3
[2026-09-26] · src Stiskalek, Desmond, Lavaux; arXiv:2609.29996 (2026-09-24)
Most local measurements of the Hubble constant chain several rungs of distance indicator together, so disagreements over any one rung's calibration are hard to isolate. Stiskalek, Desmond and Lavaux instead build a ladder with only two rungs: tip-of-the-red-giant-branch (TRGB) distances to roughly 400 Extragalactic Distance Database hosts, calibrated directly on the geometric anchors of the Large Magellanic Cloud and NGC 4258. They forward-model tip magnitudes and host redshifts - including selection effects and density-dependent galaxy bias - conditioned in turn on each of 80 Manticore-Local reconstructions of the local density and peculiar-velocity field, since most hosts sit close enough (cz below roughly 1000 km/s) that peculiar velocities matter more than calibration. The redshift likelihood and density smoothing each shift H0 by less than 1 km/s/Mpc, and the inferred velocity monopole is 5.4 +/- 3.4 km/s/Mpc, but the field-to-field scatter across the 80 realizations is 5.2 km/s/Mpc - larger than any of those. Weighting the 80 posteriors equally gives H0 = 68.4 +/- 5.9 km/s/Mpc; weighting by Bayesian evidence gives 72.3 +/- 2.6 km/s/Mpc, though that number is set by a single high-evidence realization and is sensitive to how finely the reconstruction posterior was sampled. The paper's own conclusion is that this ladder cannot yet arbitrate the Hubble tension.
The Hubble-tension debate usually turns on calibration - Cepheids versus TRGB versus masers - but this result relocates the bottleneck: at the distances where most TRGB hosts actually sit, the dominant uncertainty is imperfect knowledge of the local peculiar-velocity field, not the standard candle itself. That reframes where the next gain has to come from - extending the TRGB sample to greater distance, where peculiar velocities matter less, and improving local-volume reconstructions - rather than further refining TRGB photometry. It is also a useful caution about the evidence-weighted number, the one that lands closer to the SH0ES value: it is a single reconstruction's opinion wearing a small error bar, not five independent lines of evidence converging.
[2026-09-26] · src Lee, Magueijo, van de Bruck, Di Valentino; arXiv:2609.29206 (2026-09-24)
Thermogravity theories clip the Einstein equations to a trace-free version, remove the cosmological constant's gravitational coupling, and permit controlled violations of energy conservation that block the usual trick of reinstating Lambda as an integration constant. Lee, Magueijo, van de Bruck and Di Valentino set Lambda = 0 outright and ask whether late-time cosmic acceleration can come solely from that energy non-conservation, with no extra background parameter relative to flat LambdaCDM. Tested against supernova distances, BAO (including DESI) and CMB/BBN calibration, a universal implementation - every species undergoes non-conservation - is strongly disfavored, with DESI BAO alone worsening the fit by Delta-chi2 of about 51 relative to LambdaCDM (a caveat the authors flag themselves: universal non-conservation would also alter the standard observational dictionary, so the comparison isn't fully apples-to-apples). Restricting non-conservation to cold dark matter alone - leaving baryon and photon observables untouched at the background level - gives a self-contained one-parameter extension of LambdaCDM that actually fits the background data slightly better: Delta-chi2 of -4.85 and a maximum Bayesian preference of ln B = 2.77, a weak-to-moderate edge. But once the BAO ruler is calibrated with BBN or CMB information, the reduced effective early-time CDM density raises the sound horizon and pushes H0 lower - widening, not narrowing, the gap with distance-ladder measurements of H0.
This is a calibrated near-miss rather than a discovery: one branch of the model is ruled out cleanly by DESI-era BAO, and the branch that survives the background-level test only does so by making the Hubble tension worse rather than better, which is precisely the problem a plausible alternative to Lambda would need to solve. It is a useful data point for anyone tracking the post-DESI graveyard of no-cosmological-constant models - most fail on distance data alone, and the rare survivor here fails on the tension it might have been hoped to fix.
[2026-09-26] · src Ouyang, Wu, Xiao, et al.; arXiv:2609.29873 (2026-09-24)
S4 0954+65 is a TeV-detected, intermediate-synchrotron-peaked BL Lac blazar at z = 0.3694 that entered an exceptional gamma-ray high state in early 2025. Ouyang et al. combine gamma-ray, X-ray, optical/UV, radio and 43 GHz VLBA observations, using the z-transformed discrete correlation function to cross-correlate bands and VLBA imaging to track the parsec-scale jet. The gamma-ray light curve leads the optical by 3.22 days and the radio by possibly 18.26 days, with no significant gamma-ray/X-ray correlation; both the gamma-ray and X-ray spectra harden as the source brightens, consistent with enhanced particle acceleration during the active state. During the flare, VLBA imaging catches a new superluminal radio knot emerging from the jet core, and its extrapolated ejection time lines up with the gamma-ray peak. Broadband spectral-energy-distribution modeling shows that a leptonic synchrotron self-Compton plus external-Compton model, with the dusty torus supplying seed photons, reproduces the multi-wavelength data with physically plausible parameters; a lepto-hadronic origin remains possible in principle but would require substantially higher jet power than the leptonic fit needs.
Tying a VLBI-resolved structural event - the birth of a specific new jet knot - directly, in time, to a source's brightest-ever gamma-ray state is a more direct test of the shock-in-jet picture than the light-curve correlations blazar models are usually judged against. The energetics also favor a leptonic engine over a hadronic one for this flare, without ruling hadronic emission out entirely - a distinction that matters beyond this one source, since it is exactly the leptonic-versus-hadronic question that decides whether a given blazar should be expected to produce the high-energy neutrinos IceCube and its successors are looking for.
2
[2026-09-26] · src twelve-keys/src/exercises/half_whole/exercise.py
Piano gets the hour today: three days since its last session on the 23rd is the longest real gap of the three instruments right now, alto went yesterday and trumpet the day before, continuing the alto/piano/trumpet cycle this rotation has held since mid-September. It closes lap three's modes-and-scales slot with half_whole, the eight-note symmetric scale built by alternating half and whole steps — a different construction principle from altered_scale and dorian, the two modes this panel has already covered, and a different technical problem too: the scale repeats every minor third, so learning it on one root hands you three more roots for free. Key is F, the last of the twelve still untouched by this panel, and it lands on exactly the root the exercise wants: F7 is the V of B♭, and the point of the scale is what it does resolving home.
5 minutes: the F half-whole scale, one octave, hands separately, slow enough to name each interval as it's played — "half, whole, half, whole" — four times up and down. Then play the same eight-note shape starting a minor third higher, on A♭. It is the identical set of pitches in a different order, which is the whole reason this scale only needs three fingerings to cover twelve keys; feeling that before adding tempo is the actual warmup, not the notes themselves.
Two octaves, both hands together, quarter = 72, three consecutive clean passes ascending and descending with no stop between them. Success criterion: every one of the four altered tones — ♭9, ♯9, ♯11, and the natural 13 — speaks as clearly as the plain chord tones around it, with no audible rush or swallow at the two half-step turns per octave. A pass where the half-whole spacing drifts toward an even, all-half-step diminished run doesn't count toward the three.
Build the ii–V–I this scale is actually for: Cm7 for two beats, F7 for two beats using the half-whole scale as a single descending run from the 9 down to the root, then land squarely on a Bb major scale, ascending, on the downbeat of the I chord. Repeat it in both directions — descend into the V and ascend out of the I once, then the mirror. The target is the seam: the altered tension should evaporate exactly on the barline, not a beat late.
Stop the moment the four altered tones start blurring into a plain diminished-seventh arpeggio instead of a scale — that collapse is fatigue trading a symmetric scale for the easier symmetric chord buried inside it, and reps built on the chord train the wrong shape.
[2026-09-26] · src —
hearJoe Henderson — Inner Urge · Tidal · Spotify
Today's technical target is entirely about the half-whole diminished sound holding together under tempo, and "Inner Urge" is the tune jazz players reach for first when that sound is the point rather than a passing color — Henderson's own composition, built on shifting dominant harmony that his solo treats with the diminished scale's angular, symmetric lines rather than the more familiar diatonic alternatives. Listen for how little the symmetry reads as a "scale exercise" in his hands — the same eight-note shape practiced here in isolation, played as music rather than as a run.