You opened the app to check one thing. An hour later you were still in it, and you could not say what you saw. Nothing in the feed told you when to stop.
Wanting, not liking
The chemistry of a scroll is anticipation, not satisfaction. The pull comes before the payoff, which is why finding the good post never makes you stop: it was the possibility of the good post doing the work, and there is always another possibility one thumb away.
Most of what you scroll past is nothing. You keep going because the next one might not be.
Unpredictable rewards
A reward you get every time stops driving behaviour once it stops arriving. A reward you get at random intervals keeps driving it long after.
Feeds are the second kind. You do not know which scroll surfaces the thing worth seeing, and the uncertainty is the hook. A feed that gave you exactly what you wanted every time would be one you finished and left.
The missing stop signal
Websites used to have pages. You reached the bottom, and deciding whether to click Next was a real decision, which made it a natural exit.
Infinite scroll removed the only stopping point in the experience.
Your brain runs on end signals. A book has a last page, an episode has credits, a meal has an empty plate. A feed has none of those, so the activity never resolves and you keep looking for an end that is not coming.
It costs more than the minutes
Put the phone down and part of your attention stays behind with whatever you just saw. The next thing you try to do gets a smaller version of you, for longer than the interruption lasted.
The bill is not just the time in the app. It is the quality of everything after it.
This is the business model
Feeds make money from attention, and attention is measured in time in the app.
The people building the feed are optimising for how long you stay. You are trying to leave. The match is not even.
Why willpower is the wrong tool
Telling yourself to just put the phone down is like telling yourself to just walk out of the casino. Technically available. Practically, the whole room was built to prevent it.
Your judgement at 11pm is not your judgement at 9am, and 11pm is when most of the scrolling happens. The app’s engineering does not get tired.
The answer to an engineered environment is a counter-system, not more effort. Decide the terms while you are clear-headed and let something else hold them when you are not. Bedtime is where this pays first.
What actually works
Every intervention that holds has the same shape: it changes the environment, not the person.
Remove the app from the home screen. Put the phone in another room. Block it at the system level so it cannot launch at all. The pattern is consistent. Friction works and willpower does not, and the only question left is how much friction you need.
