You finish a good book or a dense article. You highlighted the best parts. You even took notes. A week later someone asks you about it, and you have almost nothing. The ideas were there, and now they are not.
This is not a memory problem. It is a process problem. And once you see what is going on, the fix is small.
The forgetting curve is doing its job
In the 1880s, Hermann Ebbinghaus measured how fast he forgot lists of syllables. The shape he found has held up for more than a century. Without review, you lose roughly half of new material within an hour, and about 70 percent within a day.
Your brain is not broken when this happens. It is being efficient. Most of what you read once does not come back, so it gets dropped. The brain keeps what it sees again.
Capturing a note creates a feeling of safety. The idea is "saved." But saving is not the same as learning, and that feeling is the trap.
Highlighting feels like learning. It isn't.
Rereading and highlighting are smooth. Nothing pushes back, so you feel fluent, and you read that fluency as understanding. Researchers call this the fluency illusion.
The methods that actually build memory feel worse in the moment:
- Active recall. Close the book and try to retrieve the idea from memory. The effort of pulling it back is what strengthens it.
- Spacing. Review just before you would have forgotten, then stretch the gap. Each successful recall buys you a longer interval.
- Elaboration. Put the idea in your own words and connect it to something you already know.
None of these are comfortable. That discomfort is the work.
Why "just make flashcards" falls short
Flashcards are the classic way to do active recall and spacing at once. They work. The friction is everything around them: you leave what you are reading, decide what to turn into a card, write the front and the back, and keep a separate app in sync with your notes.
Most people quit at the card-making step, not the reviewing step.
A smaller change
The fix is not more discipline. It is removing the gap between writing something down and being asked about it later.
That is the idea behind Anti-Agent. You write or import into a page as usual. The page can then turn parts of itself into things that come back: a flashcard that grades your answer in your own words, a short dialogue that argues against your reasoning, an exercise that checks a skill. The schedule is handled for you with FSRS, the same modern algorithm behind current Anki.
| Step | The old way | In one page |
|---|---|---|
| Capture | Highlight, file away | Write or import |
| Turn into review | Build cards in another app | The page does it |
| Remember | Hope you reopen it | It comes back on schedule |
The point is not the tooling. The point is that the thing you wrote becomes the thing that teaches you, with no second system to maintain.
Where to start
Pick one thing you are reading this week. Instead of highlighting, stop at the end of each section and write, in a sentence or two, what you would want to remember. Then let it come back to you a few days later and see how much survived.
That single loop, capture then retrieve then space, is most of what works. Everything else is just making it easier to keep doing.