Manifesto
You were born a scientist. You dropped things to see if they’d fall. You asked why until the adults ran out of answers.
Then came the syllabus. Sit still. Read this. Remember it for Friday.
Now the machines can answer anything, instantly. Good. Let them.
Because an answer isn’t understanding. Understanding is what happens when you guess, get it wrong, push on the thing, and watch it break the way you didn’t expect.
You can outsource your thinking. You can’t outsource your understanding.
Ask anything. Then go find out for yourself.
Philosophy
Learning is something you do, not something that happens to you.
Feynman School is built on a simple bet: people understand hard ideas when they act on them. They predict, test, get it wrong, explain it back and use it somewhere new. Everything in a quest exists to make that happen, and to prove it did.
- 01
Understanding that lasts
Watching a video feels like learning and is gone within a week. We aim for retention: every idea is retrieved on a fresh case during the quest, then brought back as short spaced reviews over the following weeks, at longer gaps each time it holds.
- 02
A story to think inside
Every quest keeps one running scenario, with the same quantities and assumptions, so ideas build on each other instead of resetting. A good scenario makes the abstract feel necessary.
- 03
Personal, with a feedback loop
Your goals, background and answers shape the route, the examples and the depth. The quest notices what you got wrong and what you breezed through, and adapts the next step. Feedback rewards evidence of understanding, never attendance.
- 04
Interactive explainers
Each idea comes with a live model built for it: a simulation, a graph you can drag, an algorithm you can step through, code you can break. The model runs the real mathematics, so what you see is the true consequence of what you change.
- 05
Predict, then unlock
Before the lab opens, you commit to a prediction. A wrong guess you can test teaches more than a right answer you were told. If you genuinely don’t know, saying so is a valid answer: the next step slows down for you.
- 06
Many ways into every idea
Some people need the equation, some the code, some a picture, some another experiment. At any step you can open the same idea as a derivation, a program, a diagram or a new lab, or test yourself on it.
- 07
Pacing that respects your attention
A quest starts with a two-to-five-minute spark, then each step takes about as long as the idea needs. One idea per screen; depth is a layer you open, not a wall you climb.
- 08
Design your own mission
Interview prep, an exam, a paper, a problem at work, or pure curiosity. Say what you are learning for and the quest is planned around it. You can reshape the route in your own words at any time.
- 09
Grounded in real sources
Bring a YouTube lecture, an article or an arXiv paper and the quest follows what it actually says, with its real numbers. Models state what they simplify and where they stop being true.
- 10
A map of the domain
Every quest has a place in a knowledge map of its domain, so you can see what you’ve explored and what lies next to it.
- 11
A mentor beside you
An AI mentor knows the quest, the step you are on and what you’ve tried. It gives hints without spoiling the challenge and can build you a new view when you are stuck.
- 12
Honest about mastery
We never award a “mastered” badge. We keep what you predicted, tried and explained, show how your checks went, and bring ideas back when it’s time to recall them.
Taste, not just technology
AI models get better every few months, and our engine gets better with them: it uses several models, each for the work it does best, and checks every result before you see it. What we add is judgement about what makes an explanation actually work, refined by what learners do in real quests.