Four teachers. Four little students. One big exam.

Every student starts out knowing nothing, and they all start exactly alike. Each one has to learn the same secret rule: which things are a “yes” and which are a “no”. A thing has two traits, like ☀️ how sunny it is and 🧼 how bubbly it is. Only those two traits decide the answer.

Pick a plan. Grab a free desk and pick one of five teaching plans, like ❓ Quiz first or 🧗 Challenge problems. When the bell rings, your teacher plays the plan by itself, about one lesson a second. You can change the plan any time.

Watch what really happens. You see each thing being taught, what your student guessed before the lesson, and which kinds of things it knows, is shaky on, or has a wrong idea about. Sometimes your student needs you: then you get 2 or 3 cards to pick from (10 seconds, or the plan just keeps going).

Robot teachers play a plan too: quiz first, mix it up, guesser (random examples), napper (doesn't teach).

There are 10 rounds of 15 lessons. After each round, every student studies and takes a quiz: the test paper shows 20 real questions with ✓ and ✗.

Your AI. It learns from the plans you pick. Next match it can race you at a robot desk, or teach for you when you choose “Let my AI teach”. It is marked with the exact same hidden exam as you, in percent.

After round 10, every student takes a secret 5,000-question exam. The best student wins , and the top three go on the podium.

A new match starts every few minutes, day and night. If your phone drops out during a match, your teacher keeps playing your plan.

Practise on your own