Imagine you're trying to detect whether a guitar string has shifted a hair's width sideways by shining a flashlight on it. Your flashlight beam has a profile — bright in the center, dim at the edges. If the string nudges the beam, you're measuring that nudge from the subtle change in the light pattern arriving at your detector. The question this paper answers is: given a fixed photon budget, what's the absolute best light state you can prepare to detect the smallest possible sideways shift, and can you actually build a detector that reaches that limit? The committed claim: the optimal quantum state for measuring small transverse beam displacements, under a hard photon-number constraint, asymptotically approaches a rotated NOON state — a specific superposition across two spatial modes. The authors derive this analytically using mode parameter estimation theory, working in the Hermite-Gauss and Laguerre-Gauss bases. They prove the quantum Cramér-Rao bound for this task and identify which states saturate it. The architectural choice matters. The paper works entirely within the framework of quantum parameter estimation — quantum Fisher information, symmetric logarithmic derivative, method-of-moments estimators. It contrasts two resource constraints: fixed total photon number (where the Hilbert space is finite and the optimal state is computable) versus fixed average photon number (where Gaussian states with squeezing are the natural contenders). This distinction is not academic bookkeeping — it determines which experiments are feasible. NOON states with fixed total photon number N scale as Heisenberg-limited (1/N sensitivity), but they are notoriously fragile. The paper's key practical contribution is showing that twin-Fock states — two modes each loaded with the same definite photon number — provide a viable alternative that nearly matches NOON-state performance while being far easier to prepare. On the ladder: the classical shot-noise limit scales as 1/√N for displacement sensitivity. The Heisenberg limit scales as 1/N. Gaussian states with optimal squeezing sit somewhere in between, approaching Heisenberg scaling asymptotically for large average photon number. The paper places all of these on the same quantitative footing for the beam-displacement task, which hadn't been done comprehensively before. This isn't beating a named SOTA in an experiment — it's establishing the theoretical scoreboard that future experiments will be measured against. Integrity is clean but bounded. This is a theoretical paper: proofs and analytical derivations, not experimental data. The Fisher information calculations are exact within the model; no fitting, no cherry-picking. But the validation is inherently self-contained — the authors are deriving bounds, not testing them against hardware noise, detector inefficiency, or real beam profiles. The measurement schemes proposed (method of moments with specific mode-sorting setups) are concrete enough to be experimentally testable, but nobody has tested them yet. The milestone gap is the bridge from theory to tabletop. Current spatial-mode sorter technology can handle low-order modes (first ~10 Hermite-Gauss or Laguerre-Gauss modes) with high fidelity. Demonstrating Heisenberg-limited beam displacement sensing with twin-Fock states at, say, N=10 photons on a real optical bench would validate the framework. That experiment is plausible within 2-3 years given existing photon-number-resolving detectors and mode sorters. The real unlock comes if this extends to sub-diffraction-limited microscopy or gravitational-wave detector alignment — domains where picometer-scale beam shifts matter. The experiment the authors didn't run: an actual photon-counting demonstration. The paper stays purely theoretical. The honest read is (a) — this is a theory group, and the experimental demonstration requires collaborators with spatial light modulators, photon-number-resolving detectors, and mode-sorting optics. Expect a follow-up with an experimental partner.
