Nirmātṛ

Entropy refuses monogamy: Clausius’s “inward turning” (Greek en-tropē) and Shannon’s bit-counting heir must expand into an infinite-dimensional menagerie of Rényi, smooth, and one-shot measures demanded by quantum non-commutativity. Physics invented entropy to tame heat engines; quantum information returned the favor by proving one engine cannot measure all transformations. Rubboli, Regula, Berta, Lami, Tomamichel systematically axiomatize the zoo – from isometric invariance, data-processing under channels, additivity, and normalization – and maps their operational roles across quantum communication, cryptography, resource theories, thermodynamics, and many-body physics. Von Neumann entropy (H(ρ) = −Tr(ρ log ρ)) remains the additive, strongly subadditive monarch of the asymptotic i.i.d. empire, uniquely recovered from relative entropy (D(ρ∥σ) = Tr(ρ(log ρ − log σ))) via the maximally mixed state, and tied to free energy and Holevo bounds. Yet quantum conditional entropy can go negative, certifying entanglement, while classical intuition collapses. Rényi family members (Petz, sandwiched) generalize moments of surprisal, delivering sharp Stein/Chernoff exponents for hypothesis testing; smooth min/max entropies (and information-spectrum cousins) handle the finite-n, non-i.i.d. one-shot frontier that real devices inhabit, regularizing back to von Neumann in the large-n limit. Canonical constructions from divergences yield the whole hierarchy, with dualities and chain rules (sometimes parameter-shifted) governing monogamy and recovery. Gaussian-state formulas via symplectic eigenvalues and algebraic extensions on von Neumann algebras push the framework into continuous-variable and infinite-dimensional regimes relevant to bosonic channels and quantum illumination. The review treats entropy not as a fixed thermometer but as a higher-dimensional toolkit; one Clausius, infinite operational faces. Surveying the post-2010 explosion of sandwiched Rényi and smooth entropies that made practical quantum key distribution and finite-size security proofs possible.

The Clean Paradox • Vacuum cleans the surface and amplifies the chaos. A cleaner diamond yields a louder surface storm. The molecules you hate keep the NV coherent. Shallow NVs prefer a little grime. Adsorbates act as charge-taming guardians: stripping the diamond surface bare in ultrahigh vacuum (UHV) shortens the Hahn-echo coherence time (T₂) of shallow nitrogen-vacancy (NV) centers by ~4×, because the molecules normally compensate excess surface charge that otherwise floods the low-frequency electric and magnetic noise spectrum. Charge compensation by dirt beats sterile purity. Shallow NVs – atomic-scale defects a few nanometers beneath the diamond surface – function as exquisite quantum sensors for magnetic and electric fields, yet surface-induced decoherence has long limited them. Conventional wisdom held that adventitious adsorbates (water, hydrocarbons, etc.) were a primary noise source. Yuan, Fehr, Zervas, Sangtawesin, Rodgers, Gumann, Flatté, and de Leon (Princeton, Iowa, Eindhoven, IBM) tested this by annealing and measuring single shallow ¹⁵N-implanted NVs in a custom UHV cluster tool (~5×10⁻¹⁰ mbar, adsorbate-free confirmed by in situ XPS and absence of proton NMR) versus ambient air. The result is the inverse: T₂ ratios (air/UHV) of 3–5.4, worse for shallower centers. Paradox of the pristine plane: fewer molecules, more noise. Single-quantum (SQ) and double-quantum (DQ) Hahn-echo measurements separate the noise channels. Both electric and magnetic contributions rise in UHV; magnetic noise becomes comparable or dominant. Dynamical decoupling reveals higher spectral density from ~10 kHz to 800 kHz with power-law exponents of 0.6–1.5. Estimated fluctuating surface-charge densities jump from ~8×10¹⁴ m⁻² in air to ~6×10¹⁵ m⁻² in UHV (correlation times 10⁻⁶–10⁻⁴ s). Depth scaling approaches the expected 1/d² for point-like surface charges under clean conditions. Meanwhile, DQ T₁ lengthens in UHV, indicating suppressed electric-field noise near 100 MHz • consistent with removal of adsorbate-related dipoles • while SQ T₁ changes little. Deep NVs show no environmental difference, confirming the surface origin. Charge-state stability also degrades for the shallowest centers in vacuum. The microscopic picture revises prior surface-noise models (Romach et al., de Leon group surface spectroscopy): low-frequency noise arises mainly from slow charge fluctuations and mobile charged species whose density and dynamics increase without adsorbate compensation; intermediate-frequency electric noise is tied to adsorbate dynamics; higher frequencies approach bulk limits. A finite-thickness adsorbate layer in air further softens the depth scaling relative to an ideal 2-D charged plane.

Entanglement Horizon Breached | Lattice DMT Tames Higher-Dimensional Quantum Hydrodynamics with a 1D Spatula. 1D Tensor String Colonizes 3D Quantum Hydrodynamics. Classical computers now extract accurate late-time diffusion constants and hydrodynamic crossovers from microscopic quantum models in 2D and 3D at high temperature, overturning the belief that entanglement growth rendered this classically intractable beyond one dimension. Lattice DMT turns higher-dimensional chaos into locally preserved currents. One-dimensional matrix product states outmaneuver the area-law curse. A strictly one-dimensional tensor-network ansatz obsessed with taxicab geometry unlocks three-dimensional late-time transport that pure geometric locality had made classically unreachable. From chain DMT’s linear blindness to lattice DMT’s geometric clairvoyance. Quantum many-body systems entangle so rapidly in higher dimensions that tracking late-time energy or spin flow was a computational dead end. Lattice Density Matrix Truncation (L-DMT) wraps the lattice in a one-dimensional matrix-product-state string while rigorously preserving every geometrically local operator (Manhattan distance ≤2) that controls hydrodynamics • densities, currents, and nearby correlations setting the macroscopic drift • and shifts truncation error into distant, high-weight correlations that hydrodynamics ignores. Simulations remain quantitative to Jt ≳ 100, an order of magnitude beyond prior classical methods that failed around Jt ∼ 1. On the mixed-field Ising model, 2D isotropic energy diffusion gives D = 0.326 ± 0.004 after extrapolating bond dimension and cylinder width; 3D cylinders show only weak finite-size dependence. Coupled Heisenberg chains capture the slow superdiffusive-to-diffusive crossover (KPZ-like z_S ≈ 1.5 drifting to ordinary z_S = 2) near t ∼ 80/J as interchain coupling weakly breaks integrability. Bond dimension scales only linearly (χ ≳ 6L_⊥ + 2) instead of exponentially. Energy-conservation violations drop by orders of magnitude versus standard SVD or chain DMT; C_4 symmetry is preserved on square lattices.

Light cones crystallize in the agent’s latent night sky, null geodesics as the cheapest way to keep quantum gossip coherent. Spacetime is just the memory palace an AI builds so its probes stop talking past each other. Spacetime’s causal skeleton (null structure / light cones) spontaneously crystallizes inside a model-based RL agent’s memory purely from maximizing quantum mutual information between Unruh–DeWitt probes in Minkowski space, no metric or geometry ever supplied. The causal structure didn’t fall from the sky; it condensed inside a planner’s score field like frost on a higher-dimensional window. LeMaitre & Perche (Innsbruck/Nordita) train a latent-dynamics RL agent (encoder → latent model → reward model → planner, horizon H=1) whose sole job is to drive two localized qubit detectors interacting with a massless scalar field so that their quantum mutual information (I_{ij}=S(ρ_i)+S(ρ_j)-S(ρ_{ij})) is maximized. Detector 1 sits at the origin; detector 2’s smearing center moves on a 301×301 grid in an effectively (1+1)-dimensional slice. Minkowski metric recovered at 10⁻⁴ from pure information hunger, geometry as an emergent compression artifact. Null ridges grow in silicon the way river deltas grow in water: the path of least surprise. The agent sees only operational detector outcomes and a simple reward based on a normalized, transformed mutual-information signal (higher when the probes are “close” in the information sense). After ~4000 episodes it hits 100% success; its internal planner scores then spontaneously develop sharp diagonal ridges that exactly track the light-cone branches of the mutual-information landscape. An Unruh–DeWitt pair taught a model-based RL agent the shape of “now,” and the agent answered with diagonal light. Spacetime emerges the moment the agent stops guessing and starts predicting which moves keep mutual information alive. Post-processing the planner’s softmax-average and best-action fields (neighbor-mask threshold ~70%) recovers the null structure to high precision: fitted metric components (g_{tt}^{fit}≈0.999878), (g_{tx}≈0), null slopes (dx/dt≈±0.999939) (deviation ~6×10^{-5}). The causal skeleton of Minkowski spacetime sits in the agent’s memory as the most efficient predictive scaffolding for the task. The light cone is the agent’s private theory of everything for one very specific task: don’t let the quantum conversation die. The metric is what remains after you delete every degree of freedom that doesn’t help the probes share secrets. The agent never knows relativity; it only learns that certain relative placements of the probes let quantum information flow efficiently while others kill it. The light-cone structure is the unique low-dimensional manifold that organizes those successes and failures into reliable predictions. A blind chess player inventing the geometry of the board solely by counting which piece exchanges transfer the most advantage entropy, or a bat evolving echolocation maps that recover Euclidean space without ever being shown a ruler. Geometry is just the AI’s optimized gossip protocol. Light cones grow in silicon synapses – spacetime emerges as an action-relevant compression of quantum correlations. Limited to flat space with a conformal-gauge reconstruction step; the authors flag extensions to curved spacetimes, distinguishing causal vs state-dependent correlations, and tasks that require spacetime structure (relativistic quantum communication).

Indices are for mortals, matrices are for the harmonic gods. Gravity just got matrixed, indices are extinct. We didn’t bend gravity’s rules. We deleted the subscripts. Einstein’s gravity, rewritten as pure matrix algebra. Jaume de Haro recasts General Relativity entirely inside the algebra of 4×4 real matrices, using the harmonic gauge. Traditional index-heavy tensor notation is replaced by musical isomorphisms (♯, ♭), metric adjoints, and matrix-valued derivative operators. The Ricci tensor collapses into an algebraic structure dominated by the Laplace–Beltrami operator (□_LB ≔ tr(G♯ ∂²)), so Einstein’s equations become a compact nonlinear matrix wave equation for the metric matrix G: □_LB 𝒢 = −16π G_N P̂_−(𝒯 + T̄). The gravitational energy-momentum pseudotensor T̄ emerges naturally and is algebraically linked to the classical Landau–Lifshitz complex, with local conservation written as a clean matrix divergence ∂ᵀ((−g)(𝒯♯ + 𝒯_LL♯)) = 0ᵀ. It recovers the weak-field wave equations, derives the exact Friedmann dynamics in harmonic time (H_har² = (8π G_N / 3) a⁶ ρ), and formulates linear cosmological perturbations in both Newtonian and harmonic gauges. Index gymnastics become linear-algebra bookkeeping. The novelty is the self-contained, index-free coordinate-matrix framework that makes GR look like a nonlinear matrix PDE rather than a forest of Christoffel symbols. A formal reformulation, not new physics, but the compactness could accelerate symbolic manipulation, post-Newtonian expansions, and numerical relativity codes that already exploit matrix libraries. Harmonic coordinates have long been favored for wave-like equations; packaging them this way simply makes the bookkeeping invisible. Limitations are those of the harmonic gauge itself (coordinate singularities can still appear) and the fact that the approach is coordinate-dependent by design. It does not replace covariant formulations for conceptual work, but it offers a practical computational dialect. A genuine algebraic reorganization of known equations, not a claim of new solutions or modified gravity. Sits in the long tradition of coordinate tricks that make Einstein’s theory more tractable.

Amyloid fibrils may be quantum photoprotection, not just pathology. Biology is the quantum computer – it never waited for artificial ones. Life’s warm, wet chaos does not kill quantum effects; protein fiber-optic networks form superradiant superhighways that coordinate biology from atom to organism at room temperature. 98-page landmark review from Howard University’s Quantum Biology Laboratory (Philip Kurian et al.) surveys ~600 works spanning a century and reframes the field. It moves beyond isolated test-tube experiments to whole living systems, where quantum processes (superradiance, tunneling, entanglement, coherent vibrations) shape structure, function, and behavior even at ultraweak, single-photon levels. Tryptophan networks in microtubules, actin filaments, and other protein fibers show collective superradiance; Dicke’s 1954 prediction realized decades later. Photons are shared coherently across many molecules, boosting emission rates and fluorescence quantum yield. These structures act as biological nanoantennas with giant dipole moments, enabling long-range electrodynamic signaling unscreened by cellular ions. The effect persists in thermal equilibrium in aqueous solution, defying the “warm-wet decoherence kills quantum” dogma. Similar collective behavior appears in photosynthetic antennae and green-sulfur bacteria. Other pillars: Fröhlich-style condensation: Vibrational energy concentrates in low-frequency protein modes, forming microscopic antennas that help binding partners find each other with speed and specificity in crowded cells. Quantum tunneling (protons in enzymes, DNA mutations), many-body dispersion forces, ultraweak metabolic photon emissions (Gurwitsch’s mitogenic radiation, now measurable with single-photon detectors), radical-pair magnetoreception in birds and other organisms, and chirality-induced spin effects. Historical roots: Bohr’s “Light and Life,” Schrödinger’s What is Life? (which helped inspire the DNA search), and older optical ideas. Sub-neuronal brain mapping and processing (cytoskeletal quantum channels orders of magnitude faster than classical ion flows); photoprotective roles for protein aggregates (including amyloids linked to neurodegeneration); enzyme catalysis; evolutionary dynamics; bio-inspired computing (slime molds solving hard problems without neurons). Quantum technologies (squeezed/entangled light, nanodiamond sensors, quantum dots) enable less-damaging probes of living tissue; biology inspires new quantum devices. The review argues the quantum layer is structural scaffolding for coordination across scales, not mere footnotes to classical chemistry.

Error Correction Just Got a Roommate • Noise Hates It. 230× more hard ops or bust – AMD’s GPUs just moved next door to babysit IBM’s still-drunk 120-qubit Nighthawk. IBM, AMD, and ORNL’s Quantum Science Center co-located the 120-qubit Nighthawk with Umbra (AMD MI355X GPUs + 5th-gen EPYC CPUs) at IBM’s Poughkeepsie site. The goal is real-time quantum error correction: classical hardware must decode errors and feed corrections back before noise ruins the calculation. Physical adjacency eliminates the latency that previously made hybrid workflows impractical. DOE’s Quantum Genesis Mission targets 100 logical qubits running 100,000 hard operations by late 2028. IBM is currently at ~70 logical qubits and 438 hard ops—a roughly 230-fold gap. Hierarchical codes are one path; this testbed is where researchers will stress-test decoders, pulse control, and workflows that could inform IBM’s 2029 fault-tolerant Starling system.

Electron Auroras Replace Stick Figures | Quantum Crystallography Broadcasts Biology’s Charge Weather in Real Time. Electron-cloud cartography is rewriting structural biology: quantum crystallography (QCr) extracts not just atomic coordinates but the full quantum fuzz of electrons – densities, electrostatics, and constrained wavefunctions – from diffraction data, turning stick-figure protein models into charged landscapes that explain binding, catalysis, and drug selectivity. Traditional crystallography delivers atomic positions via the independent atom model (IAM), still used for >99% of the ~2 million structures despite a century of refinement. QCr, coined in 1995 by Massa, Huang, and Karle, fuses quantum mechanics with scattering: either constraining experimental structure factors to N-representable density matrices/wavefunctions or injecting QM calculations (multipoles, X-ray constrained wavefunctions, Hirshfeld atom refinement/HAR, transferable aspherical atom models/TAAM) to enhance refinement. For biomolecules this yields precise charge distributions, hydrogen positions, bonding topology (via QTAIM), and interaction energies beyond geometry alone. Recent IUCr focus issues (commemorating the 2025 quantum-mechanics centenary) showcase bio applications: QCr of tyramine polymorphs links crystallization conditions to stability for formulation design; NSAID–COX studies quantify electrostatic drivers of selectivity, pinpointing key residues; TAAM reveals charged moieties even at moderate resolution in proteins. Quantum refinement (QM/MM or machine-learned potentials) corrects ligand strain, protonation, and coexisting bonded/nonbonded forms of nirmatrelvir in SARS-CoV-2 Mpro; cutting computational cost while reaching DFT accuracy. Kernel energy methods and fragmentation extend this toward large systems. Mariusz Jaskólski (Adam Mickiewicz University, Poznań; Max Perutz Prize 2024) embodies the experimental bedrock: ultra-high-resolution macromolecular work (subatomic proteins, Z-DNA at 0.55 Å), HIV protease structure (pivotal for AIDS drugs), asparaginases, amyloids, and modulated structures. These datasets supply the diffraction quality QCr needs for electron-density and electrostatic analysis in health-related targets. Classical crystallography is a skeleton X-ray; QCr is the glowing aura map of the electron cloud that steers enzymes and receptors. Roots in the 1970s, boom in the last decade, still maturing for routine proteins yet already guiding rational design. Structure–function mapping that predicts binding free energies and reaction pathways from experiment alone – biology’s quantum operating system finally visible.

Simplest molecule audits the omniverse, fivefold accuracy upgrade exposes theory’s lag. Measure the simplest molecule perfectly, or admit the Standard Model has holes. H₂’s bond audited to 175 kHz – five times tighter, caging BSM ghosts at 540 kHz. The simplest neutral molecule’s ionization energy for para-H₂ is now 124,417.491,1573(58) cm⁻¹ and dissociation energy 36,118.069,6646(58) cm⁻¹, both known to 175 kHz (relative 5×10^{-11}). Like measuring the Earth–Moon distance to a hair’s thickness. Fifteen new para-H₂ transitions (X → EF → high-n Rydbergs, n=48–70) plus a 45-interval network from three decades of Zurich–Amsterdam work deliver the jump; para-H₂’s zero hyperfine structure removes a prior systematic. Previous experimental value sat ~1.6 combined sigma lower with fivefold larger uncertainty. Theory still carries ~780 kHz error, so experiment now leads. Same data tighten any beyond-Standard-Model contribution to the para–ortho energy difference to ≤540 kHz; an independent stress test for two-electron QED (relevant to the helium ionization tension). Nine orders of magnitude tighter than 1910s chemical estimates; four-plus orders better since the turn of the century via the thermochemical cycle linking ionization to dissociation. Next, mandatory α^7 QED terms, sharper proton-size sensitivity, possible light-boson hunts.

P Equals NP, Yet Quantum Still Proves Itself. Quantum advantage can exist even if P = NP – and still fail to unlock cryptography. Quantum oracles that equalize rather than elevate. Princeton and Kyoto researchers (Goldin, Lombardi, Morimae) constructed classical oracles relative to which proofs of quantumness exist, yet P^O = NP^O and BQP^O = PSPACE^O. A unitary oracle further shows quantum-verifier oblivious state preparation can coexist with BQP^O = QCMA^O. These relativized worlds demonstrate that evidence of quantum superiority need not force separations between classical complexity classes, nor yield relativizing constructions of one-way functions or most cryptographic primitives. An oracle is a black-box computational helper. Relative to the first, a quantum machine can still convince a classical verifier it is quantum (via polynomial-round public-coin proofs of quantumness or two-message quantum-verifier variants), while classical nondeterminism collapses into ordinary polynomial time. Simultaneously, quantum computation matches the full power of classical polynomial-space machines. Nearly all classical cryptography collapses when P = NP; most quantum cryptography collapses when BQP = PSPACE. Thus proofs of quantumness cannot be converted, via relativizing techniques, into secure classical or quantum crypto, oblivious transfer, or proofs of quantum memory. Practical quantum hardware and concrete hardness assumptions (LWE, factoring, etc.) remain unaffected. Oracle worlds where quantum supremacy coexists with cryptographic emptiness. Quantum power without cryptographic leverage, the relativized limit.