Claude Shannon at 110: the man who invented the bit

You used a messaging app today. Listened to a song on a streaming service. Received a file by email. Every one of those actions depended on billions of operations with zeros and ones — bits. The word exists because a man called Claude Shannon decided, in 1948, that the smallest possible unit of information needed a name and some mathematics. That man was born on 30 April 1916. Today he would have turned 110.

From Gaylord, Michigan, to the foundation of the digital age

Claude Elwood Shannon grew up in Gaylord, a small town in Michigan. From early on he had the profile of the curious engineer: he built radio sets, made toy steamboats, and rigged up a homemade telegraph system between his house and a friend's across the road — using the wires of a barbed-wire fence as the transmission line.

He graduated from the University of Michigan with two degrees at once — electrical engineering and mathematics. In 1937, still a master's student at MIT, he wrote a dissertation that most specialists consider the most important engineering thesis of the twentieth century: he applied Boolean algebra to the design of electrical circuits, showing that logical operations could be implemented with switches and relays. That idea is the physical foundation of every computer in existence.

The paper Scientific American called the "Magna Carta of the Information Age"

In July 1948, the Bell System Technical Journal published a paper by Shannon called "A Mathematical Theory of Communication". Specialists consider it one of the most influential scientific papers ever published — IEEE Spectrum has described it as possibly the most important of the century.

The problem Shannon was solving looked technical: how do you transmit information reliably over a noisy channel? The answer he built was a complete theory — mathematical, rigorous, applicable to any communication system, whether telephone, radio, or anything that came afterwards.

The paper introduced the concept of information entropy: information is, fundamentally, the resolution of uncertainty. The more improbable the message, the more information it carries. A completely predictable sequence transmits nothing new. Shannon formalised that in equations with the same form as thermodynamic entropy in physics — and that was no coincidence.

Along with it came the channel capacity theorem: every communication channel has a maximum rate — the Shannon capacity — below which data can be transmitted with an arbitrarily small error rate. That isn't just theory: every modem, every audio codec, every network protocol in existence was designed inside the limits Shannon proved that summer of 1948.

The bit: a word with 110 years of weight behind it

Shannon didn't invent binary — that had existed for centuries. But he was the one who formalised the idea that the smallest unit of information is the choice between two equally probable alternatives: yes or no, 0 or 1. And he was the one who popularised the term bit — short for binary digit, a word coined by the statistician John Tukey but adopted and spread by Shannon.

Before Shannon, "information" was a vague, qualitative concept with no unit of measurement. After Shannon, you could measure it, calculate it, optimise it. You knew exactly how many bits you needed to represent a text, an image, a conversation. You knew how many bits fitted down a channel with given characteristics. You knew how to compress data without losing what mattered. The whole chain from MP3 to JPEG to ZIP to 4K video streaming runs through the theory he published in 1948.

Theseus: the metal mouse that learned mazes

Shannon wasn't only theory. At Bell Labs he built things — and some of them were decades ahead of their time.

In 1950 he built Theseus: a mechanical mouse controlled by electromagnets underneath a maze. Placed at the start, Theseus bumped into walls and explored the paths. Sensors recorded every decision. When Shannon repositioned the mouse at any point it had already explored, Theseus went straight to the exit without a single mistake.

It was, arguably, the first machine learning device in history. The memory wasn't in the mouse — it was in the relay circuit beneath the maze. But the resulting behaviour is what we'd now call reinforcement learning. Shannon named the mouse after the Greek hero who navigated the Minotaur's labyrinth — and he was entirely aware of what he was building.

The mathematician who rode a unicycle down the corridors of Bell Labs

Shannon was also, by any reasonable definition, eccentric in the best possible way.

He rode a unicycle through the Bell Labs corridors at night, juggling three or four balls at the same time. In his own back garden he strung a steel cable between two tree stumps and practised juggling while riding the unicycle along it. He even formalised a mathematical equation for the theory of juggling — because he was the sort of person who formalised absolutely everything.

In the lab he built machines with no declared practical purpose: a calculator that worked in Roman numerals, a mechanical chess machine before modern computers existed. And the device that became most famous among Bell Labs engineers: a box with a single switch. When you flipped the switch, the lid opened, a mechanical hand switched it back off and the lid closed. The machine's only function was to turn itself off.

110 years on: what Shannon left behind

Claude Shannon died on 24 February 2001, aged 84, with Alzheimer's. He spent his final years without recognising the revolution he himself had started — the cruellest possible irony for someone who devoted his life to understanding how information moves and is preserved.

Information theory is in every file compression, every wireless communication protocol, every error correction system that guarantees the data you receive is identical to the data that was sent. It's in modern cryptography. It's in machine learning. It's in every one of the technologies that define 2026.

Curiously, Shannon was born on 30 April 1916 — the same date on which, 77 years later, CERN would place the World Wide Web into the public domain. The Web that became the most important communication system in history was built on protocols that depend, at every layer, on the work Shannon published in 1948. Coincidences like that need no interpretation — they're simply nice.

If you'd like to stay on the theme of the great figures who changed the trajectory of computing, the post on Mark Weiser and ubiquitous computing is about another man who saw decades ahead — and who, unlike Shannon, didn't live to see everything he had predicted.