Brilliant.org

A commuter opens an app and does calculus. On purpose.

It is a Tuesday morning on a crowded train somewhere between Brooklyn and Manhattan, and a 34-year-old marketing manager is solving a probability puzzle on her phone. She is not preparing for an exam. There is no exam. She is doing this for the same reason other people do crossword puzzles - because the next screen is a small, satisfying click of understanding, and the one after that is another.

This is Brilliant.org's everyday miracle. Roughly ten million people have signed up to learn math, science, and computer science on a platform that, by design, refuses to lecture them. The company sits in San Francisco. It employs around 140 people. It charges roughly $150 a year for the privilege of being made to think. And it works.

The internet promised a renaissance in learning. It delivered videos.

By the early 2010s, online education had a clear, slightly embarrassing pattern. A professor in a polo shirt would explain a concept. A student would watch. The student would close the tab. The student would learn very little, but feel like they had. This is not a knock on MOOCs - it is a knock on a medium that mistook attention for understanding.

Sue Khim, the company's CEO, noticed something inconvenient about how math and science are actually taught at elite levels - they are taught by doing. You sit with a problem. You fail. You try again. You ask a small question and answer it before you are allowed the next. The lecture is the dessert, not the meal.

So the problem was not access. Access had been solved. Khan Academy had hundreds of millions of viewers. Coursera was minting credentials. The problem was that nobody was getting visibly smarter at things that are genuinely hard. People were collecting tabs.

In 2012, two founders bet that people would pay to be challenged.

This sounds obvious until you remember the alternative was free. Sue Khim and co-founder Silas H pivoted out of an earlier edtech project called Alltuition and started Brilliant in October 2012. The pitch went to the Launch Festival in March 2013. Chamath Palihapitiya, then at Social Capital, wrote a check. So did 500 Startups, Kapor Capital, Learn Capital, and Hyde Park Angels.

The early bet was strange in its specificity - that competitive math problems, the kind ambitious teenagers chase on weekends, could be reformatted for the rest of the population. That you could take the addictive structure of a puzzle game and quietly smuggle in the Pythagorean theorem.

For a stretch, Brilliant looked like a problem set with a brand. Then, slowly, it became something more interesting - a curriculum designed by PhDs and former teachers, packaged into bite-size interactive lessons, paced by the user, scored against itself.

An app that won't shut up - and won't lecture.

Open Brilliant on a phone and the experience is closer to Duolingo than to a university course. There is a streak. There is a badge. There is, by design, no half-hour video to skip through at 1.75x. Every lesson is a sequence of small problems that build on each other. Get one wrong and the app gently rolls back. Get a few right and it raises the temperature.

When the newspapers covering you also source their puzzles from you.

It is one thing to claim cultural relevance. It is another to write puzzles for The New York Times, The Guardian, and FiveThirtyEight - which Brilliant does. The same outlets that profile the company also publish its content. Few edtech companies can say that.

"One person, one question, one small commitment."

The official mission reads like something you could fit on a fortune cookie - inspire and develop people to achieve their goals in STEM, one person, one question, one small commitment at a time. The phrasing is gentle on purpose. Brilliant is not trying to replace the university. It is trying to replace the moment after the lecture, when the student goes home, opens the textbook, and freezes.

That moment - the gap between hearing and understanding - is where most of education actually happens. It is also where most of education actually fails. Brilliant has spent more than a decade engineering software whose only job is to live inside that gap.

Who shows up

Two populations dominate. Curious adults who want to keep their math sharp or pick up a new language like Python or linear algebra. And ambitious teenagers and university students who want a faster, denser path than their classroom is willing to offer. The platform is officially aimed at ages 13 to 113. It mostly is.

The AI tutors are coming. Brilliant has already been tutoring for fourteen years.

It is fashionable in 2026 to talk about AI tutors as if they are a brand-new category. They are not. The infrastructure for adaptive, problem-led, individually paced learning has been quietly built by companies like Brilliant for over a decade. The new models will plug into it. They will not invent it.

The company's recent additions - neural networks, large language models, data analysis - are not coincidence. Brilliant is positioning itself to be the place where the next generation of technical workers first encounters the tools that will employ them. If it gets that right, it stops being an edtech app. It becomes a fixture.

Back on the train.

The marketing manager from the opening scene is still on the same train, except now it is Thursday. She has just finished a lesson on Bayesian reasoning, which she did not previously know she wanted to finish. The train doors open. She puts the phone away. Tomorrow she will pick up where she left off, because the app remembers and, increasingly, so does she.

This is the part Brilliant.org got right. Not the content. Not the funding. Not the celebrity investors. The part where you close the app a little smarter than you opened it - and you can tell.

Where to go next

Watch & listen