Forgetting is the point

Cramming works, briefly. Four hours the night before produces a real effect that is largely gone within a week.

Spread the same four hours across two weeks and you will know it months later. Same total effort, different result — and the mechanism is one people find counter-intuitive.

Forgetting a little is the active ingredient

Memory fades along a curve, steeply at first. Review it while it is still fresh and retrieval is easy — and easy retrieval strengthens very little.

Wait until it has partly faded, and pulling it back takes effort. That effort is the strengthening. You are not topping something up; you are rebuilding it, slightly, and the rebuild is what lasts.

Hence the counter-intuitive part: a review that felt too easy was slightly wasted. You should be a little unsure. That discomfort is L1’s desirable difficulty in its most practical form.

What a scheduling system does

A spaced repetition system tracks each item and schedules it for just before you would forget it. Get it right and the gap lengthens — a day, then three, then a week, then a month. Get it wrong and the gap shrinks back.

Over time you spend almost all your effort on the minority of things that are genuinely difficult for you, and barely any on what you have solidly learned. That efficiency is the whole point, and it is why Lrnon’s review cards exist and why the intervals stretch the way they do.

You do not need software for it. A shuffled stack of paper cards in dated boxes is the original version and works.

Where AI helps, and where it hurts

Helps: turning material into cards quickly, which is the tedious step that stops people starting.

Turn this into flashcards. One fact per card. The front must be a question that cannot be answered by recognition. Include a few asking WHY, not only what.

Hurts: it will happily generate two hundred cards, and two hundred cards is a daily obligation you will abandon in three weeks.

Ask for fifteen. The rule that matters: make cards for things you would be annoyed to have forgotten in six months. Everything else is noise you have committed to reviewing forever.

What flashcards cannot do

Worth being clear, because people over-invest here.

Cards build recall of specific things — definitions, dates, formulas, vocabulary, the name of the thing. Genuinely valuable, and a foundation.

They do not build understanding, judgement, or the ability to do. No number of cards teaches you to write well, diagnose a fault, structure an argument, or notice that a situation is unusual.

So cards are one component. L3’s application questions and L5’s teaching-back are where understanding gets built, and L7 covers what neither can reach.

Making it survive

Same argument as E4·L7. A system you abandon has taught you nothing.

Small daily beats large weekly. Ten minutes most days outperforms two hours on Sunday, both for the spacing effect and for actually happening.

Do the due ones, not all of them. The scheduler exists to tell you what needs attention. Overriding it is doing more work for less benefit.

Delete cards you no longer care about. A backlog is the main reason people quit, and the cure is a smaller deck rather than more discipline.

Try it now (6 minutes)

Take something you learned recently. Ask for fifteen cards, one fact each, question-fronted, some asking why.

Then cut it to the ten you would genuinely be annoyed to have forgotten in six months. Review those tomorrow, not today.

Check your understanding

1. Spacing beats cramming because: It spreads fatigue more evenly Partial forgetting makes the next retrieval effortful, and the effort strengthens it You cover more material overall

2. The most common reason people abandon a review system: The intervals are too long Too many cards, creating a backlog that becomes a daily obligation The software is unreliable

3. Flashcards are poorly suited to building: Recall of definitions and formulas Understanding, judgement and the ability to do something Vocabulary in a new language

Recap

Spacing works because partial forgetting makes the next retrieval effortful, and the effort is the strengthening — so an easy review was a slightly wasted one. A scheduler concentrates your effort on the minority you find hard. Use AI to make cards quickly, then cut ruthlessly to what you would mind forgetting in six months. And remember cards build recall, not understanding or capability.