4
[2026-08-11] · src arXiv:2510.04869 (Double Chooz Collaboration); Phys. Rev. Lett. 137, 061803 (2026), doi:10.1103/dr26-j19g
With both Chooz-B reactor units fully shut down, Double Chooz's near detector logged 17.2 days of reactor-off data and found 106(+/-18) antineutrino candidate events at 5.9 sigma significance. The source isn't active fission: it's beta decay of long-lived fission products (e.g. 90Y, 92Y, 144Pr) still cooling in the core and in the adjacent spent-fuel pool. This is the first quantitative measurement of that post-shutdown flux, rather than a theoretical estimate of it.
Antineutrinos pass through concrete and water shielding that stops every other diagnostic, which is why reactor safeguards work has wanted this signal for decades — the open question was whether it was even detectable above backgrounds, not what it would show once seen. A confirmed, quantified flux turns "monitor a reactor's fissile inventory by its antineutrinos even while it's off" from a proposed technique into a demonstrated one. JUNO-TAO is running the same measurement at a different reactor and detector design; whether that reproduces is the next real test.
[2026-08-11] · src arXiv:2503.10935; Nature 656, 47-53 (2026), doi:10.1038/s41586-026-10822-y
Dual-rail erasure qubits encode each logical qubit across two microwave cavities, so a photon-loss event shows up as a heralded "erasure" at a known location instead of an undetected bit- or phase-flip. The new two-qubit gate — swap-wait-swap — temporarily parks an excitation on the coupling transmon to pick up a strong dispersive shift, runs in about 500 ns, and reaches roughly 99.9% raw two-qubit fidelity with a ~0.5% erasure rate. Post-selecting on no detected erasure, residual Pauli error falls below 0.1%: bit-flip error is suppressed to the parts-per-million level, leaving dephasing as the dominant remaining channel.
Below-threshold surface codes such as Google's Willow still spend most of their physical-qubit budget on syndrome extraction because the code can tell that an error happened but not which one. An architecture that both locates the error and holds 99.9% two-qubit fidelity cuts the overhead the threshold math demands, without a new code — it changes the error budget a code has to work with, not the threshold itself. No logical-qubit or memory experiment built on this gate exists yet; this is a component result.
[2026-08-11] · src arXiv:2509.18075 (Greiner group, Harvard); Nature (2026), doi:10.1038/s41586-026-10875-z
Using ultracold lithium-6 in a Hubbard-regime optical lattice (roughly 300 uniform sites at a target doping, embedded in a ~150-site reservoir), the group combined two independent probes at temperatures only recently reached in these simulators: compressibility measurements showing a doping-dependent maximum that sharpens on cooling, and lattice-modulation spectroscopy showing low-energy spectral weight vanishing non-uniformly across the Brillouin zone — the direct spectroscopic signature of a pseudogap, not an inference from transport. Both probes agree on where the underdoped and overdoped regimes split.
The pseudogap is the unexplained partial-gap phase sitting between the Mott insulator and the superconducting dome in the cuprates, and every rival theory of high-Tc superconductivity has to account for it. Getting both a thermodynamic and a spectroscopic signature from the same synthetic Hubbard model — no disorder, no unknown material chemistry — is a cleaner test of whether short-range magnetic correlations alone are enough to produce it, a question real cuprate samples can't isolate. This strengthens the case for "yes," but the simulator still can't reach the temperatures or system sizes where an actual superconducting dome would appear.
[2026-08-11] · src arXiv:2602.15655 (MPI for the Science of Light + U. Ottawa); Optica 13(8), 1508 (2026), doi:10.1364/OPTICA.601797
The team concentrated broadband, spatially and temporally incoherent sunlight through a Fresnel lens into an optical fiber, then fed it into a periodically-poled KTP crystal inside a polarization Sagnac interferometer — the same spontaneous parametric down-conversion (SPDC) geometry normally pumped by a coherent laser — to generate polarization-entangled photon pairs. Full quantum state tomography over three days of outdoor operation gave a Bell-state fidelity of 93.9(+/-2.7)%, concurrence of 0.905(+/-0.053), and a clean CHSH Bell-inequality violation: numbers in the range of conventional laser-pumped SPDC sources.
Nothing about the entanglement mechanism is new — SPDC is 1960s-vintage nonlinear optics, and this doesn't test or extend the theory. What changes is the resource assumption: a coherent laser pump was taken to be necessary for practical SPDC and evidently isn't, which matters for solar-powered or space-based quantum links where a stabilized laser is the most fragile, power-hungry component. Read this as an engineering result, not a physics one — the pump source is now an infrastructure choice, not a stated requirement.
1
[2026-08-11] · src build/orrery-model.el
orrery-model--node-properties (build/orrery-model.el:47-58) turns one
headline's PROPERTIES drawer into an alist:
(let ((drawer (org-element-map (org-element-contents headline)
'property-drawer #'identity nil t 'headline)))
(when drawer
(org-element-map drawer 'node-property
(lambda (np)
(cons (upcase (org-element-property :key np))
(org-element-property :value np))))))org-element-map takes six arguments here: DATA, TYPES, FUN, INFO,
FIRST-MATCH, NO-RECURSION. The first call passes (org-element-contents
headline) as DATA, 'property-drawer as TYPES, #'identity as FUN, nil
for INFO, t for FIRST-MATCH — and 'headline as NO-RECURSION.
Org's tree is recursive: a headline's contents include its property drawer,
its paragraphs, and every child headline, each of which can carry its own
property-drawer. Left to its defaults, org-element-map walks all of
that — it would find whichever property-drawer comes first in a full
depth-first walk, which is usually the parent's, but only by accident of
document order. Pass 'headline as NO-RECURSION and the walk still visits
each child headline itself (so a search for type 'headline would still
find it) but never descends into one — none of a child's own contents,
including its drawer, are ever visited. Combined with FIRST-MATCH t, the
call is now guaranteed to return this headline's own drawer, or nothing,
never a subheading's borrowed from below.
orrery-model-panels calls this once per top-level panel headline in
dashboard.org. Without the NO-RECURSION guard, a panel headline with no
drawer of its own but a child headline that happens to have one — a stray
subheading under a panel entry, say — would silently inherit the child's
:ID:, :WEIGHT:, :LIMIT: instead of erroring via
orrery-model--prop!. That is exactly the failure mode the model layer
exists to prevent: a plausible-looking dashboard built from the wrong
property drawer, with no signal that anything went wrong.
Exercise (verifiable): in a scratch copy, change line 53's nil t
'headline) to nil t) — i.e. drop the NO-RECURSION argument entirely —
and run make golden. Read the diff it prints before deciding whether to
explain it: with the fixtures as they stand today, does any panel headline
actually have a child headline with its own drawer, or does the golden
output stay byte-identical despite the change? Either answer is the
lesson — a fixture set that can't distinguish the two behaviors is itself
worth noticing. Then restore line 53 to nil t 'headline); this repo's
build/ is not yours to leave edited.
2
[2026-08-11] · src generic fundamentals (no method book chosen yet)
First session on alto in the rotation's recorded history. With no method book chosen yet, today is fundamentals only — tone and pitch center, not repertoire. The open question below is what turns this into a real syllabus.
5 minutes: long tones on low B♭, B, and C — the three throat tones that are easiest to let go thin and sharp — 4 counts each at mf, aiming for a centered pitch and a level dynamic with no swell or fade on any note.
Overtone series from fundamental low B♭: voice up through B♭–B♭–F–B♭–D using only the low B♭ fingering, changing partials with the throat and voicing, not the embouchure. Success criterion: three clean ascents in a row, each partial speaking immediately and at a steady dynamic, with no pinch or pitch smear at the change.
Take the B♭ major scale, one octave, and shape it as a phrase rather than playing it evenly — a slow crescendo to the top, a taper on the way down, like one long breath. Repeat in E♭ major (up a fourth) to check whether the shaping survives a new fingering pattern rather than being tied to the first key.
Stop when the tone turns bright or thin, or when locking in an overtone starts costing extra mouth pressure — that is embouchure fatigue setting in, and practising past it trains the wrong feel for what a centered note should require.
[2026-08-11] · src —
hearJohnny Hodges and His Orchestra — Passion Flower (Passion Flower 1940-46) · Tidal · Spotify
Almost any Ellington-era ballad feature — the sound stays centered and level through long notes with barely any attack noise, which is exactly the target of today's long-tone warmup and the overtone work built on top of it. Listen for how little the tone moves within a single note, not for anything around it.