{"id":"8ef54150-858a-4bea-ad7d-06e3e4a1b049","arxiv_id":"2506.16527","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Free energy unifies physical time and space complexity, and error-correction scaling makes black hole quantum computers physically intractable.","lead":"This paper defines physical analogues of computational complexity, measuring computation by energy-time product, negentropy, and free energy used. It applies these measures to biology and black holes, arguing that free energy is the fundamental 'currency of intelligence' and that error correction makes black hole quantum computers impractical.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Black-hole free-energy claim uses coarse-grained entropy while memory-space definition requires fine-grained entropy; F=Mc²/2 and O(M²) memory cannot hold for the same state.","rationale":"The reader's identified weakest point—that error entropy might be radiated away continuously—does not in the end rescue the black-hole intractability claim: for an evaporating Schwarzschild hole, the entropy flux is dS/dt ~ 1/M while the op rate is ~M, so the rate bound on ε is the same 1/M² scaling as the integrated bound. The paper's conclusion is therefore more robust than the reader's concern suggests. The genuinely load-bearing weak point is instead the dual entropy convention in §3.1/§6.1: the memory capacity of a programmed black hole requires fine-grained entropy near zero, while the F=Mc²/2 calculation uses the coarse-grained Bekenstein–Hawking entropy. These cannot both be used for the same state. This is an internal inconsistency, not a disagreement with external consensus, and it directly affects the headline black-hole free-energy claim. It does not, however, destroy the broader physical-complexity framework; the black-hole section needs a consistent entropy bookkeeping and a revised statement of what free energy is available. Thus the reader's CONDITIONAL verdict remains appropriate.","tokens_in":8410,"tokens_out":19061,"duration_ms":203080,"concrete_test":"Take a Schwarzschild black hole of mass M as a programmed computer in a pure state. Compute both quantities using the paper's definitions: (i) Φ-SPACE = S_max − S_fine = 4πM²/m_P² − 0; (ii) F = Mc² − T S_fine = Mc². Then repeat with S = S_BH = 4πM²/m_P²: Φ-SPACE = 0 and F = Mc²/2. Show that no single choice of S yields both the paper's memory capacity and its half-free-energy result. This can be done in a few lines and settles the inconsistency.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing flaw is an inconsistent use of entropy in the black-hole analysis. In §3.1, Φ-SPACE is defined as S_max − S, where S is the actual entropy of the computing system; a programmed black hole must have low fine-grained entropy to have memory capacity O(M²/m_P²). In §6.1, however, the free energy is computed as F = Mc² − T S_BH = Mc²/2, using the Bekenstein–Hawking entropy S_BH as if the hole were in the maximum-entropy thermal state. These two choices cannot both hold for the same state. If S = S_BH, then Φ-SPACE = 0 and the hole has no memory; if S ≈ 0 (pure programmed state), then F ≈ Mc², not Mc²/2. Thus the headline 'exactly half of a black hole's energy is free energy' and the assignment of O(M²) bits of computational memory rely on incompatible entropy bookkeeping. The same tension appears in the claim FREE-Φ = T(S_eq−S(ρ))/πħ, which is dimensionally wrong (energy divided by action) and, for a thermal black hole, evaluates to zero, contradicting F=Mc²/2.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes physical analogues of computational time and space: Φ-TIME = 2Et/πħ (the number of operations permitted by the Margolus-Levitin bound) and Φ-SPACE = Smax−S (negentropy). It introduces FREE-Φ as the free energy used during computation, argues that error correction forces spatial physical complexity to grow linearly with temporal physical complexity, and applies these measures to black holes, concluding that exactly half of a black hole's energy is free energy but that error correction makes black hole quantum computers physically intractable because total lifetime operations scale as M^4/m_P^4 while storage capacity scales only as M^2/m_P^2. The paper also estimates biological information-processing rates and discusses implications for AI development.","tokens_in":8618,"tokens_out":7743,"duration_ms":81580,"significance":"The manuscript draws together well-established physical bounds (Margolus-Levitin, Bekenstein-Hawking) into a resource-theoretic perspective on computation, and the idea of free energy as a unified computational resource is potentially useful. If the black-hole analysis were sound, it would give a striking no-go result. However, the central FREE-Φ definition contains a dimensional inconsistency, the black-hole free-energy and memory-capacity claims rely on incompatible entropy assignments to the same state, and the intractability argument conflates lifetime accumulation with instantaneous capacity. These issues are load-bearing for the paper's main claims, so the paper cannot be accepted in its current form; the core ideas are likely salvageable with a careful revision that fixes the entropy bookkeeping and supplies the missing justifications.","major_comments":[{"comment":"The displayed relation FREE-Φ = T(S_eq−S(ρ))/πħ is dimensionally inconsistent. With k_B = 1, S_eq and S(ρ) are dimensionless, T has dimensions of energy, and ħ has dimensions of energy×time, so the right-hand side has dimensions of inverse time, not energy. The correct relation between free energy and relative entropy is F(ρ)−F_eq = T(S_eq−S(ρ)); if instead FREE-Φ is intended to count operations per unit time, that should be derived from the Margolus-Levitin bound (e.g., (2/πħ)Ft) and labeled as a rate. Since FREE-Φ is advertised as representing the amount of free energy used, the formula as written would make FREE-Φ vanish for a thermal equilibrium state, which conflicts with the F = Mc^2/2 claim for a black hole in Section 6.1.","section":"Section 4, FREE-Φ definition"},{"comment":"The black-hole free-energy claim uses an entropy assignment that is incompatible with the memory-capacity claim. Section 6.1 states F = Mc^2 − TS = Mc^2/2, using S = S_BH, the Bekenstein-Hawking entropy, i.e., the maximum-entropy thermal state. But Section 3.1 defines usable memory space as Smax − S, so a state with S = S_BH has zero Φ-SPACE and cannot provide the O(M^2/m_P^2) bits of memory assumed for a programmed black hole. Conversely, if the black hole is programmed into a low-entropy state to have memory, its free energy is F ≈ Mc^2, not Mc^2/2. The paper must specify a single entropy assignment for the programmed state; the 'exactly half' claim and the O(M^2) memory claim cannot both hold for the same state.","section":"Sections 6.1 and 3.1"},{"comment":"The error-correction intractability argument compares total lifetime operations O(M^4/m_P^4) to the black hole's instantaneous entropy capacity O(M^2/m_P^2), concluding that the error rate must satisfy ε log(1/ε) ≲ m_P^2/M^2. This treats the hole's maximum-entropy capacity as a lifetime budget for error entropy and does not justify why error entropy cannot be expelled by Hawking radiation, which is the hole's normal entropy-disposal channel. The relevant constraint may instead be a rate constraint, comparing error-entropy production rate (∼ ε M in Planck units) with the Hawking entropy-emission rate (∼ 1/M in Planck units); such a rate-based derivation may yield a similar bound on ε, but it is not given in the manuscript. As written, the argument is incomplete and needs to be repaired.","section":"Section 6.2"},{"comment":"The quantitative claim that a human performs on the order of 10^20–10^22 bio-ops per second, comparable to global electronic computing, is supported only by reference [23], an unpublished self-citation. Because this number drives the paper's conclusions about biological efficiency, the 'currency of intelligence,' and the implications for AI, the authors should either provide a derivation in the present paper, cite a published source, or clearly label the figure as an order-of-magnitude estimate.","section":"Sections 3.1 and 7"}],"minor_comments":[{"comment":"The sentence 'the total number of bits of memory available to a system with entropy S is no greater than S−S_max' has the sign reversed; it should be S_max − S, consistent with the definition of Φ-SPACE.","section":"Section 3.1"},{"comment":"In the sentence 'The lifetime of a black hole with mass M is M = 5120πG^2M^3/ħc^4', the left-hand side should be t_M (or τ), not M, since M is the mass.","section":"Section 6.1"},{"comment":"The table describes Φ-TIME as 'accumulated quantum phase,' but 2Et/πħ is a dimensionless count of operations, not a phase; a single Margolus-Levitin-saturating bit flip accumulates phase EΔt/ħ = π/2. The wording is misleading.","section":"Table in Section 1"}],"recommendation":"major_revision","confidential_remarks":"The paper is engaging and clearly written, but it sits between a perspective/review and a new research result; the black-hole section would benefit from a more careful state-by-state bookkeeping. The reliance on an unpublished companion paper ([23]) for a headline biological number should be resolved before publication. The AI implications in Section 5 are largely speculative and should be framed as such."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is a synthesis of Margolus–Levitin, Landauer, and black-hole thermodynamics, dressed up with new names (Φ-TIME, Φ-SPACE, FREE-Φ). The first two are just the standard energy–time product and negentropy; FREE-Φ is a relabeling of free energy, and its defining equation, FREE-Φ = T(S_eq − S(ρ))/πħ, is dimensionally wrong — it has units of frequency, not energy. That’s not a harmless typo; it’s the definition of their new measure.\n\nWhat is genuinely useful: the point that error correction forces physical space complexity to grow linearly with time is correct and worth stating clearly, and the biological-vs-electronic ops/sec comparison is a nice back-of-the-envelope. But the bio-ops figure leans on an unpublished self-citation, which weakens that section.\n\nThe load-bearing problem is in Section 6. The black hole’s free energy is computed as F = Mc² − T S_BH = Mc²/2, treating the hole as if it were in its maximum-entropy thermal state. But the memory capacity they assign, O(M²/m_P²), is exactly that maximum entropy. Under their own Φ-SPACE = S_max − S, a thermal hole has zero memory; a programmed pure hole has O(M²) memory but F ≈ Mc², not Mc²/2. You cannot have both exactly-half free energy and O(M²) programmable bits from the same state. The intractability argument then compares total lifetime ops (∝M⁴) to instantaneous entropy capacity (∝M²), but if Hawking radiation can carry away error entropy continuously, the relevant bound is a disposal rate, not a lifetime total. The paper doesn’t justify the total-ops comparison.\n\nSo the central new claim — that black-hole quantum computers are intractable because error correction demands more negentropy than the hole has — fails on the paper’s own definitions. Sections 1–5 are a readable review of known results, but the black-hole application should not be taken at face value.\n\nThis deserves a serious referee only because the topic and author will attract attention and a referee can force the entropy bookkeeping to be cleaned up; as it stands, the conclusion doesn’t hold. I wouldn’t cite it.","headline":"The black-hole section is built on an inconsistent use of entropy that sinks the main new claim; the rest is a competent but mostly derivative summary of known physics.","tokens_in":9146,"tokens_out":3798,"would_cite":false,"duration_ms":37671,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Free energy, not logic or memory alone, is the physical currency of computation—and by that measure black hole quantum computers are physically intractable.","keywords":["physical complexity","free energy","negentropy","error correction","black hole computation","quantum speed limit","thermodynamics of erasure","thermodynamic depth"],"falsifier":"Track the entropy budget in a unitary model of black hole evaporation with a radiation bath: if the bath can absorb $\\epsilon\\log(1/\\epsilon)$ bits per operation at a rate no larger than the hole's entropy-emission rate while the hole remains programmable, then the per-operation error rate need not be as low as $m_P^2/M^2$. A concrete model exhibiting this would falsify the paper's intractability claim.","tokens_in":8189,"feed_emoji":"🕳️","tokens_out":13311,"duration_ms":116211,"temperature":0.7,"pith_summary":"This paper tries to establish that the resources bounding all computation are thermodynamic, not merely logical: temporal complexity becomes an energy–time product, $\\Phi\\text{-TIME}=2Et/\\pi\\hbar$, spatial complexity becomes negentropy, $\\Phi\\text{-SPACE}=S_{\\max}-S$, and once error correction is included both grow in proportion to the free energy consumed, which the paper calls FREE-$\\Phi$. A sympathetic reader would take the central assertion to be that free energy is the single physical resource behind speed, memory, and error correction. Applied to black holes, this yields a result: exactly half of a black hole's energy is free energy available for computation, yet its total operations over its lifetime, $\\sim M^4/m_P^4$, dwarf its information capacity, $\\sim M^2/m_P^2$, so the entropy generated by errors cannot be accommodated. The paper concludes that black hole quantum computers are in principle powerful but physically intractable to program and correct.","feed_headline":"Free energy, not logic, sets the true cost of computing","feed_subtitle":"Counting error entropy shows black hole computers cannot store their own garbage.","key_machinery":"The central object is free physical complexity, FREE-$\\Phi$, defined as the free energy used by a system over a computation. The argument runs through two identifications: $\\Phi\\text{-TIME}=2Et/\\pi\\hbar$ comes from the quantum speed limit on flipping a bit with energy $E$ in time $t$, and $\\Phi\\text{-SPACE}=S_{\\max}-S$ identifies available memory with negentropy, also called thermodynamic depth. The error-correction step is the mechanism that forces $\\Phi\\text{-SPACE}\\propto\\Phi\\text{-TIME}$: each operation at error rate $\\epsilon$ injects $\\epsilon\\log(1/\\epsilon)$ entropy, which must be pumped out at free-energy cost, so the physical resources of a noisy computation are governed by free energy. For black holes, the same machinery yields the lifetime operation count and the entropy-capacity bound that make error correction intractable.","core_discovery":"On the paper's own terms, physical computational complexity has three measures: temporal complexity $\\Phi\\text{-TIME}$ is the minimum energy–time product $2Et/\\pi\\hbar$ needed to perform a computation; spatial complexity $\\Phi\\text{-SPACE}$ is the negentropy $S_{\\max}-S$ available as clean memory; and free physical complexity FREE-$\\Phi$ is the free energy consumed, which combines the two. The load-bearing step is error correction: a bit-flip probability $\\epsilon$ injects about $\\epsilon\\log(1/\\epsilon)$ bits of entropy per operation, and the thermodynamic cost of erasing that entropy forces it out at free-energy cost $k_BT\\,\\epsilon\\log(1/\\epsilon)$, so physical memory requirements grow linearly with computation length no matter how small the error rate. For a black hole of mass $M$, the free energy is $F=Mc^2/2$, exactly half its energy, and the total number of operations over its lifetime is $\\sim M^4/m_P^4$, while its entropy capacity is $\\sim M^2/m_P^2$. Keeping the accumulated error entropy within that capacity forces a per-operation error rate below $\\sim m_P^2/M^2$; the paper concludes that the physical spatial complexity of black hole error correction exceeds the black hole's own spatial complexity, making black hole quantum computers physically intractable.","pith_inferences":["The paper's intractability conclusion uses a lifetime-total entropy bound. If the radiation emitted by a black hole can dispose of error entropy continuously, as it disposes of other entropy, the relevant constraint is the disposal rate rather than the lifetime total, and the negative conclusion could weaken.","A testable extension is to measure how error entropy is flushed in near-term quantum error-correcting processors: if the spatial complexity cost can be paid continuously by the environment rather than stored in the computer, the linear growth $\\Phi\\text{-SPACE}\\propto\\Phi\\text{-TIME}$ may be avoidable in open systems.","FREE-$\\Phi$ offers a single number for comparing radically different substrates, such as silicon, neurons, chemical reaction networks, and black holes; it could serve as a benchmark for artificial intelligence by measuring free energy consumed per reliable bit-operation.","A unitary evaporation model with a radiation bath could be checked for whether the bath absorbs $\\epsilon\\log(1/\\epsilon)$ per operation; if it does, the error-rate constraint becomes a rate condition rather than a total-capacity condition, directly testing the paper's central negative result."],"forward_implications":["Any physical computer running at finite error rate must consume fresh negentropy at a rate proportional to its number of operations, so fault tolerance is a thermodynamic requirement rather than a purely logical one.","Since $\\Phi\\text{-SPACE}$ and $\\Phi\\text{-TIME}$ are both proportional to FREE-$\\Phi$, complexity classes gain a physical reading: polynomial versus exponential logical resource gaps correspond to gaps in free energy consumed.","Black holes can perform at most $\\sim M^4/m_P^4$ operations over their lifetime while storing at most $\\sim M^2/m_P^2$ bits, so a programmable hole requires per-operation error rates below $\\sim m_P^2/M^2$; for all but microscopic holes this is unattainable in practice.","Biological information processing operates far closer to the thermodynamic limits than digital electronics do, so the roughly $10^5$ gap in energy efficiency is an architectural feature of current artificial intelligence, not a fundamental limit.","Because the universe is at critical density, the same formulas bound the total computation the universe can perform over its history."],"supporting_citations":[{"why":"Supplies the quantum speed limit that bounds the number of operations by $2Et/\\pi\\hbar$.","marker":"[1]"},{"why":"Extends the speed-limit accounting to the computational capacity of the universe, which the black hole analysis inherits.","marker":"[2]"},{"why":"Establishes the thermodynamic cost of erasing each bit of error entropy, the basis for the linear growth of physical space complexity.","marker":"[3]"},{"why":"Identifies negentropy with thermodynamic depth, the quantity used to define $\\Phi\\text{-SPACE}$.","marker":"[10]"},{"why":"Supplies the initial proposal that black holes can act as quantum computers, the target of the intractability argument.","marker":"[14]"},{"why":"Provides models of unitary black hole evaporation that justify treating the outgoing radiation as carrying the results of the computation.","marker":"[15-16]"},{"why":"Models fast scrambling and information release from black holes, used to assess how hard it is to extract or decode the computation.","marker":"[17-18]"},{"why":"Supplies the cryptographic hardness argument that decoding the radiation has very high computational complexity.","marker":"[21]"},{"why":"Gives complexity-theoretic analysis of black hole transformations reinforcing the difficulty of unscrambling the radiation.","marker":"[22]"}],"fun_headline_variants":["Black hole computers can't store their own error garbage","Error entropy makes black hole quantum computers impossible","Free energy cost kills black hole computing","Physical complexity bricks black hole quantum computers","Black hole memory too small for quantum error correction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The black-hole intractability result assumes that all entropy generated by errors over the entire computation must be stored in the hole and can never exceed its instantaneous entropy capacity, $\\sim M^2/m_P^2$; if the radiation emitted by the hole can carry that entropy away continuously, the conclusion collapses.","fun_headline_variants_meta":{"raw":{"variants":["Black hole computers can't store their own error garbage","Error entropy makes black hole quantum computers impossible","Free energy cost kills black hole computing","Physical complexity bricks black hole quantum computers","Black hole memory too small for quantum error correction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000194,"raw_usage":{"total_tokens":1315,"prompt_tokens":867,"completion_tokens":448,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":483,"completion_tokens_details":{"reasoning_tokens":395}},"tokens_in":483,"tokens_out":448,"duration_ms":4587,"temperature":1.0,"reasoning_tokens":395,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:25:31.544692+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track the entropy budget in a unitary model of black hole evaporation with a radiation bath: if the bath can absorb $\\epsilon\\log(1/\\epsilon)$ bits per operation at a rate no larger than the hole's entropy-emission rate while the hole remains programmable, then the per-operation error rate need not be as low as $m_P^2/M^2$. A concrete model exhibiting this would falsify the paper's intractability claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the quantum speed limit that bounds the number of operations by $2Et/\\pi\\hbar$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Extends the speed-limit accounting to the computational capacity of the universe, which the black hole analysis inherits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the thermodynamic cost of erasing each bit of error entropy, the basis for the linear growth of physical space complexity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Identifies negentropy with thermodynamic depth, the quantity used to define $\\Phi\\text{-SPACE}$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the initial proposal that black holes can act as quantum computers, the target of the intractability argument."},{"cited_title":"fire-wallsJournal of High Energy Physics 61-56","cited_arxiv_id":null,"evidence_quote":"Supplies the cryptographic hardness argument that decoding the radiation has very high computational complexity."}],"review_version":1}