Notes on why media stopped being something I chose and became the thing I live inside, and what it takes to get back down.
The premise
I went to bed watching short videos, woke up watching them, kept them playing while I got ready, and kept them playing while I drove. There was no gap anywhere in my morning. I noticed it all at once one day, and what struck me wasn't that I used my phone a lot. It was that there was no longer any part of the day that wasn't mediated.
Here's the thing I actually noticed, and it's the part that generalizes. Working. Driving. Lying around. At the gym. Cooking. Falling asleep. The activity underneath keeps swapping out, all day long, and the audio never does. It hands off from one to the next without a gap. I wasn't choosing something to listen to for each activity. There was one continuous stream running, and the activities were happening inside it.
Most advice about this treats it as one problem. It isn't. There are two things happening at the same time, and they move independently. Until you can see them apart, nothing you try will hold.
The horizontal axis is saturation. How much of your life has something playing in it.
The vertical axis is density. How much stimulation is packed into each second of what's playing.
Every attempt I made failed because I only ever moved on one axis. That turns out to be the whole story.
Horizontal
Saturation
For most of the last century, media was a destination. You sat down for it. It had a start time and an end time, and when it ended you got up and did something else. It occupied a bounded slot in the day.
Then it became portable, and it stopped being a place you went. It became the thing that filled the gaps: the waiting room, the line at the counter, the four minutes before a meeting. Still secondary. Still filling something.
The third stage is the one almost nobody names, and it's where I ended up. Media stops filling the gaps and becomes the continuous background, and real life becomes the thing that interrupts it. Shorts playing while I got ready weren't filling my morning. My morning was happening inside them.
The measured version: DataReportal's 2026 global report puts average daily internet use at 6 hours 38 minutes, with roughly 2 hours 21 minutes of that on social platforms alone.2 Those aren't numbers you reach by filling gaps. There aren't enough gaps.
Vertical
Density
Independently of how much of your day is covered, the stuff covering it keeps getting more concentrated. Same minute, more stimulation packed into it. That's a ladder, and it only goes one direction on its own.
Bar weight = stimulation per second
The clearest evidence that this is still moving is on your television. Broadcast was the original bounded format. It had a schedule, it ended, and when it ended you stood up. It's now one of the more common places to find an infinite feed of short video. The ladder didn't simply grow another rung at the top. It reached back down into the one format that had a stopping point built into it by physics, and took the stopping point out.
The engine under the ladder isn't mysterious and it isn't a conspiracy. It's a business model. Every second of attention is sold, so there's continuous, well-funded pressure to make each second denser than the last. You aren't competing against your own weak character. You're competing against an optimization process with a budget that grows every quarter.
The specific mechanism is old. B.F. Skinner showed that rewards delivered on an unpredictable schedule produce behavior that's stronger and far harder to extinguish than rewards delivered reliably. An infinite feed is that schedule. Most items are nothing, some are good, and you can never tell which is coming, so you keep pulling.4 A slot machine is the honest comparison, and it's the one researchers actually use.
The mechanism
Here's the part that took me years to see. The axes aren't independent in practice. They ratchet.
Higher density raises your baseline. Once your reference point moves, unstimulated time feels worse than it used to, so you cover more of the day. More coverage means more hours at high density, which moves the baseline further, which sends you up another rung. Round and round.
This explains the failure mode I repeated for a decade. Deleting an app is a step down the ladder while saturation stays at one hundred percent. The appetite is untouched. It reroutes immediately into something that feels virtuous, usually a podcast or an audiobook, and because that substitute feels productive, nothing in you resists it. The demand keeps getting fed and trained the entire time you believe you're quitting. Then a hard day arrives, willpower dips, and it goes straight back to the densest thing available.
It also explains the motivation problem, which I think is the most underrated cost of all this. If you spend your day at maximum density, ordinary things go flat by comparison. Real work, real conversation, a real evening. None of it's boring. Your reference point moved. That's a measurable, reversible thing, not a character defect, and it's the reason "I want to do this so badly and I can't start" is such a common sentence among people in stage three.
Background
Research
I'm not a scientist, so I've tried to keep every claim here tied to work somebody else did and anybody can go read. Four findings carry most of the load.
We're bad at being alone with nothing. In 2014, Timothy Wilson and colleagues ran eleven studies in which people were left in a room for six to fifteen minutes with nothing to do but think. Participants disliked it consistently. In one condition, given the option, many chose to administer themselves a mild electric shock rather than sit with their own thoughts, including a majority of the men tested.1 This was before short-form video existed. The appetite was already there. What changed since is that something arrived to feed it perfectly.
Unpredictable rewards build the most durable habits we know how to build. Variable-ratio reinforcement produces behavior that resists extinction better than any other schedule, and the design of infinite feeds maps onto it almost exactly: intermittent likes, unpredictable payoffs, no natural stopping point.4 This isn't an accident of the technology. It's the part that works.
Phones in bed cost you real sleep, by more than one route. A 2024 systematic review and meta-analysis of electronic media use and sleep quality found consistent negative associations, operating through at least three mechanisms: straightforward displacement of sleep time, light exposure suppressing melatonin, and emotional arousal from the content itself making it harder to come down.3 The arousal pathway is the one people forget, and it's the one that makes "I'll just watch something relaxing" not work.
Boredom is doing something, and you lose it when you cover it up. Mann and Cadman found that people made deliberately bored beforehand generated significantly more creative responses afterward, with passive boredom producing a stronger effect than active boredom.5 The proposed mechanism is the default mode network, the diffuse mode of thought that runs when you're not focused on anything in particular. It's where planning, memory, and unprompted insight seem to live. Fill every gap and you don't just lose rest. You lose the mode of thinking that only runs in the gaps.
Protocol
Two axes means two kinds of move. A horizontal move reclaims a window of your day. A vertical move drops a rung of density. Neither one survives on its own, which is the single most important thing on this page.
The sequence alternates for that reason, and the first two moves are both horizontal on purpose. Dropping a rung with no reclaimed windows underneath it's the dashed arrow in that picture, and it's the version that keeps failing.
The plan is fixed. The content is yours. Everyone's day holds the same five kinds of window, but they sit in different places, so name your version of each before you start.
Be specific. "Evenings" isn't a window. "10:40 until whenever I fall asleep, in bed, phone on the pillow" is one.
Independence
Less screen time is a metric, not a goal, and chasing a metric is how these attempts turn into a scoreboard you eventually stop looking at. Worth being clear about what the point is instead.
The same device is the best tool anyone has ever built for making things happen, and the worst thing anyone has ever built for occupying time. Those two facts aren't in tension. They're two completely different uses that happen to share a slab of glass, and almost all of the damage comes from failing to tell them apart.
A tool is something you pick up with an intent and put down when the intent is served. It waits for you. It doesn't ring, suggest, refresh, or autoplay. Judged as a tool, the phone is extraordinary: it holds your work, your maps, your people, your camera, your money. Nobody sensible is arguing you should give that up, and this site isn't going to.
An occupant is different. An occupant doesn't wait to be picked up. It arrives, it fills, it hands off from one activity to the next, and its entire job is to be the thing between you and any unfilled minute. It isn't measured in how much it helps you. It's measured in how much of you it holds.
That's what independence means here, and it's a smaller and more achievable thing than quitting. You aren't renouncing anything. You're separating a tool from an occupant that currently share a body, and then keeping only one of them within arm's reach when you've nothing to do.
What comes back isn't virtue. It's hours, and with them the diffuse, unfocused mode of thinking that only ever runs in an unfilled minute. That's where planning happens, where the thing you had been stuck on resolves itself, where you notice you've been unhappy about something for a month. None of it's available while something is playing. It isn't that you can't hear it. It's that it never starts.
The list
I'm doing all of this in public, including the parts that don't work. If you want the notes as they come, leave an email. Occasional, no more than a couple a month, and nothing else is ever done with the address.