In 1984, the educational psychologist Benjamin Bloom published a short paper with a startling result. Students taught one-to-one, by a tutor, performed about two standard deviations better than students taught the same material in a conventional classroom. Two sigma. In plain terms: the average tutored student outperformed roughly 98% of the students in the ordinary class.
Read that again, because it's one of the largest effects anyone has ever found in education. Not a nudge — a different league. The same child, given a tutor instead of a classroom seat, becomes a top-2% student.
And then Bloom named the catch, which is why the paper is called The 2 Sigma Problem. We know the thing that works. We just can't afford it. You cannot hand every child on earth a personal tutor — the economics don't remotely work. So the "problem" he set the field was this: find a way to teach groups that gets somewhere close to one-to-one tutoring, without the one-to-one cost. Forty years of education research have, in one way or another, been chasing that ghost.
Why tutoring is so much better
It isn't mysterious. A classroom moves at one pace — roughly the average — which means it's too fast for some and too slow for others, most of the time, for everyone. A tutor moves at your pace. A tutor notices the exact sentence where you got lost and stops there. A tutor asks you questions instead of talking at you, which forces you to retrieve and think rather than nod along. A tutor gives feedback in seconds, not when the marked homework comes back a week later. A tutor is patient with the specific thing you find hard.
Strip it down and tutoring is really just attention — continuous, responsive, individual attention — plus the willingness to make you do the cognitive work yourself. That combination is the most powerful teaching tool we've ever measured, and for most of history it has been a luxury good: the quiet preserve of princes and the wealthy, who have always hired it.
One honest note, because this site doesn't traffic in overclaims: the precise "2 sigma" figure has been argued over, and it's proven hard to reproduce exactly under every condition. But the direction is not seriously in doubt. Individual, responsive teaching massively outperforms one-to-many instruction. Bloom's number can be debated; his point stands.
The first technology that can actually do it
Here is what's new. For the first time, there is a technology that can deliver something genuinely tutor-like at a marginal cost of approximately nothing.
A capable AI will teach one-to-one, at your pace, in your language, at whatever level you need, for as long as you like, at three in the morning, without impatience and without judgement. It is not a perfect human tutor — it can be wrong, it doesn't know you as a person, and it needs watching. But it can do the core of what makes tutoring work — respond to you, question you, calibrate to you — and it can do it for everyone at once. That is a genuinely new thing in the history of education. Bloom's problem has a candidate answer for the first time, and it arrived from an unexpected direction.
The catch AI brings with it
There's a sting in the tail, though, and it's the reason this whole site exists.
On its own, an AI is not a tutor. It's an answer machine. Ask it a question and it will simply tell you — which, as we've argued elsewhere, teaches you almost nothing, because understanding comes from being questioned, not told. To get the two-sigma effect out of a general-purpose model, you have to actively turn it into a tutor: make it withhold answers, question you, test you, pitch to your level, argue both sides. Most people never do — so most people get an expensive search engine instead of a tutor.
That's the gap ModernEncyclopedia exists to close. Every subject in the library comes with the questions that flip a chatbot into a tutor already written and tuned — a Socratic partner, an examiner, a debate engine — so that the thing Bloom found, and the thing that's always cost too much, is simply available. To anyone. For free.
The two-sigma problem was never really about whether tutoring works. It works. It was about whether we could ever afford to give it to everyone. For the first time, the answer might be yes — and the only missing piece is the questions. Which is exactly what we hand you.