Teach Me The Concepts
Explain it in your own words. See exactly where your understanding breaks.
Two sentences in. Ritsu checks them against your own material — and asks the next question from the gap.
Free account · One chapter or lecture · Your first question in a couple of minutes
Compound interest
Why does money earning compound interest grow faster over time than money earning the same rate only on the original amount?
A few sentences beats a few words.
Try it here — this page can't grade you; a session does. · Two sentences is enough to start · Your words never leave this page
The exchange
One answer tells you what you know. The next question finds what you don't.
Personal Finance 101 — Chapter 4: InterestExample session · illustrative
Why does money earning compound interest grow faster over time than money earning the same rate only on the original amount?
A few sentences beats a few words.
“Interest on a bigger number” — that's the key idea, and you've named the mechanism correctly: each year's interest joins the base, so the next year's interest is calculated on more. You've also drawn the right contrast with a flat rate, where the base stays fixed. What you haven't said yet is why that difference gets larger every year rather than staying the same.
This page
The feedback repeats the learner's own words back. It graded what was said, not what it expected.
This page
Accuracy, coverage, depth, clarity — each out of 5, weighted into one number. The card is the product's.
You explained that the base grows. Now the why behind the shape: if both accounts earn the same 10%, why does the gap between them widen every single year instead of just staying open?
Why this question
The first answer said the base grows — but not why the gap accelerates. The follow-up went straight at that missing piece.
“The extra interest itself starts earning” — that's the widening, and it's exactly right: every year the compound base carries last year's interest, so the same 10% produces a little more than it did the year before. One assumption is sitting underneath your explanation, though: that the rate stays put.
Your explanation assumes the rate never changes. Suppose it dropped to 1% for both accounts — does compounding still win, and would the gap open as fast?
“The two look almost identical for years” — yes, and you reasoned through it rather than guessing: the mechanism doesn't depend on the rate, but the size of the effect does. That's the last piece: compounding always wins eventually, and the rate decides how soon it shows. This is a solid explanation you could give without notes.
This page
At 75 the concept ticks, and the tick stays even if a later answer is weaker. Every graded answer moves the number; the toast shows the change.
That question came from your answer.
Ritsu doesn't just move to question two. It listens for what you explained, what you skipped, and what may be wrong — then uses that to decide what to ask next.
The conversation gets harder where your explanation gets thinner.
Free account · One chapter or lecture · Your first question in a couple of minutes
Sound familiar?
Explaining something out loud can feel like active learning. Until nobody tells you what you missed.
- You explain the idea to yourself.
- It sounds coherent.
- Nothing interrupts.
- Nothing asks why.
- Nothing points out the assumption you skipped.
- Nothing notices the part you quietly talked around.
So the explanation ends when you feel satisfied.
A monologue can expose confusion. A good follow-up exposes much more.
The missing half
Teach-back gets powerful when something pushes back.
Everyone recommends explaining it out loud.
And it stops working the moment you try it alone.
The reason is simple: nobody answers back.
Explaining forces you to retrieve the idea instead of simply recognising it.
But the real test begins when someone asks:
Gap-targeting
Catches an explanation that names the rule but not the reason.
That's where a polished explanation turns into tested understanding.
Don't just explain until it sounds right. Explain until it survives the questions.
Not a prompt
This isn't a list of questions. It's a conversation that remembers what you said.
It checks against the material
The grading spine is built before the first question
Personal Finance 101 — Ch. 4 Interest (lecture notes).pdf
Never shown to the learner
Before the first question, the reference is already built.
On your own document, Ritsu pulls out each concept's key points and likely misconceptions before the first question — up to 7 and 3 per concept. You never see the list. Every answer is judged against it.
It grades the explanation
Four dimensions on every answer, not one verdict
- Accuracy — Is what you said right?
- Coverage — How much of the concept's spine did you reach?
- Depth — Mechanism and reasons — not just labels.
- Clarity — Could a classmate follow it?
Every graded answer is scored on accuracy, coverage, depth and clarity, each out of 5, weighted into one number — so a fluent answer that skips the mechanism reads differently from a rough one that has it.
It adapts
Deeper when you're strong, easier when you're stuck
The level of the next question
Compound interest
The opening teach-back has no level — it's the bar.
Entropy
The opening teach-back has no level — it's the bar.
Strong answers push the next question up a level. Two weak answers in a row bring it down a level and rebuild the idea from a concrete case — the product's own rule. The Entropy example on this page is exactly that.
A prompt can keep chatting. A coach has to know what changed.
Progress that stays
The conversation ends. The progress stays.
Personal Finance 101 — Chapter 4: Interest
How interest is charged and earned, why compounding changes the shape of growth, and how to read a rate before you sign.
- 1Simple interestInterest charged only on the original amount, the same every period.82
- 2Compound interestInterest earned on interest — why a growing base makes growth accelerate.
- 3APY vs APRTwo ways to state a rate, and why the yield is the honest one.
- 4The Rule of 72A quick estimate of how long money takes to double at a given rate.64
- 5Inflation and real returnWhat a rate is worth once rising prices are taken out.58
- Every concept keeps its own mastery state.
- Strong concepts earn a tick.
- Weak concepts stay visible.
So the session ends with something more useful than a chat history: a map of what you can explain — and what still needs work.
You don't finish with more text. You finish knowing what you can actually defend.
The honest part
This AI doesn't do the thinking for you.
- 1It gives you the question.
- 2You supply the explanation.
- 3It responds to your reasoning.
- 4Then you have to defend it again.
AI removes the need for a study partner. It doesn't remove the need to think.
Built from your material
On your material, the material is the reference.
Ritsu uses the source you provide to identify expected ideas, key points and likely misconceptions.
Your explanation is judged against that structure rather than against a generic conversation about the topic.
Pick a source — Same session, four sources. Watch the concepts arrive from the material.
Example session · illustrative
Sized to the source
A session holds 4 to 20 concepts in up to 6 groups, each with up to 7 key points and 3 likely misconceptions pulled from your material — never a generic syllabus.
Every answer graded on four dimensions
Accuracy, coverage, depth and clarity, each 0 to 5, on every graded answer — the opening explanation and every follow-up alike. Answers under 40 characters are capped, so a few sentences always beats a few words.
Progress that sticks
Mastery is kept per concept, from 0 to 100. A concept ticks at 75 and the tick stays. Restart a concept any time — the old dialogue moves to History and your best score is kept.
Teach Me The Concepts finds the weak edge of your explanation. Ritsu keeps training it until it holds.
You already did the hard part: you proved that retrieval, explanation and pushback expose gaps rereading can hide. Ritsu applies that loop across your material — and brings the weak concepts back before they fade.
- Understand
- Explain
- Get challenged
- Practice
- Review
- Master
Questions
What people ask before their first answer.
Two sentences is enough to start. The tool caps the score of anything under 40 characters, so a few sentences always beats a few words — and later questions may need more detail.
One more time
Don't just review what you studied. Try to defend it.
Pick one concept. Explain it in your own words.
Let Ritsu find the part that still needs work.
Free account · One chapter or lecture · Your first question in a couple of minutes
- File
- YouTube
- Website
- Paste text
Your material. Your words. Your progress.