Science
267 billion years? How to tell a breakthrough from a beautiful wrong answer
A telescope opened its eye, and there they were: galaxies that, by every textbook, should not yet exist — too big, too bright, too finished for a universe only a few hundred million years into its life. So a physicist did the boldest thing imaginable. He doubled the age of everything. Twenty-six point seven billion years, the headline said. Not 13.8. And here is the question that should haunt you when a claim like that goes viral: how do you tell a real breakthrough from a beautiful wrong answer, without a physics degree?
You don't need one. You need a method — and the best one I know was written for schoolchildren stuck on geometry problems. George Pólya's How to Solve It lays out four steps: understand the problem, devise a plan, carry it out, and — the step almost everyone skips — look back. Most people think solving is finding an answer. Pólya knew better. The answer is a hypothesis; looking back is where you find out whether it's true.
An answer that fits one square
Start by understanding the actual problem, the way Pólya insists: what's the unknown, what's the data, what are the conditions? The James Webb Space Telescope handed astronomers a genuine anomaly — early galaxies that look surprisingly mature and massive for their age. Rajendra Gupta, an adjunct professor at the University of Ottawa, devised a plan to absorb it. In a paper in Monthly Notices of the Royal Astronomical Society, he proposed a hybrid: take Zwicky's old "tired light" idea — that light simply loses energy crossing space — and bolt it onto "evolving coupling constants," the notion that the fundamental constants of nature drift over cosmic time. Stretch the clock with both, and galaxies suddenly have billions of years, not millions, to grow up. The math runs. The anomaly dissolves.
The whole condition, not the convenient one
Here is where Pólya's fourth step earns its keep, with the question that decides everything: does it agree with everything else you know? Pair that with his other quiet demand — did you use all the data, the whole condition, not just the part that was easy to satisfy? Because the age of the universe is not measured one way. It's pinned down by the cosmic microwave background, the faint afterglow of the Big Bang itself. It's corroborated by the oldest stars and globular clusters, which clock in around 13-plus billion years. It's cross-checked against the expansion rate and against Big Bang nucleosynthesis — the recipe that set the cosmic abundance of hydrogen and helium. Four independent witnesses, agreeing on 13.8.
Gupta's model was engineered to explain one anomaly. But a doubled age cannot quietly fix the galaxies and leave the rest untouched. It must also re-explain the microwave background, the stellar ages, the element abundances — and it doesn't. Think of it like a Sudoku. A number that completes one square can feel triumphant, inevitable, clearly right. But if that same number breaks three rows and a column, it is simply wrong, however elegant it looked in the square. That is the test the figure below traces: odd galaxies, a daring guess, a perfect fit to the one thing it was built for, and then collapse against everything it forgot to check.
"But it got published in a real journal"
I can hear the objection, because it's the one everyone reaches for: it appeared in MNRAS, a serious astronomy journal — doesn't that make it accepted science? No. Publication means a claim is interesting enough to argue about, not that it survived the argument. Peer review is the door, not the destination. And the discrediting detail is sharp: "tired light" is not a fresh idea but an old, long-abandoned one. As Brad Tucker of the Australian National University put it, if light really lost energy that way, we would already have noticed — it would have to show up, uninvited, in countless other observations we've calibrated to the bone for a century. It doesn't. The silence is the evidence.
And to be fair — Pólya would insist on this — proposing is not a crime. He distinguishes a plausible conjecture from a proof, and a scientist is entirely allowed to explore a wild conjecture, to chase a stretched clock and see where it leads. Gupta proposed; that is honest work. The proposal simply hasn't passed the looking-back checks. Exploring is permitted. Adopting is earned.
What this means for you
The transferable skill is a single change of question. When a stunning claim lands in your feed — a cosmology that rewrites the calendar, a supplement that cures the incurable, a tidy theory of who really runs the world — your instinct is to ask, "What does this explain?" Stop. Ask instead: What does this have to NOT break — and does it? That one inversion is the whole of Pólya's looking-back, and it travels everywhere. Run any viral claim through three quick gates. First: does it fit all the evidence, or only the one anomaly it was built to explain? Second: who else has checked it, and what did the silence or the chorus say? Third: to believe this, how many well-tested things would I have to throw away — and is the trade honest? A claim that buys one mystery by discarding ten settled facts isn't a bargain. It's a Sudoku square that breaks the board.
The whole-board test
A real answer fits every row and column you already trust; a clever wrong one solves a single square and quietly breaks the rest.
Method drawn from George Pólya, How to Solve It — his four steps (understand, plan, carry out, look back), the looking-back questions "does it agree with everything you know?" and "did you use all the data / the whole condition?", and his distinction between a plausible conjecture and a proof. Cosmology: Rajendra Gupta's ~26.7-billion-year model in Monthly Notices of the Royal Astronomical Society, built from Zwicky's discredited "tired light" plus evolving coupling constants, motivated by JWST's surprisingly mature early galaxies; the consensus 13.8-billion-year age rests on independent measurements — the cosmic microwave background, the oldest stars and globular clusters, the expansion rate, and Big Bang nucleosynthesis — and Dr. Brad Tucker's point that real "tired light" would already have shown up elsewhere. Popular-science commentary.