The Quiet Revolution: How Three Ideas From a Thousand Years Ago Built the Modern World

I write JavaScript. JavaScript runs on a virtual machine. The virtual machine compiles to assembly. Assembly runs on a CPU. The CPU executes operations on numbers. The numbers are encoded in binary, which is a positional notation.

The operations are arithmetic and logical, defined algorithmically, manipulating symbolic representations.

It's the same three ideas, all the way down.

I went looking for those three ideas the other night. The trigger was a name I had seen a hundred times and never really looked at: al-Khwarizmi. I knew it was where we got the word "algorithm." That was the extent of it. So I clicked. The further I dug, the more it felt like I was looking at the load-bearing beams of basically everything I do for a living. Every line of code. Every calculation my phone does in the background. Every model that gets trained on a GPU somewhere. All of it sits on top of a stack, and three of the biggest ideas in that stack came from a stretch of about 800 years between roughly 500 BCE in India and 850 CE in Baghdad.

This essay is about those three ideas. Algebra, zero, and the algorithm.

Algebra: math used to be a drawing problem

For most of human history, doing math meant doing geometry. The Greeks were the gold standard. Euclid's Elements, written around 300 BCE, is basically the most successful textbook in history, used in classrooms for over two thousand years. And it's almost entirely about shapes.

Want to find the area of something? Draw it, decompose it into triangles. The Pythagorean theorem isn't really a statement about $a^2 + b^2 = c^2$ in the way we think of it now. For the Greeks it was a statement about the areas of literal squares drawn on the sides of a triangle.

This was beautiful. It was also a cage. Some problems are really hard to draw. If your equation involves a fourth power, you'd need to reason about a 4-dimensional shape, and Greek math didn't have great tools for that. They had a deep philosophical commitment to the idea that math had to correspond to physical, drawable reality. Anything you couldn't construct with a compass and a straightedge was suspect. Geometric proofs also didn't generalize easily. A proof that worked for one specific triangle didn't necessarily tell you how to handle a different one. Every problem was bespoke.

The systematic break came from al-Khwarizmi, working at the House of Wisdom in Baghdad around 820 CE. His book was called The Compendious Book on Calculation by Completion and Balancing. The "completion and balancing" is the technical content: he described two operations, al-jabr (restoring, or moving a subtracted term to the other side as a positive) and al-muqābala (balancing, or canceling like terms on both sides). If you've ever done algebra homework, you've done both many times over. The Arabic al-jabr is where we got the word "algebra."

The genius wasn't really in those two operations. It was in the framing. He was saying: here is a procedure. Apply step one. Apply step two. You will arrive at the answer. The answer doesn't depend on you being clever or having a flash of geometric insight. It depends only on whether you followed the steps correctly.

This shift is so deeply baked into modern math that it's almost hard to see now. Calculus, linear algebra, differential equations, statistics, all of it lives in a world that al-Khwarizmi made habitable.

Zero: the number that almost wasn't

While that was percolating in Baghdad, something else was happening in India that turned out to be just as important.

For a long time, humans had counting systems but no proper concept of zero as a number. The Babylonians used a placeholder symbol for "no value in this column" by around 300 BCE, but they didn't think of it as a number you could do arithmetic with. The Greeks didn't have one. Aristotle had argued that zero couldn't really be a number because you can't divide by it, a position that didn't age particularly well.

Indian mathematicians cracked it open. By the 5th century, Indian astronomers were using a decimal place-value system with a symbol for zero. In 628 CE, a mathematician named Brahmagupta wrote a book called the Brāhmasphuṭasiddhānta that did something nobody had done before. It set out the rules for arithmetic with zero as a number. A number plus zero is the number. Zero times any number is zero. He even tried to figure out what happens when you divide by zero, got it wrong, but the fact that he was asking puts him centuries ahead of his time.

I want to pause here, because the importance of zero is one of those things that gets glossed over because the modern world makes it feel obvious. It's not obvious. Zero is the number that represents the absence of a thing. You can't pick up zero apples. You can't draw zero of anything. To treat that as a number, something you can add and subtract and multiply with, required a level of mathematical maturity that took most civilizations a long time to develop.

And here's the part that really matters: zero is what makes positional notation work. The digit "3" means different things in different places. In 3 it's three. In 30 it's thirty. In 300 it's three hundred. The position determines the value. Most ancient civilizations didn't write numbers this way. The Romans wrote numbers as strings of letters where each letter had a fixed value. To write 1888 you wrote MDCCCLXXXVIII. Try multiplying that by MCDXLII without converting to modern numerals first. It's genuinely difficult, and that difficulty is why most Roman-era arithmetic was done with physical counting boards rather than on paper.

But positional notation only works if you have zero. Without zero, you can't distinguish 18 from 108 from 1008. The zero holds the place open. It says, yes, this column exists, and yes, it has nothing in it.

The system spread from India to the Islamic world over the 7th and 8th centuries through trade and translation. Al-Khwarizmi himself wrote a treatise on it around 825 CE. He was upfront about the origin. The Arabic name for these digits was arqām hindiyya, "Indian numerals." We call them "Arabic numerals" today only because that's the route by which they reached Europe. And Europe didn't adopt them quickly. There's about a 400-year period where merchants and scholars argued over whether to use the new system or stick with Roman numerals. Some Italian city-states actually banned the new digits because people worried they were too easy to forge. A 0 could be turned into a 6 or a 9 with a quick stroke of the pen. It took the printing press in the 15th century to really lock the new system in.

Even genuinely better technologies can take centuries to displace the ones people are used to.

The algorithm: what happens when you combine the two

We now have two ingredients on the counter. Symbolic problem-solving and a number system that makes computation fast. The third idea is what happens when you put them together.

The word itself comes from history's most direct route. When al-Khwarizmi's book on Hindu numerals was translated into Latin in the 12th century, the translator rendered his name as "Algoritmi." The book opened with the line Algoritmi dixit, "Algoritmi says." European scholars started using "algorism" to refer to the method he described. Over the centuries it drifted to "algorithm," and the meaning generalized from "the specific method of decimal arithmetic" to "any step-by-step procedure for solving a problem."

So when you say "the YouTube algorithm" or "we need a better sorting algorithm," you are, in a literal etymological sense, invoking the name of a 9th-century Persian mathematician.

But the linguistic accident is the least interesting part. The interesting part is what an algorithm actually is.

An algorithm has three properties that distinguish it from just "a way of doing something." It's finite. It terminates. You don't run forever. It's deterministic. Given the same input and the same steps, you get the same output. There's no "and then you intuit the rest." Every step is specified. And it's general. It solves a class of problems, not just one specific instance. A recipe for one specific cake is not an algorithm. A recipe that works for any cake within some category is.

Al-Khwarizmi's methods had all three properties. His procedures for solving quadratic equations would terminate, would give the same answer every time, and would work for any quadratic equation, not just the example he happened to use.

The thousand years between al-Khwarizmi and modern computing is a long story I won't tell here. The short version: Persian mathematicians extended algebraic methods to higher-order equations. Fibonacci's Liber Abaci in 1202 taught European merchants how to do business math with the new numerals. Symbolic notation matured in the 16th and 17th centuries with Viète and Descartes. Calculus showed up. George Boole formalized logic as algebra in 1854. Charles Babbage designed the first general-purpose mechanical computer. Alan Turing, in 1936, defined what an algorithm formally is, and proved that some problems are genuinely uncomputable. Within a few decades you have FORTRAN, then C, then the internet, then the laptop on which this article was written.

Every step of that ladder is built on the rung below. And the bottom rung is a 9th-century book that wrote down step-by-step procedures for solving equations.

A note on credit

This story sometimes gets told as a "great man" narrative. Al-Khwarizmi invented algebra, end of story. That's not quite right.

What actually happened is more like a thousand-year relay race across multiple civilizations. Indian mathematicians developed positional notation and zero. Persian and Arab scholars at the House of Wisdom translated Indian and Greek texts, synthesized them, extended them. Latin scholars in medieval Europe translated the Arabic works and built on them. Renaissance Europeans pushed the symbolic tradition into modern algebra and calculus. 20th-century logicians and engineers turned it into computer science.

Al-Khwarizmi gets a starring role partly because his name happens to have stuck to the word "algorithm," and partly because he was working at a genuinely pivotal node. The House of Wisdom under the Abbasid caliphate was, for a couple of centuries, the most important place in the world for math and science. But he was synthesizing and extending traditions, not creating from nothing. Brahmagupta deserves equal billing. So do hundreds of mathematicians whose names we don't even know because their work survived only in translations of translations.

Big ideas are usually collective, slow, and international. The algorithm, foundation of the most powerful technology our species has ever built, was developed over a millennium by people speaking Sanskrit, Greek, Persian, Arabic, and Latin, working in cities from Alexandria to Baghdad to Toledo to Pisa. None of them knew they were building computer science. They were just trying to do math better.

Wrapping up

The three ideas are so deeply embedded in modern life that they've become invisible. They're hard to see because we're standing on them. But each one was, in its time, a genuinely radical break from how humans thought before.

Algebra freed math from geometry. It said you don't need a picture, you need a procedure. That made it possible to reason about quantities and relationships that no human can visualize, which turns out to be most of the interesting ones.

Zero made computation tractable. It turned arithmetic from a craft into something a literate ten-year-old can do. Making things easy is what makes them ubiquitous.

And the algorithm, the descendant of those two ideas, is the bridge between human thought and machine execution. It's what lets us hand a problem to a piece of silicon and trust the answer. Without it there is no software, no internet, no AI, no modern world.

The next time you unlock your phone with your face, follow a GPS route, or read something an LLM helped produce, the foundation underneath all of it was laid by people whose names most of us have never learned. Twelve hundred years ago, scholars wrote down rules for moving symbols on a page, and those same rules are now running every piece of software on Earth. There's something quietly remarkable about that, about ideas being durable enough to outlast the civilizations that produced them, the libraries that housed them, and the languages they were first written in.

We owe them more than we know.

TL;DR

Three ideas, developed between roughly 500 BCE and 850 CE across India, the Mediterranean, and the Islamic world, are the substrate of computer science.

Algebra broke math out of geometry. Before al-Khwarizmi's book around 820 CE, hard math problems usually meant drawing shapes. He introduced the idea of solving problems by manipulating abstract symbols according to fixed rules. The word "algebra" comes from al-jabr, one of the operations he described.

Zero wasn't always a number. Indian mathematicians, most famously Brahmagupta in the 7th century, treated zero as a real number you could do arithmetic with. Combined with positional notation, this is what made fast computation possible.

Algorithms are step-by-step procedures with three properties: finite, deterministic, general. The word comes from the Latin spelling of al-Khwarizmi's name. The concept is what every program on earth is, all the way up to the LLM you're probably using to do your job right now.

No single person did this. It was a thousand-year relay race across civilizations. We're standing on it.