The Number That Broke Infinity — And Still Won't Be Caught
Humanity's 4,000-year pursuit of a number that goes on forever
Every civilization that ever looked at a circle felt compelled to chase this number — and none of them could catch it.
Pi has been computed to 314 trillion digits — a number so vast it would take roughly ten million years to recite aloud — yet we are no closer to "finishing" it than the Babylonian scribe who scratched 3.125 into a clay tablet four thousand years ago. That is the central absurdity of humanity's longest-running mathematical obsession. Pi is not merely a number. It is a mirror, reflecting back our compulsive need to measure what cannot be measured, to complete what provably cannot be completed, and to find order in the fabric of reality itself.
What makes this story remarkable is not just the mathematics. It is the characters: an ancient Egyptian scribe named Ahmes, a self-taught genius from southern India who died at thirty-two, a pair of brothers who built a supercomputer in their Manhattan apartment, and a physicist at the San Francisco Exploratorium who invented a holiday around a constant. Pi weaves through clay tablets and quantum mechanics, through Congressional farce and the most beautiful equation ever written. It connects Archimedes to Google Cloud. And after four millennia of pursuit, it still will not sit still.
A Clay Tablet, a Papyrus Scroll, and the First Algorithm
The earliest known approximation of π comes from an Old Babylonian clay tablet dating to roughly 1900–1600 BCE, excavated near Susa in modern Iran. The tablet describes a geometric relationship that, when worked backward, yields π ≈ 25/8 = 3.125. Not bad for the Bronze Age — off by less than 0.6%. A common misconception links this to the famous Yale tablet YBC 7289, but that tablet actually concerns the square root of 2. The Babylonian π tablet is a humbler artifact, but it opens our story: somebody, somewhere in ancient Mesopotamia, looked at a circle and tried to pin down the ratio of its boundary to its width.
Around 1650 BCE, an Egyptian scribe named Ahmes copied eighty-seven mathematical problems onto what we now call the Rhind Papyrus. Problem 48 gives a recipe for finding the area of a circle: take the diameter, cut off one-ninth, and square the result. In modern notation, this makes the area of a circle with diameter d equal to (8d/9)², which implies π = 256/81 ≈ 3.1605. Whether Ahmes understood he was approximating a universal constant is debatable — he was probably just trying to measure grain silos. But his method was accurate to about 1%.
Then came the breakthrough. In his treatise Measurement of a Circle, Archimedes did something no one before him had done: he bounded π with mathematical certainty. He inscribed a regular polygon inside a circle and circumscribed another around it, then calculated their perimeters. Starting with a hexagon, he doubled the sides — 12, 24, 48, 96 — computing each step using only the Pythagorean theorem and angle bisection. With a 96-sided polygon, he proved that:
This pinned π between 3.14085 and 3.14286 — not an estimate, but a guarantee, backed by rigorous logic. Mathematicians often call it the first true algorithm: a systematic, iterative, extendable procedure that could achieve arbitrary precision if you had enough patience and papyrus. Think of it like binary search for a number that doesn't exist as a fraction — you keep tightening the bounds, and the target keeps slipping between them. Every digit-hunting quest that followed stands on Archimedes' shoulders.
The baton passed east. In 263 CE, Chinese mathematician Liu Hui independently developed his own polygon method, calculating areas rather than perimeters. With a 3,072-sided polygon, he reached π ≈ 3.14159 — five correct decimal places, surpassing Archimedes. Two centuries later, Zu Chongzhi (429–500 CE) pushed Liu Hui's algorithm to a polygon with 24,576 sides, proving that 3.1415926 < π < 3.1415927 — seven correct decimal places. He also produced the fraction 355/113, known as Milü ("close ratio"), accurate to six decimal places. It remains the best rational approximation of π with a denominator under 16,600. Europe would not rediscover this fraction until Adriaan Anthoniszoon stumbled upon it in 1585 — more than a thousand years later.
Most fractions with small denominators are lousy approximations of π. But 355/113 = 3.14159292... is accurate to the 7th decimal place. To do better, you need a denominator of at least 16,604. It is a "best rational approximation" in the strict mathematical sense — no fraction with a smaller denominator gets closer.
Perhaps the most underappreciated chapter belongs to Madhava of Sangamagrama (c. 1340–1425), a mathematician from Kerala, India. Madhava and his school discovered infinite series for π, including what we now call the Madhava–Leibniz series:
He also developed power series for sine and cosine and devised correction terms to accelerate convergence — sophisticated techniques that anticipated calculus by roughly 250 years before James Gregory and Gottfried Leibniz independently rediscovered them in the 1670s. Using these series, Madhava computed π to 11 correct decimal places. His original manuscripts are lost, but later Kerala school texts unambiguously credit him. It is one of the great "what if" stories in the history of science: what if Madhava's work had reached Europe a century earlier?
Why Pi Can Never Be Caught
For millennia, mathematicians held out hope that π might eventually resolve — that with enough polygon sides or enough series terms, the decimal expansion would terminate or settle into a repeating pattern. Think of it this way: the fraction 22/7 gives 3.142857142857..., where "142857" repeats forever. If π were rational, it would do something similar. Maybe the pattern was just really, really long?
In 1761, Swiss mathematician Johann Heinrich Lambert destroyed that hope. Using generalized continued fractions, he proved that if x is a nonzero rational number, then tan(x) must be irrational. Since tan(π/4) = 1 — which is rational — π/4 itself, and therefore π, must be irrational. Its decimals would never repeat. Never terminate. The pursuit was provably infinite.
But irrationality was just the first blow. In 1882, Ferdinand von Lindemann proved something far more devastating: π is transcendental. It is not merely irrational — it is not the root of any polynomial equation with integer coefficients, no matter the degree. The proof was elegant in its indirection: Euler's identity tells us that eiπ = −1. If π were algebraic, then iπ would also be algebraic (since i is algebraic, being a root of x² + 1 = 0). But Lindemann had proved that e raised to any nonzero algebraic power must be transcendental — and −1 is plainly not transcendental. Contradiction. Therefore π cannot be algebraic.
Rational numbers are roots of linear equations (ax + b = 0). The square root of 2 is a root of x² − 2 = 0 — irrational, but algebraic. A transcendental number is not the root of any polynomial with integer coefficients, of any degree. It sits outside the entire algebraic hierarchy. Think of algebraic numbers as the addresses on every street in an infinite city — transcendental numbers are the points between the addresses, and there are uncountably more of them than there are addresses.
This single proof annihilated a 2,300-year-old industry. The ancient Greek problem of "squaring the circle" — constructing a square with the same area as a given circle using only compass and straightedge — required constructing a line segment of length √π. Since only algebraic numbers can be constructed with those tools, and π is transcendental, squaring the circle is forever impossible. Hippocrates, Anaxagoras, and two millennia of would-be circle-squarers had been chasing a phantom. Professional mathematicians moved on. Amateurs, characteristically, did not.
There is a deeper strangeness still. Georg Cantor showed in 1874 that algebraic numbers are countable while real numbers are not — meaning almost every number on the number line is transcendental. They are the rule, not the exception. Yet proving any specific number is transcendental remains extraordinarily hard. We still don't know whether π + e or π × e is transcendental, though at least one must be. Pi is also conjectured to be a "normal" number — meaning every finite sequence of digits appears with equal frequency. If true, then buried somewhere in π's expansion is your phone number, every possible sentence ever written, and every possible arrangement of every possible sequence. But proving normality remains beyond reach.
Obsession Made Flesh: The Manual Computation Era
Knowing that π's digits stretched to infinity did not deter anyone. If anything, it sharpened the obsession.
A German-Dutch fencing instructor turned mathematics professor, van Ceulen spent much of his life computing π by Archimedes' polygon method, pushing to polygons with 262 sides — a number in the quintillions. By his death in 1610, he had reached 35 correct decimal places. He requested the digits be engraved on his tombstone at the Pieterskerk in Leiden. The original stone vanished around 1800; in 2000, a replica was unveiled. For generations, Germans called π the Ludolphsche Zahl — the Ludolphine number.
An English boarding-school proprietor who spent fifteen years computing π to 707 decimal places, finishing around 1873. He used John Machin's formula — π/4 = 4 arctan(1/5) − arctan(1/239) — calculating new digits every morning and checking them every afternoon. He published his results and basked in the achievement. It was the high-water mark of hand computation.
The brutal twist: In 1946, mathematician D.F. Ferguson took a desk calculator and an independent formula and quietly recomputed π from scratch. His result disagreed with Shanks starting at the 528th decimal place. Shanks had made a single error — likely omitting a zero — and every digit from 528 onward was wrong. Fifteen years of meticulous work, undone by one slip. Ferguson extended the computation to 808 correct digits. It was both the last great manual record and a testament to the merciless precision π demands.
The Man Who Saw Infinity
In January 1913, a twenty-five-year-old clerk at the Madras Port Trust earning twenty pounds a year wrote a letter to G.H. Hardy, the leading mathematician at Cambridge. The letter was nine pages of mathematical results — over 120 theorems on infinite series, continued fractions, number theory, and integrals — presented without a single proof. Hardy's first instinct was to dismiss it as the work of a crank. He set it aside. Then, that evening, he picked it up again.
They must be true, because, if they were not true, no one would have the imagination to invent them.
— G.H. Hardy, on first reading Ramanujan's theoremsHardy showed the letter to his collaborator J.E. Littlewood. They spent hours examining the results. Some were already known. Some were clearly wrong. But many were, in Hardy's words, results that had defeated him completely — he had never seen anything like them. Bertrand Russell wrote that he found Hardy and Littlewood in a state of wild excitement because they believed they had discovered a second Newton — a Hindu clerk in Madras.
Srinivasa Ramanujan (1887–1920) had taught himself mathematics almost entirely from a single book — George Shoobridge Carr's Synopsis of Elementary Results in Pure and Applied Mathematics, a compendium of 5,000 theorems with minimal proofs. He had failed out of college twice because he refused to study anything but mathematics. He was deeply religious, crediting the goddess Namagiri with his insights: "An equation for me has no meaning unless it expresses a thought of God."
Hardy brought Ramanujan to Cambridge in 1914. That year, Ramanujan published a paper containing seventeen extraordinary series for computing 1/π, including:
Each term of this series yields approximately eight correct decimal digits. The very first term (k=0) alone gives six digits. In 1985, William Gosper used this exact formula to compute 17 million digits of π — the first computation to rely on Ramanujan's work, more than seventy years after it was published. The formula had been sitting there, waiting for machines fast enough to exploit it.
When Hardy was asked to rate mathematicians on a scale of pure talent from 0 to 100, he reportedly gave himself 25, Littlewood 30, the great David Hilbert 80 — and Ramanujan 100.
Ramanujan's health collapsed in England. A strict vegetarian Brahmin in wartime Britain, poorly nourished and far from home, he developed what was diagnosed as tuberculosis — though modern analysis suggests hepatic amoebiasis, a treatable parasitic infection his English doctors never thought to investigate. He returned to India in 1919 and died on April 26, 1920, at age thirty-two. He left behind notebooks containing nearly 3,900 results, almost all of which have since been proven correct.
In 1976, mathematician George Andrews discovered Ramanujan's "lost notebook" — 138 pages of formulas from his final year — in a box at Trinity College Library, Cambridge. The discovery was compared to finding Beethoven's tenth symphony.
The Man Who Knew Infinity — Robert Kanigel's 1991 biography told Ramanujan's story to a wide audience. The 2015 film starring Dev Patel as Ramanujan and Jeremy Irons as Hardy brought it to millions more. Hardy's own reflection may be the most poignant epitaph: "My association with him is the one romantic incident in my life."
The Chudnovsky brothers, David and Gregory, built directly on Ramanujan's work. In 1988, they published a refined formula where each term produces roughly fourteen digits — nearly double Ramanujan's rate. Their algorithm, based on deep connections to modular forms and the Heegner number 163, has been used for virtually every π world record since 2009. They famously computed over a billion digits on a homemade supercomputer assembled in their Manhattan apartment. Every time a new trillion-digit record is announced, Ramanujan's ghost is in the machine.
When Silicon Joined the Hunt
In September 1949, the ENIAC — one of the world's first electronic computers — calculated π to 2,037 decimal places in about seventy hours. John von Neumann directed the project, partly to test the machine's reliability and partly to investigate whether π's digits were statistically random. It was the beginning of exponential acceleration.
The modern era belongs to y-cruncher, a program written by Alexander Yee that started as a high school project and became the engine behind every π record since 2010. The software uses the Chudnovsky algorithm with binary splitting, FFT-based multiplication, and disk-swap capabilities that let relatively ordinary hardware crunch extraordinary numbers.
Emma Haruka Iwao, a Google Cloud developer advocate who was inspired as a child by Kanada's records, computed 31.4 trillion digits on Pi Day 2019 — the first record set using cloud computing. She returned in 2022 to reach 100 trillion. StorageReview then pushed to 202 trillion in 2024 and, as of late 2025, the current record stands at 314 trillion digits — a number symbolically echoing 3.14 — computed on a single Dell server with 2.46 petabytes of Micron NVMe SSDs.
Here is the glorious absurdity: NASA's Jet Propulsion Laboratory uses exactly 15 digits of π for all interplanetary navigation. With 15 decimal places, you can calculate the circumference of a circle 20 billion miles across — the distance to Voyager 1 — and be off by only 1.5 inches. With 39 digits, you could measure the observable universe to the width of a hydrogen atom. We have computed five trillion times more digits than we will ever need. And we keep going.
The Number That Haunts Physics
Pi's appearances in physics extend far beyond geometry. It surfaces in places where no circle is in sight — like an uninvited guest who turns out to know everyone at the party.
The Basel Problem (Euler, 1734)
A twenty-seven-year-old Leonhard Euler stunned the mathematical world by solving a problem that had defeated the Bernoulli family for ninety years: what is the exact value of 1 + 1/4 + 1/9 + 1/16 + ...?
The sum involves only integers and their squares — pure arithmetic, no geometry — yet π² emerges as if from nowhere. This was the first glimpse of π's reach beyond circles and directly inspired Bernhard Riemann's zeta function and the still-unsolved Riemann Hypothesis.
Buffon's Needle (1777)
The Comte de Buffon published a proof that if you drop a needle of length L onto a floor ruled with parallel lines spaced d apart, the probability it crosses a line is 2L/(πd). This means you can estimate π by repeatedly dropping needles and counting crossings — a physical experiment that produces a transcendental constant. In 1901, Mario Lazzarini claimed to have done exactly this with 3,408 tosses, obtaining the suspiciously precise result 355/113. Statistical analysis strongly suggests he fabricated the data. Pi, apparently, punishes cheaters.
Pi Everywhere You Look
Einstein's field equations of general relativity contain the factor 8πG/c⁴. Coulomb's law places 4π in the denominator because electric fields radiate spherically — 4π steradians of solid angle. Heisenberg's uncertainty principle involves ℏ = h/(2π) because position and momentum are Fourier conjugate variables, and Fourier analysis is built on sinusoidal functions with period 2π. The normal distribution — the famous bell curve — carries a factor of 1/√(2π) because the Gaussian integral ∫e−x²dx from −∞ to ∞ equals √π. Even the period of a simple pendulum is T = 2π√(L/g), because oscillation is circular motion projected onto a line.
A friend spots the symbol π in a paper on population trends. "The ratio of the circumference to the diameter," the statistician explains. "Well, now you are pushing your joke too far," the friend replies. "Surely the population has nothing to do with the circumference of the circle."
— Eugene Wigner, "The Unreasonable Effectiveness of Mathematics in the Natural Sciences" (1960)And then there is the Mandelbrot set. In 1991, graduate student Dave Boll discovered that π hides inside the most famous fractal: approaching the "neck" at c = −0.75 along the imaginary axis, the number of iterations before escape multiplied by the offset ε converges to π. The explanation involves differential equations whose solutions pass through the tangent function — which has period π. Even chaos cannot escape the circle constant.
Congressional Farce, Cultural Celebration, and the Tau Rebellion
The Indiana Pi Bill (1897)
In 1897, an Indiana physician named Edward J. Goodwin convinced a state legislator to introduce House Bill 246, which would legislate a "new mathematical truth" — Goodwin's claim to have squared the circle. The bill's garbled text implied π = 3.2 (among other contradictory values). It passed the Indiana House 67 to 0. Senators were poised to pass it too, until Professor C.A. Waldo of Purdue University, at the statehouse on entirely unrelated business, caught wind of it and coached senators on the bill's absurdity. The Senate spent a cheerful half-hour mocking it, then tabled it indefinitely. Technically, House Bill 246 has never been defeated — and could still be brought to a vote.
Pi (1998) — Darren Aronofsky's black-and-white debut, made for $60,000, won the Sundance Directing Award. It follows a number theorist's descent into madness while hunting for patterns in π and the stock market.
Contact (1985/1997) — In Carl Sagan's novel, the protagonist discovers a perfect circle drawn in 1s and 0s hidden deep in π's base-11 expansion — a creator's signature embedded in the architecture of mathematics itself. The idea that a message might be woven into the digits of a universal constant remains one of science fiction's most haunting premises.
Star Trek: "Wolf in the Fold" — Spock defeats an evil entity by ordering the Enterprise computer to "compute to the last digit the value of pi" — an impossible task that traps the entity in an infinite loop.
Pi Day
Physicist Larry Shaw invented Pi Day at the San Francisco Exploratorium on March 14, 1988 — 3/14, matching 3.14. He led parades around a circular brass "Pi Shrine" at 1:59 PM (the next three digits), with participants reciting digits to the sound of Elgar's "Pomp and Circumstance." March 14 also happens to be Albert Einstein's birthday (1879) and, in a cosmic coincidence, Stephen Hawking's death date (2018). Congress formally recognized Pi Day in 2009, and UNESCO designated March 14 as the International Day of Mathematics in 2019.
The most devoted pi disciples memorize digits competitively. Rajveer Meena holds the Guinness record: 70,000 digits, recited blindfolded over nearly ten hours at VIT University in India in 2015. Japan's Akira Haraguchi claims 100,000 but lacks Guinness verification.
The τ (Tau) Rebellion
Not everyone worships π. In 2001, mathematician Bob Palais published "π Is Wrong!" in The Mathematical Intelligencer, arguing that the true circle constant should be τ (tau) = 2π ≈ 6.283185..., since τ relates circumference directly to radius (C = τr). Michael Hartl's 2010 Tau Manifesto formalized the argument: a full turn equals τ radians, making fractions of a circle intuitive — a quarter turn is τ/4, not the befuddling π/2. Vi Hart's spirited YouTube videos brought the debate to millions. MIT cheekily announces admissions decisions on Pi Day at 6:28 PM — "Pi Day at Tau Time." The debate is partly serious, partly mathematical performance art — and entirely delightful.
At position 762 in π's decimal expansion sits a sequence of six consecutive 9s — nicknamed the "Feynman Point" after physicist Richard Feynman's alleged desire to recite π to that spot and then declare "nine nine nine nine nine nine, and so on!" — cheekily implying π is rational. Feynman probably never actually said this. The joke, like π, propagates endlessly.
Five Constants Walk Into an Equation
We end where all roads through mathematics converge. In 1748, Leonhard Euler published the formula eix = cos x + i sin x. Set x = π, and you get:
Five numbers. Three operations. Every major branch of mathematics represented in a single line. Zero, the additive identity. One, the multiplicative identity. e, the base of natural logarithms and the heartbeat of growth and decay. i, the imaginary unit that extends numbers into a second dimension. And π, the circle constant — the ratio that will not end.
A 1990 poll in The Mathematical Intelligencer ranked it the most beautiful theorem in mathematics. A 2004 Physics World survey placed it alongside Maxwell's equations as the greatest equation ever written. A 2014 neuroscience study found it activated the medial orbitofrontal cortex — the brain's beauty center — more consistently than any other formula shown to mathematicians. It is not a metaphor to say this equation is beautiful; the brain literally processes it as beauty.
Gentlemen, that is surely true, it is absolutely paradoxical; we cannot understand it, and we don't know what it means. But we have proved it, and therefore we know it must be the truth.
— Benjamin Peirce, after deriving the identity at the Harvard blackboard, 19th centuryRichard Feynman, in his Lectures on Physics, called Euler's formula "one of the most remarkable, almost astounding, formulas in all of mathematics" and concluded simply: "This is our jewel."
Stanford mathematician Keith Devlin put it perhaps most eloquently: "Like a Shakespearean sonnet that captures the very essence of love, or a painting that brings out the beauty of the human form that is far more than just skin deep, Euler's equation reaches down into the very depths of existence."
∞
Four thousand years after a Babylonian scribe pressed a reed stylus into wet clay and wrote down 3.125, we have computed 314 trillion digits of π and proved that the number extends forever without pattern. We have shown it is transcendental — unreachable by any finite algebraic process — and yet it appears unbidden in gravity, electricity, uncertainty, probability, and the structure of spacetime itself. We need 15 digits to navigate between planets and 62 to measure the cosmos, yet we keep computing trillions more.
Ramanujan saw God in it. Euler found beauty. Archimedes found rigor. Feynman found a jewel. And still, at the 314-trillionth digit and beyond, π keeps going — indifferent to our devotion, irreducible, infinite, ours to chase but never to catch.