The forty-card wall
Ask anyone who has abandoned a flashcard app when they stopped, and you get a remarkably consistent answer: somewhere around thirty or forty cards, a few weeks in. Not on day one, when enthusiasm was high, and not at four hundred, when the habit would have been real.
What is interesting is why they stopped. It is almost never the reviewing. Reviewing is five minutes on a bus and it feels good, because recalling something you half-remembered is genuinely satisfying.
They stopped because of the card-writing.
The asymmetry nobody warns you about
A flashcard habit has two jobs stapled together, and they are not equally pleasant.
Reviewing is short, bounded and rewarding. The app tells you exactly what to do, you do it, and you are finished.
Authoring is open-ended and joyless. You have a word. Now you need a definition that is short enough to recall but complete enough to be worth recalling. Then usage. Then, if you are being thorough, an example sentence. For one card. You have forty words from this chapter.
So the deck stops growing. And a deck that stops growing is a deck you finish — you review the same forty cards until the intervals stretch out to months, the app goes quiet, and you delete it in the next storage cleanup.
Why “just make fewer, better cards” doesn’t fix it
The standard advice is to make fewer cards and make them well. It is correct advice about card quality and useless advice about the actual failure.
The problem is not that people make too many cards. It is that the cost of making one card is high enough to require a decision — and any step that requires a decision, every single time, at the end of a long day, is a step that eventually doesn’t happen.
You cannot fix a per-card cost problem by reducing the number of cards. You fix it by reducing the cost.
Move the work off the person
This is the specific bet Toly makes: type one line and the back of the card is written for you in about a second — translation and usage for a word, an explanation for a concept, an answer for a question. You read it, change anything you disagree with, and save.
The card is still yours. You chose the word, you approved the back, you edit anything wrong. What is gone is the blank field and the ten minutes of composition it used to demand.
When adding a card costs one line of typing, you add the word you met on the bus. That is the whole difference between a deck that dies at forty cards and one that is still growing in March.