3
[2026-09-04] · src Berthusen, Lavasani, Benhemou, Allman, Dreiling, Estey, Foltz, Jacobs, Mills, Park, Reed, Hayes, Metodi, Potter (Quantinuum); arXiv:2609.03194 (2026)
On Quantinuum's 98-qubit trapped-ion processor Helios, the team encoded two logical qubits in twenty physical qubits at code distance 6 (the 20,2,6 "C4-Helix" code) and ran repeated rounds of quantum error correction, measuring a logical error of 4.6(+6.2/-2.6)x10-5 per logical qubit per QEC cycle. They then benchmarked the full Clifford group on those two logical qubits under active correction (2.8(+1.0/-1.6)x10-4 error per two-qubit logical Clifford gate), and built a fault-tolerant "chain-map" interface joining a C4-Helix block to a separate distance-5 surface code to prepare a three-logical-qubit GHZ state spanning both codes, with a fidelity lower bound of 99.925(+0.068/-0.245)%. In every one of these tests the encoded, corrected result beat its corresponding unencoded physical baseline, with no postselection on favorable outcomes.
Low logical error rates on their own are not new - Google's Willow already showed error suppression that scales with code distance. What this adds is the rest of the stack shown working together on one machine: an efficient code (20 physical qubits per 2 logical qubits, far leaner than a comparable surface-code block), logical two-qubit gates under correction, and - the harder piece - a working bridge between two different error-correcting codes, which is what a real fault-tolerant computer needs to reach the non-Clifford resources a single code family cannot supply cheaply on its own. None of the circuits demonstrated here (Clifford-group benchmarking, a three-qubit GHZ state) are beyond classical simulation, so this is an architecture proof rather than a computational-advantage claim - the question it answers is whether these pieces compose without the overhead eating the benefit, and here they did not.
[2026-09-04] · src Zeng, Qin, Fan, Ahmed, Fang, Long, Shan, Chen, Xu, Deng, Chu, Bao (University of Houston); arXiv:2609.02995 (2026)
A University of Houston group demonstrates one-step superconductivity validation using a room-temperature, cryogen-free optically pumped atomic magnetometer to detect the magnetic field of a sample's own remnant supercurrent, with no electrical contacts, coils or integrated sensors on the sample itself. The transition is read directly off two signatures: the abrupt disappearance of the supercurrent's field above Tc, and reversal of the supercurrent's direction when the applied field is reversed. Tested on YBCO microcrystals and REBCO tape, the technique picks up picotesla fields from supercurrents induced by nothing stronger than Earth's ambient field, in a millimeter-scale REBCO disk and in sub-100-micrometer YBCO crystals sized to fit inside a diamond-anvil cell; a ferrite flux guide extends sensing out to centimeter distances.
High-pressure superconductivity claims keep stalling on the same two weaknesses: a zero-resistance reading can come from a short across electrical leads that barely survive diamond-anvil pressures, and Meissner- effect confirmation normally needs delicate, hard-to-replicate SQUID or NV-diamond setups built around the specific sample. A contact-free technique validated on known superconductors at exactly the sample sizes a DAC uses does not resolve any current claim by itself, but it removes one of the recurring reasons those claims stay unresolved for years: an independent lab could in principle check a sample without first reproducing the original group's contacts and wiring.
[2026-09-04] · src Ahn, Kim, Park, Lee, Kwon, Chung, Byun, Uchaikin, van Loo, Nakamura, Youm, Youn, Semertzidis (CAPP/IBS); arXiv:2609.03879 (2026)
Axion haloscopes need a high-Q cavity inside a strong magnet, and standard superconductors suffer severe vortex dissipation at multi-tesla fields while copper is capped around Q ~ 105 by the anomalous skin effect. The CAPP/IBS group built a pole-to-pole three-dimensional cavity from strain-controlled, delaminated REBCO tape, stripping the lossy metallic substrate and using the remaining copper stabilizer as a conductive backing so that the seams between tape segments act as waveguides below cutoff rather than as leakage paths. A 5.4 GHz cavity built to test the material alone reached an unloaded Q0 of 1.4x107 in an 8 T field, two orders of magnitude above the copper baseline. A second, tunable 2.3 GHz version, integrated into the live Pilot Axion Cavity Experiment (CAPP-PACE), reached a 180 mK noise temperature and a 5-fold Q gain over copper, for a combined ~8.4-fold acceleration of the experiment's scan rate.
An axion haloscope's reach is scan-rate limited: how much of the axion's allowed mass range it can rule out per unit time is set directly by cavity Q and system noise. An order of magnitude of Q recovered specifically in the high-field regime where superconductors normally fail is not a new idea - HTS tape cavities for axion searches have been under development for several years, including by parts of this same collaboration - but going from a bench demonstration to a cavity running inside CAPP-PACE now is what turns an 8-fold number into search coverage rather than a promise of it, for one of the best-motivated dark-matter candidates and the leading solution to the strong-CP problem.
2
[2026-09-04] · src —
Alto takes today's slot — three days since its last session is the longest gap of the three instruments right now, and trumpet went yesterday, so the two chops-dependent instruments don't land back to back. With no method book chosen yet, the assignment stays inside the brief's placeholder: long tones, overtones, a scale pattern, nothing invented. Today's fundamental is low F, continuing the upward climb the survey has held since it passed E♭, and both the warmup shape and the musical-target pattern open a second lap rather than reaching for novelty for its own sake — messa di voce is back after four sessions cycling through the other three breath shapes, and the whole-tone scale gets the broken treatment the major-scale patterns already had.
5 minutes: long tones on low F and low F♯, messa di voce on each — pp building to f over four counts, then receding back to pp over four more, quarter=60, two full cycles per note. The target is a swell whose peak doesn't sharpen and whose return doesn't go breathy — both ends of the shape holding center, not just the attack or the release alone, which is what the last four single-direction sessions each isolated separately.
Overtone series from fundamental low F: voice up through F–F–C–F–A using only the low F fingering, changing partials with throat and voicing, not embouchure — the same method as every prior session, now a fifth semitone higher than where the survey started this climb. Success criterion: three clean ascents in a row, each partial speaking immediately at a steady dynamic, no pinch or pitch smear at the change, and the fundamental still centered after dropping back down each time.
F whole-tone scale (F–G–A–B–C♯–D♯), one octave, in broken groups (1-3-2-4-3-5-4-6-5-1 through the six-note scale) rather than the straight ascent and descent 08-23 played on this same scale, or the broken thirds, triads and fourths of the last three sessions on the major scale. There is still no half-step anywhere in it and still no note that resolves — shape it as a phrase anyway, leaning into the top of each broken pair and tapering into the next, so the absence of a resting tone reads as a choice rather than as a run being gotten through.
Stop the moment either half of the messa di voce swell loses its center — a sharp peak or a breathy tail — or an overtone needs a jaw shove or throat squeeze to speak. Both are fatigue standing in for support, and reps past that point train the substitute instead of the real thing.
[2026-09-04] · src —
hearJohnny Hodges — Warm Valley (Passion Flower 1940–46) · Tidal · Spotify
Cut with the Ellington orchestra in October 1940, "Warm Valley" gives Hodges one long unaccompanied-feeling melodic arc built almost entirely out of swells placed inside sustained notes — the same messa di voce shape today's warmup is drilling, except played as the entire point of the performance rather than as an exercise. Listen for how the crescendo and the taper both stay centered in pitch, with neither end of the swell announcing itself as effort.