orthogonal-machine-state structure actually holds.","tokens_in":16072,"feed_emoji":"⚛️","tokens_out":14680,"duration_ms":133731,"temperature":0.7,"pith_summary":"This paper asks whether 'magic'—the non-stabilizerness that makes quantum circuits hard to simulate classically—can be cloned or broadcast across many subsystems. It first proves that stabilizer operations, the free operations that cannot create the resource, cannot clone magic in any finite dimension, even when the auxiliary copy is only required to become non-stabilizer while the original keeps its magic. It then proves that the impossibility persists for unrestricted operations in a qualified sense: no fixed unitary can broadcast the magic of all qudit states, and a qubit unitary built to broadcast the magic of given reference states cannot broadcast any state with higher magic than those references. The paper closes with explicit conditions under which specific cloning machines broadcast magic perfectly, and numerical evidence that magic broadcasting demands more magic-generating power than ordinary state broadcasting.","feed_headline":"Quantum magic cannot be freely broadcast","feed_subtitle":"Even unrestricted operations can't clone magic beyond the reference level, so magic distillation remains necessary.","key_machinery":"The load-bearing objects are the magic monotones used for each part of the argument: the order-2 stabilizer entropy and its extended version, which is additive under tensor products and monotone under partial trace, and the single-qubit robustness of magic $R(\\rho)=\\max\\{1,\\sum_j |m_j|\\}$. The unifying mechanism is the broadcasting unitary $U(|\\psi_i\\rangle_S |0\\rangle_A |\\mu_0\\rangle_M)=|\\tilde{\\psi}_i\\rangle_{SA} |\\mu_i\\rangle_M$, together with the assumption that the machine states $|\\mu_i\\rangle_M$ are mutually orthogonal for orthogonal reference inputs. Orthogonality converts the output reduced states into convex mixtures of fixed reference outputs; convexity then caps the magic of any broadcast copy at the reference magic, which is exactly what blocks universal and above-reference broadcasting.","core_discovery":"The central claim is that non-stabilizerness is not freely replicable, even when no restriction is placed on the operations used. Theorem 1 rules out stabilizer cloning in all finite dimensions by showing that any successful clone would require the system's magic to fall below its initial value, contradicting the cloning condition. Theorem 2 rules out a universal qudit broadcasting unitary: because the output auxiliary state depends only on the squared amplitudes of the input superposition, it cannot reproduce the phase-dependent magic of every input. Theorem 3 sharpens this for qubits: convexity of the robustness of magic forces the broadcast output to have no more magic than the reference states the unitary was designed for, so an operation that broadcasts maximal-magic states still fails on generic lower-magic states. The same methods yield geometric conditions for perfect broadcasting and a characterization of how the original copying machine and the state-independent cloner broadcast magic.","pith_inferences":["The phase-versus-amplitude mismatch behind Theorem 2 suggests a broader pattern: any resource measure that depends on complex phases of superpositions cannot be universally broadcast when outputs see only squared amplitudes, so the no-go may extend beyond magic to other coefficient-sensitive resources.","The geometric polytope condition for perfect qubit broadcasting may generalize to higher-dimensional stabilizer polytopes, giving explicit construction recipes for qutrit and qudit magic broadcasters; the paper does not work out this generalisation.","The numerical gap in magic-generating power between magic broadcasting and state broadcasting could be turned into a proof that magic-broadcasting circuits necessarily require more non-stabilizer gates than state-broadcasting ones, sharpening the resource-cost comparison with distillation.","If the no-go theorems are correct, the useful figure of merit for a magic-broadcasting protocol is not output fidelity but the pair of reference magic and input magic; protocols should be designed by first fixing the target magic level and then searching for unitaries within that level."],"forward_implications":["Stabilizer circuits cannot be used to bypass magic distillation in any finite dimension: a blank register plus free operations can never end with nonvanishing magic while preserving the input magic.","Every unrestricted magic broadcaster is state-specific and bounded by its reference magic, so practical protocols must match the target magic content to the machine design rather than aiming for a universal cloner.","The original copying machine, when set to a non-stabilizer reference state, can perfectly broadcast the magic of whole families of lower-magic states, not just the reference states themselves.","The state-independent cloner broadcasts maximal-magic states at a fixed 2/3 magic ratio, while low-magic states can be broadcast at ratios approaching 0.9 by tuning the machine parameters.","Perfect magic broadcasting and perfect state broadcasting are different tasks: magic can be copied with fidelity as low as 3 times 10^{-3}, and magic-broadcasting unitaries have higher average magic-generating power than state-broadcasting unitaries."],"supporting_citations":[{"why":"Supplies the prime-dimensional stabilizer no-cloning result that Theorem 1 extends to arbitrary finite dimensions.","marker":"[54]"},{"why":"Defines the extended order-2 stabilizer entropy whose additivity and monotonicity under partial trace carry the proof of Theorem 1.","marker":"[69]"},{"why":"Defines the stabilizer entropy and the magic-generating power used to compare magic broadcasting with state broadcasting.","marker":"[64]"},{"why":"Introduces the original cloning transformation used in the unrestricted broadcasting setup and in the copying-machine analysis.","marker":"[78]"},{"why":"Establishes the no-broadcasting definition and product-output criterion that the paper recovers when the final state is product in the system-auxiliary split.","marker":"[44]"},{"why":"Defines the robustness of magic, the single-qubit monotone used in Theorem 3 and in the perfect-broadcasting conditions.","marker":"[79]"},{"why":"Provides the state-independent cloner whose free parameters are tuned to broadcast magic and whose 2/3 scaling is used as a benchmark.","marker":"[62]"}],"fun_headline_variants":["Magic can't be broadcast: even unlimited ops fail","Broadcasting magic? Even unrestricted ops can't","No universal magic broadcaster, even with full power","State-dependent cloners weaken magic: broadcast limits","Magic broadcast fails for generic lower-magic states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The weakest load-bearing step is the claim that orthogonal reference inputs force the machine states to be mutually orthogonal for every pair; without that, the output of the broadcasting unitary is not a convex mixture of the reference outputs and the proof does not go through.","fun_headline_variants_meta":{"raw":{"variants":["Magic can't be broadcast: even unlimited ops fail","Broadcasting magic? Even unrestricted ops can't","No universal magic broadcaster, even with full power","State-dependent cloners weaken magic: broadcast limits","Magic broadcast fails for generic lower-magic states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000707,"raw_usage":{"total_tokens":3155,"prompt_tokens":882,"completion_tokens":2273,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":498,"completion_tokens_details":{"reasoning_tokens":2200}},"tokens_in":498,"tokens_out":2273,"duration_ms":13885,"temperature":1.0,"reasoning_tokens":2200,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:56:38.223630+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search numerically over two-qubit unitary parameters for any single-qubit pure state whose input robustness of magic is larger than the reference-state robustness but whose output auxiliary copy, after the designed broadcasting unitary, still has the same magic as the input; finding one would refute Theorem 3, and measuring the inner product of machine states for two orthogonal reference inputs would check whether the assumed orthogonal-machine-state structure actually holds.","supporting_citations":[{"cited_title":"Barnum, J","cited_arxiv_id":null,"evidence_quote":"Supplies the prime-dimensional stabilizer no-cloning result that Theorem 1 extends to arbitrary finite dimensions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the extended order-2 stabilizer entropy whose additivity and monotonicity under partial trace carry the proof of Theorem 1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the no-broadcasting definition and product-output criterion that the paper recovers when the final state is product in the system-auxiliary split."},{"cited_title":"Grassl, A","cited_arxiv_id":null,"evidence_quote":"Defines the robustness of magic, the single-qubit monotone used in Theorem 3 and in the perfect-broadcasting conditions."},{"cited_title":"Patel, S","cited_arxiv_id":null,"evidence_quote":"Provides the state-independent cloner whose free parameters are tuned to broadcast magic and whose 2/3 scaling is used as a benchmark."}],"review_version":1}