Rose Horowitch's Atlantic piece landed in the fall of 2024 and a lot of teachers read it twice. The title was "The Elite College Students Who Can't Read Books." It argued that students at top universities are showing up unable to sustain attention across a novel. They can parse paragraphs. They can write a thesis. They can't stay with three hundred pages.
The piece got criticized, fairly, for being anecdotal. Mostly interviews with professors. No comprehensive dataset. The criticism is right.
But the people who teach reading for a living are watching it happen.
My first reaction was the one most teachers had: yes, that's what we're seeing. My second took longer to arrive.
We trained them for this.
A few years ago David Rickert wrote a piece called "Is Annotating Necessary?" He compared annotation, as it's currently assigned, to making a bed. Students do it because an authority told them to. Not because it makes any internal sense to them. They produce a clean, performed, compliant artifact. The bed is made. Nothing is happening inside it.
That image sat with me for a year. It was the cleanest description of what I'd asked hundreds, maybe even thousands of students to do, year after year, without quite understanding why the work in front of me felt empty.
We asked them to annotate. They made the bed.
The memory I keep coming back to is from 2011. Early in my teaching career. I was teaching The Giver.
How was I teaching it? Book in their hand. Pen in the other. Annotation requirement: a certain number of marks per chapter, due Monday, checked Tuesday.
What I got back was exactly what I asked for. Pages full of underlines, "interesting," "important," marginal arrows pointing at random sentences. Students who hit the count. Students who didn't. Students who'd clearly filled the count up in the parking lot before first period.
I'll never forget Jason. He was fourteen. Most of the time I did annotation checks by flipping pages and counting marks, the same compliance posture I was asking from him. One afternoon, for no particular reason, I stopped on one of his pages. Then another. Scribbled words. Pages dog-eared in places that didn't land on anything. I don't think Jason had been reading the book at all. He'd been producing the artifact I was grading. That was the moment it clicked.
The Giver is a book about people going through the motions of a life without seeing it. I was asking students to go through the motions of reading the book without seeing it. The assignment was reproducing the exact failure mode the novel was warning them about.
This wasn't a personal failure. It was a design failure of the assignment.
The kids were optimizing for a rubric. The rubric rewarded marks-per-page density. Density is what they produced. The kind of annotation that actually builds reading, writing-while-reading that makes a student stop at a critical moment and take a position, predict, doubt, connect, argue back, is hard to assign legibly. You can't put a count on real thinking. So most teachers, myself included, fell back to what we could grade.
What students produced was almost all restatement. "Jonas is the main character," underlined. "Jonas's father releases the twin." "Important." None of these are thinking events. They're evidence that the student saw the words on the page, which we already knew because the words were on the page.
A 2013 review of common study strategies in Psychological Science in the Public Interest rated highlighting and underlining among the least effective techniques studied. The two that scored highest were practice testing and elaborative interrogation, the academic name for stopping every few paragraphs and asking yourself why something in the text is true.
The bed-making wasn't building readers.
What we should ask for instead isn't more annotations. It's different ones.
A real annotation has the student writing down something that didn't exist in their head a second before they wrote it. A question. A doubt. A prediction. A connection. A small argument with the author. Not "the twin is released," but "why doesn't Jonas's father look horrified?" Not "Jonas is special," but "I think the Giver is going to die soon and Jonas is going to have to leave. The book has been priming me for it since chapter eight."
That kind of writing can't be graded by count. It has to be read by a human and engaged with as thinking, which is harder for the teacher.
We mostly took the easier deal.
I taught a book about a society that stopped asking questions to a roomful of teenagers, using an assignment design that asked them not to ask any. Hit the count, get the grade, move on. The book and the assignment were arguing with each other, and the assignment was winning.
We trained students to produce evidence of compliance with a reading assignment. We didn't train them to be in a fight with the page. That's the missing skill, and it's the one Horowitch's piece is watching go missing in college classrooms.
In 2026 the cost is higher than it was in 2011. We're watching an entire culture decide what to do with tools that produce confident-sounding text on any topic in seconds. Students who can argue with a sentence are going to do something different with those tools than students who learned to underline anything that looked important.
I'm not in the classroom anymore. I'm building tools that try to make the writing students do while reading be the actual cognitive work, not the proof-of-reading work. Prompts that demand prediction, position, pushback. Not marks per chapter.
I don't think it solves what Horowitch was watching. The chasm is bigger than any tool. But assignment design is the lever we have, and assignment design has been wrong for a long time.
Most of the students I taught The Giver to in 2011 are in their late twenties now. I hope they're better at arguing with the page than I trained them to be. I hope they figured out, on their own, what I was supposed to teach them. Jason most of all.
The book is about people sleepwalking through their own society. The teacher was asking them to sleepwalk through the book. It took most of my career, and a piece in the Atlantic about kids who couldn't finish a novel, for me to see what Jason had been showing me on those pages all along.
Tools, demos, and what I'm working on now: readbookpulse.com