The problem
Relativity. Genetics. How the economy actually works. How neural networks learn. Why your code works sometimes and not others. There’s something you’ve been meaning to understand — really understand — for years. You tried.
The textbook was impenetrable. It was written for people who already understood the thing, and it punished you for not being one of them. The YouTube video was better, but it assumed prerequisites you didn’t have, and you spent more time pausing and Googling than actually learning. You asked ChatGPT and got a perfectly fluent answer that scratched the itch for ten minutes and then evaporated. You couldn’t remember any of it by Thursday.
You never actually learned the thing.
You just brushed past it, again, and the wall got a little taller.
This is not a motivation problem. You’re curious — curiosity is abundant, almost involuntary. The problem is that curiosity keeps running into walls, and every failed attempt makes the next attempt feel more pointless.
What everyone gets wrong
They all treat learning as labor to be managed. They differ only in whether they gamify the labor, credential it, catalogue it, or optimize it. Not one of them creates the feeling of sitting with a teacher who cares specifically about you.
What we believe
The oldest and most effective technology for human learning ever invented wasn’t a textbook or a video or a flashcard app. It was a patient teacher who knew you.
Aristotle tutored Alexander, and Alexander conquered the known world. Socrates taught Plato in the agora, one question at a time, and Plato built the foundation of Western philosophy. Dronacharya trained Arjuna under the banyan tree — one teacher, one student, unlimited patience — and Arjuna became the greatest archer who ever lived.
One thread runs through all of them: a teacher who knew the student — their strengths, their gaps, their fears — and refused to move on until it clicked.
In 1984, Benjamin Bloom proved what these stories already told us. He called it the “2 Sigma Problem”: students who receive one-on-one tutoring perform two standard deviations better than students in conventional classrooms. The 50th percentile becomes the 98th. Not because the students were smarter. Because the teaching was personal.
Only princes got Aristotle. Only royals got Dronacharya. Everyone else got the textbook.
What Teach Me Something does
Teach Me Something turns your AI into that teacher.
It gives your AI a personal curriculum, a growing tutorial library, and spaced repetition quizzes. Your AI learns who you are — your background, your goals, the way you think. Then it writes tutorials pitched at exactly your level, using your own projects as examples. Not generic explanations. Articles written for you, the way a great teacher would write a chapter of a textbook if you were the only student.
When you read a tutorial and struggle with section three, your AI notices. It rewrites. It approaches from a different angle. It doesn’t give up on you. And when you’re about to forget what you learned, it quizzes you — not with gamified flashcards, but with real questions that probe whether you understood or just read.
Your understanding deepens. Your library grows. Physics sits next to Python sits next to music theory. It all compounds in one place, taught by someone who knows you.
Learning is one of the deepest human pleasures. What’s been missing isn’t motivation. It’s a good teacher.
