Sorted by lag. Three of these seven were news that week only in the sense that somebody re-ran them — the median gap between publication and coverage was 36 days, and exactly one item was same-day. That distribution is why this digest is weekly and not daily.
Verified
Published 2026-09-02
Journal Science Advances
Read the paper, in full
Lag 0 d
Saturn grew a ten-sided pole, and the famous six-sided one stopped being a law
Saturn's northern hexagon — a jet stream bent into six straight sides, wider than Earth — has been in every image anyone has taken for more than forty years. That durability is exactly why it read as a rule about how gas giants behave. Hubble has now tracked a decagon, ten sides, encircling the south pole: the first large regular-sided jet pattern seen in Saturn's southern hemisphere. It has been visibly developing across 2023, 2024 and 2025, and the authors offer two candidate causes — an instability of the jet itself, or a 4,000-km red anticyclone at 55°S — without being able to choose between them.
The star stuff frame
One planet, one atmosphere, one set of physics — and two poles that resolved it differently, six sides at the top and ten at the bottom. Neither is the standard; the hexagon only looked like the standard because it was the one we had measured. Forty years of stability was mistaken for a rule, and the correction came from looking at the other end of the same object. The universe doesn't pathologize its own variation — and here it declines even to standardise a planet against itself.
Where it goes
Glimmers, gold accent. It also pairs hard with
No. 50,
The Sky Was Not Regular Enough — that piece is about the sky failing to be steady enough to keep time by; this is the sky failing to be regular enough to generalise from. A companion note, not a rewrite.
Sánchez-Lavega, A., et al. “A decagon wave around Saturn's south pole.” Science Advances 12, eaee4251, 2 September 2026. DOI 10.1126/sciadv.aee4251, CC BY 4.0. Upgraded to verified 2026-09-02, when the paper itself was obtained and read in full. The figures the NASA release did not carry are all in it: latitudes 58°S–63°S, mean zonal wavelength 16,782 ± 1,100 km, drift +2.5 ± 0.2 m/s against a 116 m/s jet, vertex oscillation 32.5 d at 4.6°–8.4°. Two corrections to what was filed here first: “nobody knows why” overstated it — the authors name two candidate mechanisms and say they cannot discriminate; and the formation window is bounded by the south pole being unobservable from mid-2012, not 2017 as some coverage has it. Now a piece: Until There Were Ten.
Contested
Announced 2026-08-28
Journal Astrophysical Journal Letters
Read Portsmouth release
Lag 5 d
The clock on “you are made of star stuff” may start 13.7 billion years ago
The first stars were enormous, and some ended as pair-instability supernovae — explosions that can throw out more than a hundred solar masses of heavy elements at once. Simulations from Portsmouth follow that debris as it settles around a later, smaller star, roughly 70% of the Sun's mass, and find a water-rich disc capable of building planetesimals: the precursors of rocky worlds. Carbon, oxygen, iron. Water at a few times less than the Solar System had. All of it about 100 million years after the Big Bang, before the first galaxies had finished assembling.
The star stuff frame
This is the founding fact of the whole collection with its date moved. The atoms in a body were forged in dying stars and scattered — and the scattering, and the assembling of them into rock and water, may have begun almost immediately. Sagan's line stops being a story about the late universe and becomes a description of how it worked from the start. If it holds, the ingredients were never scarce and never late.
Why contested, not plausible
This is a simulation chain, and the release describes one example disc meeting the criteria. Coverage elsewhere is already framing it as a debate. The gap between “planetesimals could form” and the habitability language in the headlines is wide, and it is the kind of gap this site exists to hold open rather than close.
Where it goes
Glimmers if the paper supports the narrow claim, with the wide claim named and refused on the page. It is also the strongest zine candidate here for
Star Stuff proper — one settled-enough fact, followed honestly, and the belonging claim already inside it.
Jessop, C., & Whalen, D. J. “Planet Formation at Cosmic Dawn: Planetesimals in H2O-Rich Disks Around Low-Mass Stars.” The Astrophysical Journal Letters, 2026. Read: University of Portsmouth. Not read: the paper. Not obtained: DOI, full author list, redshift or metallicity thresholds. The title came from a search summary rather than the release — confirm it before quoting.
Plausible
Published 2026-08 (epub)
Journal Autism
Read listings only
Lag ~7 d
Somebody asked Autistic adults what we are good at, in our own words
Autism research overwhelmingly measures deficits; instruments ask what is hard and rarely what is there. This study put open-ended questions to 127 legally independent Autistic adults and coded what they said about their own strengths and talents. The answers spread across problem-solving, creativity, character, academics and getting along with others — and close to a third named good people skills, which is the specific thing the literature is most confident we lack. The authors argue strengths should be measured during assessment, not just difficulties.
The star stuff frame
A spiky profile has two sides, and the field has instruments for one of them. This is not a claim that we are secretly fine; it is a claim about who holds the pen. A third of a sample reporting social strength, inside a paradigm built on social deficit, is variation showing up the moment the question changed — and
difference-first is the frame this whole site is built on.
Handle carefully
Self-report from 127 people recruited through SPARK — a genetics research registry some of our community declines to join, which makes the sample selected in a way any use of it has to state. And the finding writes itself straight into the superpower framing if left alone. That trap is already named across the collection; a piece here would have to name it again.
Where it goes
Glimmers is a stretch — this is a paper, not an occasion. Better as a card in the neurodiversity field guide, or the evidence spread of a zine about who gets to write the instrument.
Lampinen, L. A., Singer, J., Wang, X., VanHook, B., Wilkinson, E., & Bal, V. H. “Self-reported strengths and talents of autistic adults.” Autism, 2026. DOI 10.1177/13623613251364361 · PMID 40856551. Read: journal and PubMed listings, plus a SPARK research-match summary. Not read: the abstract in full, or the paper. Recruitment, coding method and limitations all need checking at source. Note the house rule: Autistic is capitalised as an identity term, and the paper's own lowercase title is quoted verbatim.
Contested
Published 2026-07-28
Journal Nature Medicine
Read both papers, in full
Lag 36 d
The largest fibromyalgia study ever run found something — and much less than the headlines said
A multi-ancestry meta-analysis across 2,563,755 people — 54,629 cases, 87.7% of them female — reports 26 risk loci, heritability enriched exclusively in brain tissue and neural cell types, and genetic correlations above 0.7 with eighteen conditions including low back pain, PTSD and IBS. The strongest cell-type signal is the hippocampal dentate gyrus, which the authors connect to the symptom patients call brain fog. For a condition routinely met with there is nothing wrong with you, that is not a small thing.
And a second study of comparable size, published the same day in Nature Communications by a different group, reaches materially different conclusions on two of the headline questions. It is not a companion paper. It is a second opinion.
The star stuff frame
The good part is real: pain that has been treated as fabrication has a genetic architecture, and it sits in the nervous system. But the finding worth the most here is the one nothing we read reported. The Nature Medicine group found the genetic architecture to be the same in men and women — inter-sex correlation 1.03, statistically indistinguishable from perfect — while 87.7% of the diagnosed are women. Their own conclusion: the prevalence gap is “unlikely to be driven by sex-specific genetic risk variants, and may instead reflect nongenetic biological factors, environmental exposures or diagnostic bias.” That is the authors, in Nature Medicine, putting who gets believed on the list of explanations. The other paper puts that correlation at 0.73, so even this is contested — but it is the question worth asking.
And the belonging claim was never waiting on any of it. If a finding is what makes disbelieved people credible, the next null result un-makes them. Down to is not because of.
The critique, now checked line by line — and its central charge is misattributed
Five of the critic’s numbers hold, and every one of them is the paper’s own, stated plainly: the HTT variant carries OR 1.090 (“about 9% increased odds”), the polygenic score reaches AUC 0.59, and prevalence really does range sixfold across cohorts, 1.2% to 7.5% — disclosed in the second sentence of the Results. One number is wrong: the critique says heritability explains “less than 10%”; the paper reports 10.4% (95% CI 9.8–11.0). One is unresolved: we could not confirm the one-billing-code claim without the Methods supplement. And the central charge misses. The critique frames “a central nervous system disorder” as a headline exceeding the evidence. It is the paper’s own abstract — “This study provides robust genetic evidence defining fibromyalgia as a central nervous system disorder” — said twice, alongside two named therapeutic targets. That may still be over-reach, but it is a scientific disagreement rather than a science-communication failure, and the two need different treatment.
Two papers, one condition, one day, two answers
Read side by side, the Nature Medicine study and Bright et al. in Nature Communications (85,139 cases; Yale and the VA Million Veteran Program) disagree where it matters most. On sex: an inter-sex genetic correlation of 1.03 and “no sex difference in the genetic architecture”, against 0.73 ± 0.064 — significantly short of identical. On autoimmunity: Nature Medicine finds no MHC signal and no immune-cell enrichment and concludes fibromyalgia is “not primarily an autoimmune disorder”; the Nature Communications structural model places it “mostly as a pain and autoimmune-related trait.” On heritability: 10.4% against 6.9%, measured on different scales. The coverage this item was filed from reported one of these papers and not the other.
Where it goes
Both papers are now read and the critique is checked, so the open items are closed — but the piece this points at has changed shape. It is no longer a
Too Good to Check story about press over-reach, because the strong framing is the authors’ own. Two candidates remain:
two studies, one condition, opposite answers, published the same day — or the quieter and harder one,
identical genetics and a diagnosis rate that is not. Editorial call, not ours.
“The genetic architecture of fibromyalgia across 2.5 million individuals.” Nature Medicine 32, 3060–3070, published online 28 July 2026; received 18 September 2025, accepted 28 May 2026. DOI 10.1038/s41591-026-04492-6. And Bright, U., Beck, S., Levey, D. F., Deak, J. D., the VA Million Veteran Program, Gaziano, J. M., Stein, M. B. & Gelernter, J., “The genetics of fibromyalgia and its relationships to psychiatric and medical traits,” Nature Communications 17:6248, 28 July 2026. DOI 10.1038/s41467-026-75256-6. Both read in full on 2026-09-02, from copies supplied rather than fetched — every publisher route refused on the compile date. The grade stays contested, for a better reason than before: not because we could not check it, but because two studies of this size disagree in print.
Plausible
Published 2026-07
Journal Neuron
Read Columbia coverage
Lag ~37 d
Taking four to eight years to finish is the mechanism, not the delay
Microglia are the brain's resident immune cells, and they do developmental work: pruning synapses, tuning how circuits respond. In a mouse they reach maturity in about three weeks. In a human they take four to eight years. Columbia's group traces that to human-specific duplications of the gene SRGAP2 — reportedly around ten times more abundant in microglia than in neurons — which slow the maturation of both cell types. The slowing is synchronised: glia and neurons kept on the same long clock, and the group's reading is that the long clock is what human cognitive complexity is built on.
The star stuff frame
One gene, doing the same thing to two different cell types so they stay in step, and the result of the delay is capacity. That is neoteny read as an achievement rather than a deficit, and it lands directly on a word our children get measured with. Developmental delay describes a schedule — and this is a species whose signature trick is being years behind the mouse on purpose.
Where it goes
A numbered
Star Stuff zine rather than a Glimmers entry: one settled mechanism, followed until the belonging claim is already in it. Neoteny is
not unspent here — check what
The Play You Keep and
Thirteen Changes already argue. The unspent part is the
synchrony: that being slow
together is the point.
Diaz-Salazar, C., … Polleux, F. “Human-specific SRGAP2 paralogs synchronize neotenic microglial maturation and synaptic development.” Neuron, 2026. Article S0896-6273(26)00536-2 · preprint bioRxiv 10.1101/2024.06.28.601266. Read: Neuroscience News and Columbia Zuckerman Institute coverage. Not read: the paper — Cell returned 403. The “ten times more abundant” figure and the four-to-eight-year range come from that coverage, not the paper.
Plausible
Published 2026-05-21
Journal Science Advances
Read HKU release
Lag 104 d
Diamond was ruled non-piezoelectric for a century. Its irregularities turned out to be the mechanism
Squeeze a quartz crystal and it makes a voltage; that is piezoelectricity, and it requires a crystal structure without a centre of symmetry. Diamond has one, so for over a century diamond was simply not a piezoelectric material. A Hong Kong group used edge exfoliation to make diamond membranes thin enough to bend, bent them, and measured stable, repeatable voltages. Their calculations put the cause in the grain boundaries — the messy joins between crystal domains in a polycrystalline film. Charge polarisation gathers there as the bending increases, leaving a potential difference between the two faces. Named applications include self-powered implants and deformation sensors.
The star stuff frame
This is the Yasuda line arriving from the opposite direction. Bone is piezoelectric — pressure into charge, the fact this whole collection stands on. Here the hardest material there is turns out to do the same thing, and it does it because it is not uniform. A perfect single crystal of diamond stays silent. The signal exists only where the structure is irregular. Difference is not tolerated in this system; difference is the transducer.
Numbers not confirmed
A five-micrometre membrane holding output across more than 7,000 bending cycles at 0.35% strain: those figures came from a search summary. The HKU release carries the mechanism and no numbers, and Science Advances would not open. No piezoelectric coefficient was obtained at all. Don't print any of these until the paper is read — and mind the unit, because µm is micrometres and a stylesheet that uppercases it turns the figure into millimetres and a thousandfold error.
Where it goes
The frame is too good to spend on a recirculated item without the paper. This is the companion to
Bone Song — a zine whose one fact is that asymmetry is what makes the charge.
Jing, J., et al. “Uncovering piezoelectric effect in polycrystalline diamond membranes.” Science Advances, 21 May 2026. DOI 10.1126/sciadv.aea8318. Group led by Zhiqin Chu and Yuan Lin, University of Hong Kong. Read: the HKU press release. Not read: the paper, the full author list, or any measured value. Surfaced by a news feed on 29 August 2026, 100 days after publication.
Plausible
Published 2026, vol 12 no 17
Journal Science Advances
Read Georgia Tech release
Lag ~130 d
There is no lizard inside your head, and 182 species say what is there instead
The triune brain — a reptilian core for survival, a limbic layer for emotion, a rational neocortex stacked on top — was proposed in the 1950s and has been contradicted by comparative neuroanatomy since the 1970s. It is also still in circulation, and appeared in introductory psychology textbooks as a real model as late as 2017. This analysis across 182 species finds the limbic system and neocortex do not stack; they trade. The two are wired on different principles — spatially organised circuits against distributed networks — and expand and contract together as an animal's life demands. When limbic regions grow, cortex tends to give ground.
The star stuff frame
This is the
How We Got Here motive in one finding. The triune brain is good science's clothes on a bad idea: it has a neuroanatomist's name on it, it produced a diagram everybody can draw, and pointed at people it does exactly the flattening that collection was built to answer — it ranks parts of a nervous system by how recently they arrived and calls the older ones primitive. Every appeal to somebody's
lizard brain descends from it. And the replacement is not a ladder at all: it is a trade-off, with no top.
Where it goes
The best chain candidate in this edition, and unspent — nothing on this site says triune or lizard brain. MacLean's proposal, the 1970s contradictions, the textbook survival to 2017, the pop-neuro descendants, and this paper: one link per spread, joints marked documented / contested / leap. It is an argument for the method and against a model that borrowed its authority — if a draft starts reading as anti-science it has gone wrong.
Imam, N., Kielo, M., Trude, B. M., & Finlay, B. L. “Dual computational systems in the development and evolution of mammalian brains.” Science Advances 12(17), 2026. DOI 10.1126/sciadv.aec6112. Read: a news release citing the paper, plus Georgia Tech Research News (29 June 2026). Not read: the paper. Date approximate — the issue number implies spring 2026 and was not confirmed; get the real date before it goes on a page. Barbara Finlay is a long-standing critic of the triune model, which is context rather than a caution.