Pioneers & Inventors
The Woman Who Taught Machines to Remember
Every time a computer runs a program without crashing the one next to it, it's because a Navy admiral figured out in the 1950s that machines could share a brain.
The Idea
Before Grace Hopper, computers were essentially enormous, expensive calculators that spoke only in raw machine code — strings of ones and zeros unique to each machine. If you wanted to run the same calculation on a different computer, you rewrote everything from scratch. Hopper found this absurd. Not just inconvenient — philosophically wrong. She believed that if a task could be described clearly in English, a machine should be able to translate it into its own language automatically. That belief produced the compiler: a program that converts human-readable instructions into machine code. The idea sounds obvious now, which is precisely the measure of how radical it was then. Her colleagues were skeptical to the point of dismissal. The common objection was that computers couldn't 'understand' anything — they processed symbols, not meaning. Hopper's counter was pragmatic: understanding isn't the point; reliable translation is. She didn't need the machine to grasp intent. She needed it to follow rules consistently. That distinction — between meaning and instruction — is one that still shapes how we think about programming languages, and increasingly about AI. The compiler she built in 1952 eventually led to COBOL, a language so durable that it still processes a significant share of the world's financial transactions today, quietly humming under systems most people assume are ultramodern.
In the World
In 1959, a committee of computer manufacturers and government officials gathered to solve a crisis: the US military was drowning in incompatible software. Every branch ran different machines; every machine ran different code. Grace Hopper, then a reservist at the Naval Reserve Midshipmen's School, was brought in as a technical advisor. She arrived with a decade of compiler work already behind her and a very clear argument: the only way to make software portable — able to run on more than one type of machine — was to design a language that sat above the hardware entirely. The result was COBOL, the Common Business-Oriented Language. Hopper pushed for it to read almost like plain English: 'MULTIPLY HOURS BY RATE GIVING GROSS-PAY' rather than hexadecimal sequences. She wanted programmers to be able to read their own code six months later without a decoder ring. That choice was mocked at the time as unsophisticated. Decades on, it looks like genius. Estimates suggest that more COBOL code is executed on any given day than code written in Python, Java, and JavaScript combined — most of it buried in banking infrastructure, pension systems, and government databases. When a financial institution announces a 'digital transformation,' there is usually a very old COBOL core it is quietly trying to replace, and usually failing.
Why It Matters
Hopper's story reframes where technological revolutions actually happen. We tend to imagine breakthroughs as hardware moments — the first transistor, the first chip, the first smartphone. But Hopper's contribution was entirely conceptual: she changed what we believed a machine could be asked to do. Her insistence that computers should accommodate humans, rather than the other way around, is the philosophical foundation of every high-level programming language ever written. It's also a live argument. Every time someone debates whether AI tools should be designed to match human language or whether humans should learn to prompt more precisely, they are rehearsing a version of the dispute Hopper settled in 1952. She settled it in favour of the human. Worth remembering too: she did this while navigating a military culture that routinely sidelined women, returning from mandatory retirement multiple times until she was finally commissioned as a rear admiral at 79. The lesson isn't just about compilers. It's about what happens when someone refuses to accept that the way things are done is the way they must be done.
A Question to Ponder
Where in your own work or life are you adapting yourself to a system that was never actually designed to suit you — and who would benefit if you pushed back?
Get a new one of these every morning.
Start learning with Thinkable