Does the endless-feed format train you to read worse over time?

Honest answer: that is still a hypothesis. The shallowing account says fast, fragmented digital reading nudges you toward shallower processing, and the evidence behind it is small and mixed. What is well measured is behavior: most bookmarks never get reopened and most reading time stays at the top of the page.

Why does the feed make me scroll past articles I actually wanted to read?

Because saving is cheaper than reading right now, and the feed always has a next thing ready. You are not deciding against the article, you are postponing it, and the save click feels enough like progress to end the thought. One reader described the loop exactly: scroll past, save for later, never come back.

Why can I picture the screenshot but not remember the article?

Because you kept the souvenir, not the idea. Recognition feels like knowledge. You remember the highlight, the layout, where you saved it, and none of that is the argument. Recall is the different test, and the only one that tells you what landed.

Why You Scroll Past Articles and Remember Nothing

Why do I scroll through articles and remember none of them?

Because a feed has no ending and no middle. Eyetracking data puts about three quarters of reading time in the first two screenfuls, so you absorb the top of every article and the argument in the middle goes unread. Nothing ever closes, and saving quietly replaces the reading you meant to do.

The number comes from Nielsen Norman Group's eyetracking research on scrolling: across more than 130,000 fixations from 120 participants, 57% of page-viewing time landed above the fold and 74% inside the first two screenfuls. The remaining 26% was spread down the page in a long tail, which is where the argument is actually made. Inside a feed, every article's fold is somebody else's headline, so the top of everything gets read and the middle of nothing does.

The missing ending matters just as much. A printed article ends, a chapter ends, and that end is not decoration. It signals your brain that nothing more is coming and it is safe to summarize. Infinite scroll deletes the signal. Each article's bottom is the next article's top, so the reading never terminates and the quiet "what did I just read" moment never arrives. You end up with a strange amnesia: you have seen all of it and closed none of it.

Saving substitutes for reading instead of enabling it. Bergman, Whittaker and Schooler asked 50 participants to refind 250 web pages they had bookmarked, and found that only 41 of them, about 16%, were ever retrieved through the browser's bookmark facility (Journal of Librarianship and Information Science, 2021). The save click delivers the completion feeling the article was supposed to deliver, and the article waits for a version of you with more time.

What happens to comprehension when you read inside a scrolling feed?

You lose coverage, not memory. You are not forgetting a finished piece of reading. You are reading the opening of five pieces and completing none of them, so there is no whole shape for your brain to store. The feed turns reading into a stack of beginnings, and beginnings are easy to remember and useless to quote.

Two things make it worse, and neither is about screens damaging your brain. The first is time pressure. In one experiment, 140 undergraduates read in print or on screen, with or without a deadline. The print readers tightened their focus when time got tight, while the screen readers under the deadline understood less than every other group (Delgado and Salmerón, Learning and Instruction). A feed is a deadline you never set, with something new waiting in every pause, and related studies find digital readers more overconfident about what they took in when time is short (Ackerman and Lauterman, 2012). The honest caveat: the paper-versus-screen gap overall is small and uneven, around two tenths of a standard deviation in Delgado's 2018 meta-analysis, and largest for factual text against the clock.

The second is the shape of the reward. In a feed, the next post is always cheaper than the article in front of you, so the piece you actually wanted gets postponed rather than rejected. The newest read-later tools push that logic one step further: they chop saved articles into feed-sized snippets and slide them back into your timeline so the backlog clears without leaving the scroll. That solves the backlog by removing the last reading surface you had.

How do I break the scroll-without-comprehension loop?

Give one article a surface with an end, and then retrieve it. Read one piece on its own, finish it, and write three lines in your own words before you swipe away. Regenerating an idea from memory is what makes it stick, and it is the step a feed skips every single time.

The shortcut came from a PhD student on a productivity forum who could not retain papers. He described it exactly: "I'll read something closely, genuinely follow the argument, feel like I've got it. A week later I know I read it, I have the highlights, and the content is gone." The fix that worked was not another app. It was three lines in his own words right after finishing, and it matches the classic evidence: in Roediger and Karpicke's testing-effect experiments, students who tested themselves recalled 61% of a passage a week later, while students who only restudied recalled 40%, even though the rereaders had read it 14 times and the testers 3.4 (Psychological Science, 2006). Retrieval is what makes an idea durable. Re-exposure mostly makes it familiar, and familiarity is what lets you walk away believing you understood.

Every thread on this topic ends at the same gap. In one minimalism discussion, the top answer was save everything to Instapaper and read in batches, followed by the admission: "I can't really help you with the remembering part... you have to do that one on your own." Batching fixes volume and does nothing for retention unless the batch ends in three lines of retrieval.

Policy has started to treat the format itself as the problem. California's AB 1709, signed on September 10, 2026, bans infinite scroll and autoplay for users under 16. Adults get no such protection and no equivalent measurement, which is the real gap: your reading environment keeps changing and nothing tells you what actually landed. That is the surface Absorb is built for: one reading, listening, or analysis challenge a day, about 10 to 12 minutes, a real end, and three scores for what you actually absorbed. See Absorb's daily comprehension workout on the App Store.