
The Long Read
The page that never ends.
Infinite scroll was designed to spare us a pointless click, and it succeeded so well that regulators now treat it as a hazard. Its history is a lesson in how small design choices decide who is in control of a screen.
Written by CONTROL
This article runs 1,360 words and takes about six minutes to read.
For most of the history of reading, the page ended. A book had a last line at the bottom, a newspaper had a fold, and a magazine had a back cover. The early web inherited the habit. Search results came ten to a page, forums were numbered, and when you reached the bottom of something you had to decide, however briefly, whether to click for more. That small decision was a stopping cue, a point at which the reader was handed back the choice. Almost nobody thought of it that way at the time. It looked like friction, and friction was the thing designers were paid to remove.
Two inventors and one idea
The man most often credited with removing it is Aza Raskin. His father, Jef Raskin, started the Macintosh project at Apple and spent his later career arguing that computers should adapt to people rather than the reverse. After Jef Raskin died in 2005, his son founded a small company called Humanized to carry on that work, and it was there, in 2006, that Aza Raskin designed a page that loaded new content automatically as the reader approached the bottom. His reasoning, as he has told it since, was a designer’s reasoning: every time a person is forced to make a choice he does not care about, the interface has failed him. The “next page” button was such a choice, so he took it away.
The credit is not his alone. Hugh Williams, who led search work at Microsoft, has written that he and two colleagues, Nick Craswell and Julie Farago, built continuous loading into MSN’s image search in 2005, and Microsoft filed a patent on the approach in early 2006. Their reason came from data. On web search, most queries never went past the first page of results, but image searchers kept going, and a pagination button only got in their way. That story is worth keeping because it corrects a myth. Infinite scroll was not a stroke of genius or of villainy. It was an obvious improvement, arrived at independently by several people, for readers who clearly wanted more.

The feed arrives
What changed its meaning was the thing it was attached to. In September 2006, Facebook introduced the News Feed, which gathered friends’ activity into a single running column. Twitter’s timeline, Tumblr’s dashboard and Instagram’s grid followed the same logic. At first these feeds were chronological, and a chronological feed has a natural floor: eventually you reach the post you saw yesterday and know you are caught up. Then the platforms began ranking. Twitter introduced an algorithmic timeline in early 2016, and Instagram announced in March 2016 that posts would no longer appear in the order they were published. Once a feed is ranked rather than dated, there is no yesterday to reach. There is only the next thing the system predicts you will stop for.
Infinite scroll plus a ranked feed is a different machine from either part. The scroll removes the pause, and the ranking guarantees that the material on the other side of the pause is chosen to hold you. TikTok, which merged with Musical.ly in 2018 and brought the full-screen vertical feed to the West, perfected the combination by removing even the decision about what to watch. The video starts; the next one is waiting beneath it; the thumb does the rest.
Once a feed is ranked rather than dated, there is no yesterday to reach.
The economics of attention
None of this was mysterious to economists. In 1971 Herbert Simon, who later won the Nobel Prize, observed that “a wealth of information creates a poverty of attention.” Advertising-funded platforms turned that observation into a business model. If revenue depends on the number of advertisements seen, and the number seen depends on minutes spent, then every design choice will be tested against minutes spent, and the choices that survive will be the ones that keep people longest. The designers who made those choices were not, for the most part, cynics. They ran experiments, kept the versions that produced more engagement and moved on. The outcome was an interface optimized, variant by variant, against the user’s own ability to put it down.
Raskin has spent much of the past decade saying so. In a 2018 BBC interview he compared some engagement techniques to “behavioural cocaine” sprinkled over an interface. In 2019 he wrote that one of his lessons from infinite scroll was that “optimizing something for ease-of-use does not mean best for the user or humanity.” He has described being blind to the way his good intentions became irrelevant once the feature was placed inside a system with an incentive to capture attention. He went on to co-found the Center for Humane Technology with the former Google design ethicist Tristan Harris, which argues that platforms should be judged by what they do to people rather than by how long people stay.

Who holds the controller
Anyone who grew up with a console understands the distinction at stake. A game is also designed to hold attention, and some games are shameless about it. But the classic game had a shape. It had levels that ended, a save point where the player could stop, a boss that could be beaten and credits that rolled. Even the arcade, which was built to take quarters, had a game-over screen that returned the decision to the person standing at the cabinet. The design assumed there was someone outside the machine who would, at some point, want to leave, and it made room for him.
The endless feed makes the opposite assumption. It has no level to finish and no credits. There is no moment built into it at which the person holding the phone is invited to consider whether he wants to continue. That is the plainest way to describe what design decides: whether the system hands the choice back to the user at regular intervals or keeps it for itself. A “load more” button, a daily digest, a feed that says you are caught up, autoplay that is off by default, a timer that ends a session: these are not features that stop anyone doing anything. They are moments at which the person is asked, and being asked is the whole of what control means on a screen.
Regulators discover the stopping cue
For years this argument lived in essays and documentaries. It has now reached law. On February 6, 2026, the European Commission issued preliminary findings that TikTok had breached the Digital Services Act through addictive design, naming infinite scroll, autoplay, push notifications and its highly personalized recommender system. The Commission said TikTok would have to change the basic design of the service, including by disabling infinite scroll over time and introducing effective screen-time breaks. On July 10, 2026, it reached similar preliminary findings about Instagram and Facebook. Under the act, a final finding of breach can carry a fine of up to 6 percent of a company’s worldwide annual turnover.
It is striking that the remedy under discussion is, in effect, the restoration of the page. After twenty years, the feature that was removed as a nuisance is being reconsidered as a safeguard. That does not make the original designers wrong about the nuisance. The pagination button really was an irritation on an image search, and nobody wants to go back to clicking through thumbnails a page at a time. The mistake was in treating every pause as waste, as though the only thing a reader could want at the bottom of a page was more of the same.
The practical lesson for anyone holding a phone is modest and a little dull. Choose tools that end. Prefer the newsletter that arrives once to the feed that refreshes forever, the album to the autoplay queue and the game with a save point to the one with a daily streak. Turn off autoplay wherever the setting exists, and notice which applications refuse to offer it. The page used to end on its own. Now the reader has to build the ending himself, and the people who do so are the ones still holding the controller.