You highlight half a book, then never see those highlights again. Not because you meant to abandon them - you fully intended to go back. But "go back" never turned into a scheduled moment, so it just didn't happen, and the highlights sit there, technically saved, functionally gone.
This isn't a discipline problem. It's what happens whenever capture and review are two separate, disconnected steps.
In short
- Highlighting is capture. It is the first half of remembering, not the whole of it.
- Highlights and saved articles are built to be filed, not to be brought back, so nothing ever asks you about them.
- Memory fades quickly after you read something, whether or not you marked it.
- To retain what you read you need two things: a small selection of what is worth keeping, and a scheduled way of being asked about it.
- Reviewing from the source, instead of a retyped copy, keeps the context that made the passage make sense.
Highlighting feels like progress
There's a reason highlighting is so satisfying in the moment: it feels like work. You're marking what matters, building a record, being the kind of reader who pays attention. And capture genuinely is the first half of remembering something - you can't retain a passage you never noted in the first place.
But capture isn't retention. It's the raw material for retention, and raw material left untouched doesn't do anything on its own. A highlighted sentence sitting in a book you've already returned to the shelf has the same effect on your memory as a sentence you never read at all: none, past the first pass.
Part of the trap is that highlighting produces a feeling of fluency. Passages you've marked look familiar when you see them again, and familiarity is easy to mistake for knowledge. Recognising a sentence is a different, much weaker thing than being able to say what it claimed. The test of whether you remember a book isn't whether its sentences look familiar; it's whether you can reconstruct its ideas with the page closed.
What happens after you close the book
Take a nonfiction book you finished a few months ago. Try to list its three main arguments, then one example the author used for each. Most people manage the first, partly, and rarely the second. This isn't unusual and it isn't about intelligence. It's the ordinary shape of forgetting: steep right after learning, slower later, and never zero. The forgetting curve describes it, and it applies to ideas as much as to word lists, though meaningful ideas tend to decay more gently than arbitrary strings.
Reading itself is also a one-shot exposure. You meet each idea once, in a single sitting, at the moment when the first-day loss is about to begin. Without a return trip, the best case is that a few vivid points survive by chance.
The gap between capturing and reviewing
The actual failure point isn't the highlighting. It's everything after it. Highlights, clipped articles, and saved quotes all share one property: they were built to be filed, not to be reviewed. There's no prompt that brings them back, no moment where the app asks you about that thing you marked three weeks ago. So the only way they resurface is if you go looking - and going looking is exactly the step that never happens on its own.
A highlight you never see again did less for your memory than no highlight at all - it just cost you more attention getting there.
Most note-taking tools are optimized for the capture half of this problem and silent on the second half. They'll hold onto anything you throw at them, indefinitely, without ever bringing it back up. That's a filing cabinet, not a memory system - and a filing cabinet you never reopen might as well be empty.
Rereading your highlights in bulk doesn't solve it either. A long list read top to bottom is passive: your eyes move, nothing is retrieved. The version that works is being asked a question and trying to answer it first. That distinction is the subject of active recall vs spaced repetition.
Choosing what is worth keeping
Not every highlight deserves a place in a review queue. A useful filter is a single question: would I want to be able to recall this in six months? Quotes that are merely nice, facts you can look up in seconds, and details that serve only one chapter usually fail it. Ideas that change how you think, definitions you will use, arguments you may need to make, and numbers you will actually cite usually pass.
A rough habit that works: after a reading session, pick the two or three passages that pass the filter and capture only those. A small, well-chosen set that you review beats a large archive you never open. If you do end up with a long backlog of material to bring in at once - say, a list of key terms from a textbook chapter - bulk import means getting it in isn't a second job.
Reviewing straight from the source
The fix isn't "highlight less" or "try harder to revisit your notes." It's closing the gap between capturing something and being asked about it again - automatically, without you having to remember to go looking.
In Cirello, a card doesn't have to be something you typed out by hand. It can prompt straight from the real source instead - a book, a specific page, a link - with your own note on what to focus on, so the review pulls you back to the actual material instead of a flashcard-ified paraphrase of it. You're not retyping the highlight into a separate system and hoping you maintain it. The source itself becomes the thing you get quizzed on, on a schedule that actually brings it back before you'd have forgotten it.
The step-by-step version of this is in how to turn book highlights into spaced-repetition cards, and there is a short overview on the remember book highlights page.
When a plain question-and-answer card is better
Source-linked cards suit ideas and passages. For crisp facts - a term, a date, a formula - a small front-and-back card is usually the right tool. The skill is in writing it so each card asks for one thing; how to write good flashcards covers that. Think of the two as complements: source cards keep the argument tied to its page, and short cards pin down the details you need to produce quickly.
Three passages, three treatments
Suppose you finish a book on how people form habits and you flagged three passages.
- A definition - "a habit is a behaviour repeated until it needs little deliberate thought". This is a crisp fact. A small question-and-answer card does the job: front "What defines a habit?", back the definition in your own words.
- An argument - a page that walks through why changing the environment beats relying on willpower. This one loses its force if you shrink it to one line. A source-linked card works better: it points at the page, with a prompt such as "summarise the argument about environment", and you check your summary against the real text.
- A striking example - a story that illustrates the argument. You don't need to memorise it word for word; you need to be able to retell it. A card that points at the page with the prompt "retell the example" gets you there.
Three passages, three different cards, and about ten minutes of work. Nothing was retyped wholesale, and each card tests the thing you actually want to keep.
A reading routine that actually sticks
- While reading: mark freely, but keep it light. Don't stop the flow to build cards.
- Right after a session: spend five minutes choosing the few passages that pass the filter and capture them with a page reference.
- Daily: a few minutes of review. FSRS-6 scheduling brings each item back near the point of forgetting, so the daily queue stays small.
- Monthly: glance at what you are no longer reviewing and decide whether anything should leave your library. Selective beats exhaustive.
That's the difference between a highlight and something you'll actually remember: not how carefully you marked it, but whether anything ever brings it back to you again.