The only person in the building who thinks the floor is fake.
And you have to keep standing on it anyway — every day, using the same instruments and journals as everyone who isn’t. This lab isn’t about cosmology. Cosmology is just the vessel. It’s about the swing.
The swing
Picture one scientist picking up a sledgehammer and aiming it at the load-bearing pillar of their entire field — the assumption everything else is bolted to. Swing and connect, and the building comes down. Swing and miss, and you’ve spent your one life vandalizing a wall that was holding the roof up.
The cruelty is that from inside the moment, the vandal and the structural engineer make the exact same swing. Same temperament, same certainty, same instruments. The conviction is almost involuntary — once you think you’ve seen the thing, you can’t un-see it. Whether that ends in prophet or fool is decided by reality, later, often after you’re gone.
The bet is rigged both ways
You wasted your one life on a delusion, and you took students, funding, and reputations down with you. The field was right to resist.
You may not live to see it. Wegener died on the Greenland ice ~30 years before plate tectonics was accepted. Boltzmann hanged himself before atoms were settled science. Zwicky was “the crank” for ~40 years before dark matter caught up to him.
So this lab does something uncomfortable: it refuses to tell you who’s right. Instead it puts three real figures on the table who bracket the whole cruel space — and then it puts you in the chair.
The three who bracket it
- The live swing — a physicist swinging right now, this week. Crank or Copernicus, genuinely unknown.
- The proven swing — a man who was provably, card-flippingly right… and got robbed anyway.
- The gatekeeper — the one who did the resisting, who was both the necessary defense and the corruptible villain.
The one honest posture
- Weight the sources. Look for convergence. Update when the evidence moves.
- Admit what you can’t know — out loud, including to yourself.
- There is no bedrock to stand on while you decide. Truth here is provisional and probabilistic; you work with incomplete maps in hand. That’s not a failure of the method. That is the method.
Francesco Sylos Labini takes a swing at the cosmological principle.
On Wednesday, 24 June 2026, Nature published a paper arguing the universe has special directions — that it is not the same no matter which way you look. If it holds, it cracks the founding assumption under almost everything. It is real, peer-reviewed, and fiercely contested. We are not going to pretend he’s right.
The claim, plainly
The cosmological principle says the universe is homogeneous (every patch holds about the same amount of stuff) and isotropic (no direction is special). It is the mathematical premise that lets cosmologists write one equation — the FLRW metric — for the whole cosmos. Inflation, dark energy, the age of the universe, the curvature: all of it assumes the floor is flat.
Using the Dark Energy Spectroscopic Instrument (DESI) — a map of roughly 47 million galaxies across ~11 billion light-years — Sylos Labini and Marco Galoppo report coherent anisotropic structure persisting out to gigaparsec scales: structures large and directional enough that, they argue, a universe with no preferred direction shouldn’t produce them.
The method — ADPD, and why the slice is everything
Their statistic is the Angular Distribution of Pairwise Distances (ADPD) — parameter-free: measure the angles formed between pairs of galaxies. A uniform spread of angles means isotropy; preferred angles mean a direction. But ADPD is a 2-D statistic, so it requires slicing the survey into thin, quasi-flat sheets. And that is exactly where the smell is.
Travis’s first instinct — “did he handpick a subset?” — is sharp and half-right. They did not run this on all 47 million galaxies. The headline analyses are thin slices: a BGS sub-sample of 36,290 galaxies; another slice of ~27,685 galaxies, about 40 Mpc/h thick. Tens of thousands sliced out of tens of millions. The selection step is the crux.
Any thin slice looks stringy. So what?
Slice the cosmic web thin enough and it always looks filamentary — the famous 1986 “Slice of the Universe” (the CfA Stick Man) made that obvious 40 years ago. A stringy slice proves nothing by itself. The entire claim rests on one move: comparing each real slice to a ΛCDM mock slice with the same geometry and the same number of galaxies, and asking whether the real one carries more coherent directional structure than the standard model can fake. Drag the slice. Flip real vs. mock. Watch how easily a direction appears either way.
Where the reviewers would push hardest
This passed real peer review — Nature’s referees were Ruth Durrer and Kostas Migkas, both serious cosmologists, and the code and data are public (GitHub + Zenodo), so the cherry-pick question is directly checkable, not faith-based. The two sharpest questions live in the mock comparison:
The questions that decide it
- How many mocks? One “representative” mock is not enough. You need many realizations to know the real variance of ΛCDM before you can call the real slice an outlier.
- Chosen before or after looking? If slice orientations and thicknesses were picked after seeing the data, the look-elsewhere effect manufactures signals out of noise.
What keeps it out of the fringe bin
- The disagreement with the mainstream is quantitative — where does the universe smooth out, ~100–260 Mpc/h vs. his ~400+ — not flat-earth.
- It’s a 30-year research program with Luciano Pietronero (the “fractal universe”), not a fresh anomaly. That doesn’t make him right. It means it needs independent hands and standard methods — which the public code finally allows.
The wall on the other side of the table
Meanwhile, a great deal of independent evidence says isotropy holds, and the working cosmologists were quick to say so on the record:
Beyond the quotes, the standing evidence for a smooth universe is real: the near-perfect isotropy of the cosmic microwave background; the uniformity of the cosmic X-ray background; the absence of the spectral distortions and kinetic Sunyaev–Zel’dovich signatures a genuinely lumpy universe should leave; and SDSS / Planck analyses that find a transition to homogeneity at ~100 Mpc/h with no support for endless fractality. Katherine Freese (UT Austin) put the stakes simply: it could challenge “the basic scaffolding for the universe that we all assume in our work” — if it survives.
Graded honesty — the stakes scale with how true it is
Mild version
- Isotropy is approximate; the homogeneity scale is just bigger than the textbook number. Annoying, not apocalyptic — recalibrate, and dark energy survives. Many cosmologists already accept small departures.
Strong version
- His fractal claim: no transition to homogeneity, coherent anisotropy all the way to a gigaparsec. Then the FLRW metric is wrong, “dark energy” could be partly an artifact of fitting a smooth equation to a clumpy cosmos, and precision cosmology gets a much bigger error bar. A genuine paradigm break.
That spread is why the evidence bar is brutal. A claim that would break the foundation has to be far stronger than a claim that just nudges a number. Right now it sits in the open — reviewed, published, public, and not yet believed.
Dmitri Mendeleev predicted elements that didn’t exist yet.
This is the purest commit-before-reveal in the history of science. He left blank squares in his table and told the world what would fill them — weights, densities, properties — for elements nobody had ever seen. Then reality flipped the cards, one by one, his way. And he still didn’t get the Nobel.
The commit, and the reveal
Mendeleev publishes the periodic system and leaves gaps. He commits hard predictions for “eka-aluminium,” “eka-silicon,” and “eka-boron” — elements that did not exist on any shelf.
Gallium is discovered — eka-aluminium. Mendeleev corrects the discoverer on its density, from his prediction. He is right.
Scandium (eka-boron) and germanium (eka-silicon) turn up, matching the predictions. Reality flips card after card. The blank squares fill in.
This is the dream case. Not “maybe a Copernicus” — demonstrably a Copernicus, confirmed in his own lifetime. So the obvious lesson would be: be right, and you’ll be honored. Except that’s not what happened.
Mendeleev was nominated for the Nobel Prize in Chemistry nine times across 1905–1907. In 1906 the Nobel Committee for Chemistry voted 4–1 in his favor and recommended him to the Royal Swedish Academy. The Academy — which has the final say — refused. It added four more members to the committee and made it vote again. The re-stacked vote came back 5–4 for Henri Moissan (for isolating fluorine). The Academy took that.
Then he ran out of time. Mendeleev died in February 1907, of influenza, at 72 — and there is no posthumous Nobel. As his predicted elements kept showing up, the snub curdled into what’s now called one of the Academy’s biggest mistakes.
Which raises the obvious question: who, exactly, stacked that committee?
Footnote, present day · Mendeleev left blank squares for elements that didn’t exist yet. A century and a half later, a thirteen-year-old built her own periodic table and deliberately left slots gold-less — fifteen written, room kept for the sixteenth. Same move: the gap is the point. → what jackie thinks
Svante Arrhenius wore both coats at once.
Someone has to guard the pillar. Most heretics are wrong, and a field with no gatekeepers is a field that believes anything. Arrhenius was that guard. He was also, in at least one famous case, a man who may have used the gate to settle a score — and from inside the room, you cannot tell those two things apart.
The case against Mendeleev
Arrhenius was not even on the Chemistry committee — but he was enormously influential in the Academy, and he pressed hard for Mendeleev’s rejection. His stated argument: the periodic system was “too old” to honor in 1906, having been published back in 1869. (A dissenting committee member, Peter Klason, put Moissan’s name forward; Arrhenius supplied the weight.) The two 1907 nominations were frustrated again by his opposition — right up until Mendeleev died.
Not a one-off — a pattern
Arrhenius also worked against Kristian Birkeland, whose theory that charged particles from the Sun drive the aurora was correct and decades ahead of its time. Birkeland was nominated for the Nobel seven or eight times and never won; Arrhenius, who considered auroral physics his own turf, was among those who kept him out. Gatekeeping-as-grudge looks, in the record, like a habit.
The complication that makes it honest
Here is what keeps this from being a clean villain story: Arrhenius was no fool. He was a Nobel laureate himself (Chemistry, 1903). He was the first person to calculate how carbon dioxide warms the planet — the foundation of climate science. The gatekeeper and the heretic can be equally brilliant.
Resisting the swing was right…
- Most challenges to a foundation are wrong. A field that honored every bold claim immediately would be a field with no spine. Skepticism is the immune system. Arrhenius doing his job was, most of the time, the system working.
…and could still be a crime
- The exact same posture — “this isn’t proven enough, not yet” — can carry principled caution or a powerful ego settling a score. They wear the same coat. No clean heroes, no clean villains, no answer key.
Would you swing?
Here is the live case, stripped to its bones. You are going to commit a position before you see how it turns out — which is the only honest way to do it, because the real world hasn’t turned out yet either. There is no answer key in this tab. There isn’t one in reality this week.
A 30-year horn blown against the wind. A real Nature paper that passed real referees. A signal pulled from thin slices — tens of thousands of galaxies out of tens of millions. Public code, so the cherry-pick question is checkable. And a wall of independent evidence — the CMB, the X-ray background — saying no, loudly, from the other side of the table.
There’s no key. Here’s why that’s the point.
You committed before the reveal. But the reveal isn’t “you were right” or “you were wrong” — because nobody alive knows. The paper is days old. The verification hasn’t happened. The real experts split exactly the way a healthy field should:
The second twist — the Mendeleev clause. Suppose you somehow knew you were right. Mendeleev knew. The cards flipped for him and the crown still never came. So the real question was never just “are you right?” It was: would you swing anyway — knowing you might be wrong and waste your one life, or be right and never live to see it, or be right and simply get robbed by someone wearing the gatekeeper’s coat?
NULL the Penguin watched the scientist swing, and the crowd step back, and nobody — not the crowd, not the scientist, not NULL — able to say yet which one history would keep. NULL said nothing. NULL marked the date and kept watching.