The Ultra-Processed Media Diet: Why Short-Form Video Feels Like Junk Food for Attention

I finished a stretch of short videos the other night and could not tell you what a single one of them was about. Not one. Forty minutes, gone, and no residue — no idea I could turn over the next day, no image that stayed, not even a strong opinion. Just the vague sense of having eaten something.

That last part is what interests me. The sensation is not "I wasted time." Wasting time on a long novel or a bad film at least leaves you with the film. This is something else: consumption that goes down easily, ends without a natural stopping point, and leaves almost nothing behind. Psychologists and media researchers have started reaching for the same comparison, and it turns out to be more precise than a throwaway insult.

Short-form video, the argument goes, is to attention what ultra-processed food is to nutrition.

Where the comparison actually comes from

"Ultra-processed" is not a synonym for unhealthy, and it is worth getting the definition right before borrowing it, because the borrowed version is usually sloppier than the original.

The term comes from the NOVA classification, developed by Carlos Monteiro and colleagues at the University of São Paulo. NOVA sorts food not by nutrient content but by degree and purpose of processing. Group 1 is unprocessed or minimally processed — an apple, rice, milk. Group 2 is culinary ingredients like oil and salt. Group 3 is processed foods: bread, cheese, tinned beans, things made by combining groups 1 and 2 in ways a home kitchen could manage. Group 4, ultra-processed, is the interesting one: industrial formulations built largely from substances extracted from foods, plus additives whose job is to make the result hyper-palatable, shelf-stable, and convenient. The defining feature is not that it is bad for you. It is that it is engineered, and engineered against an objective that is not your wellbeing.

What makes the category more than a taxonomy is that it predicts things. Kevin Hall's team at the US National Institutes of Health ran a tightly controlled trial in 2019 in which participants lived in a research facility and ate either ultra-processed or minimally processed diets matched for calories, sugar, fat, fibre and sodium. On the ultra-processed diet, people ate roughly 500 more calories a day and gained weight. Same nutrients on paper. Different behaviour in practice. The processing itself was doing something — softer textures, higher energy density, faster eating rate, weaker satiety signals.

That is the finding the media version of the metaphor is really borrowing. Not "junk food is bad." Something sharper: a formulation optimised for consumption will get consumed past the point where it is doing you any good, and the eating will not feel like overeating while it happens.

Why it holds up better than most food analogies

Media metaphors usually collapse under a little pressure. This one survives four specific parallels, and it is worth being precise about which ones.

It is formulated, not merely produced. A short-form feed is not a channel that happens to be popular. It is a system with a measurable objective — watch time, completion rate, session length — and a fast feedback loop for hitting it. Every element is a variable: the length of the clip, the cut on the first frame, the audio hook, whether the thing autoplays. The comparison to a snack food engineered in a lab against a bliss point is not rhetorical. It is a description of the same practice applied to a different substrate.

It is hyper-palatable. Ultra-processed food concentrates the sensory rewards — salt, fat, sugar, crunch — and strips the parts that slow you down. Short-form video does the analogous thing to narrative: the punchline without the setup, the reveal without the build, the emotional payload with the context removed. A three-minute clip of a film's best scene is genuinely more immediately pleasurable than the film. It is just not the film.

It has almost no satiety signal. This is the parallel that convinces me most. Whole foods tell you when to stop, through volume and fibre and the sheer time chewing takes. A liquid meal with the same calories does not. Long-form media has natural terminators: the chapter ends, the episode ends, the credits roll, and in that gap you get a moment to decide whether to continue. An infinite feed has removed the gap deliberately. There is no unit. Nothing ever ends. The decision to keep going is never actually presented to you, so you never make it.

It displaces rather than poisons. This is the honest version of the harm claim, and it is less dramatic than the usual one. The problem with ultra-processed food, in population terms, is largely that it crowds out the vegetables. Nobody is claiming a single biscuit is toxic. Likewise, forty minutes of short video is not damaging you; it is occupying the slot where the thing that would have fed you might have gone. The cost is an opportunity cost, and opportunity costs are invisible by nature, which is exactly why they accumulate unnoticed.

Where I would stop the metaphor

Metaphors are useful right up to the point where people start treating them as mechanisms, and this one has an obvious failure mode.

There is no nutritional science of media. There is no attentional equivalent of a calorie, no measured deficiency state, no randomised controlled trial in which people are housed in a facility and fed short-form video for fourteen days. The Hall trial is real; there is no counterpart for the feed. Anyone telling you exactly how much short-form video is safe per day has invented the number.

I would also push back hard on the dopamine talk that usually rides along with this comparison. The popular version — that short video "spikes dopamine" and leaves you "dopamine depleted," requiring a "dopamine detox" — is not what the neuroscience says. Dopamine is central to motivation and to learning from prediction error; it is not a fuel tank that empties. You cannot detox from a neurotransmitter your brain manufactures continuously. When someone reaches for that framing, they have usually stopped describing your experience and started decorating it with borrowed vocabulary.

And the food comparison carries moral baggage worth refusing. Diet culture has done real harm by attaching shame to eating, and importing that structure into media use is a mistake. The interesting question is not whether you are being disciplined about your feed. It is whether what you are consuming is giving you back anything proportionate to the time you are handing it.

Rest and this are not the same thing

Here is where I think most people, myself very much included, get genuinely confused.

The feed gets reached for at moments of depletion. The end of a long day. The ten minutes after putting a child to bed when you are too tired to start anything. It presents itself as rest — it is passive, it asks nothing, it requires no decisions. And it is not rest, in the same way that an energy drink is not sleep.

Rest, as far as I can tell from paying attention to it, has a specific quality: your attention stops being pulled and comes back to you. A walk without headphones does it. Sitting quietly does it. Actual boredom, uncomfortable as it is at first, does it — the mind, given nothing to consume, starts producing. This is not mystical. It is close to what people report from any contemplative practice, and my own experience with Heartfulness meditation lines up: the first few minutes of sitting are dominated by the residue of whatever I was looking at beforehand, and if that residue is forty minutes of short video, the sitting is noticeably harder. Content that arrives in three-second cuts leaves an afterimage that takes a while to settle.

Scrolling does the opposite of rest. Attention stays externally driven, just at low intensity. You end the session neither restored nor stimulated, in that specific flat state that is hard to name and easy to recognise. Which is why "I'll just relax with my phone for a bit" so often ends with you feeling exactly as tired as before, only later.

The distinction I have found useful: rest is when the input stops. Entertainment is when the input is good. Scrolling is neither, and it costs the same time as both.

Building a more whole diet without going cold turkey

Abstinence advice is popular because it is easy to write and easy to fail at. Delete the apps, throw the phone in a drawer, take a thirty-day detox. Some people do this successfully and I would not argue with them. Most reinstall within a week and add a small failure to the pile.

The nutritional framing suggests a better move, and it is the one that actually worked for me: add before you subtract. Nobody improves a diet by staring at the biscuits harder. They improve it by having something else in the house that they actually want to eat.

Some practical versions.

Keep something long-form genuinely within reach. Not aspirational — actual. A book you want to read, not one you think you should have read. The reason the feed wins at 9pm is not that it is more compelling than a good novel. It is that it is closer. Reduce the distance to the thing you would rather do and the competition changes.

Watch for the swap, not the total. Fifteen minutes of short video while queuing for coffee costs you almost nothing. Fifteen minutes in bed after you had decided to read costs you the reading and some of the sleep. The same quantity, wildly different price. Track where it lands, not how much of it there is.

Restore the endings. Since the feed has removed all natural stopping points, you have to reinstall one. A timer works. So does the cruder version: watching in a context that ends on its own — standing up, waiting for the kettle, the train arriving. Anything with a built-in terminator beats a decision you have to make repeatedly against a system designed to prevent it.

Notice the residue, not the moment. During, it is fine. It is always fine during; that is the whole design. The signal is afterwards. Ask yourself an hour later what you actually retained. That question is the closest thing to a satiety cue this medium has, and it is the one habit I would keep if I could only keep one.

Eat some rough food on purpose. A long article that takes effort. A film with a slow first half hour. A conversation with no phone on the table. This is the fibre. It is not more virtuous, it is just structurally different — it demands something of you, and the demand is what makes it satisfying.

A weekly media audit, modelled on a food log

Food logs work not because the log itself changes anything, but because attention does. People eat differently when they are writing it down, and stop when they stop. The same trick transfers, and the same limitation applies — so do this for one week, not forever.

For seven days, keep a note on your phone or a page in a notebook. Every time you consume media deliberately — not every glance, but any session of five minutes or more — write one line with four things:

1. What it was. Two or three words. "Short video," "podcast," "novel," "news," "long YouTube essay."

2. Roughly how long. Estimate. Precision does not matter and chasing it will make you quit by Wednesday.

3. Why you started. This is the important column and the one people skip. Was it boredom, avoidance, tiredness, genuine interest, or a notification? Be blunt with yourself. "Avoiding the thing I said I'd do tonight" is a real and common answer.

4. What you had an hour later. One phrase. Nothing at all is a legitimate entry and will be your most frequent one. "An idea I want to test." "A scene I keep thinking about." "Irritated for no reason." "Blank."

At the end of the week, do not add up the hours. That is the diet-culture move and it teaches you nothing except how to feel bad. Instead read down column three and column four together, and look for the pairing.

What you are hunting for is the combination of a depleted reason and a blank residue. That is the pattern worth changing, because it is the one where you paid real time and received nothing at all. Sessions that started from genuine interest usually leave residue, even short ones. Sessions that started from tiredness or avoidance usually do not, regardless of what you watched.

And you will almost certainly find that the medium matters less than the entry point. The same app is a decent tool at 11am with a purpose and a hole in the day at 11pm without one. That is a much more useful thing to learn than a total number of hours, and it points at a fix you can actually make.

Questions worth answering honestly

Is short-form video actually harming my attention span?
The evidence is much weaker than the confident headlines suggest. Attention is not a single measurable capacity that erodes like enamel, and the studies people cite for this usually show correlation in self-reported measures. What I would say with more confidence: heavy feed use trains an expectation of frequent novelty, and things that violate that expectation — a long chapter, a slow film, an unglamorous problem at work — start to feel more effortful than they used to. Whether that counts as damage or as a habit is a real question, and I lean toward habit, which is a more hopeful diagnosis.

Isn't this just the same panic people had about novels and television?
Partly, and that history should make anyone cautious. But there is a difference in kind worth naming: previous media were not adaptive. A television schedule did not learn from you and reorganise itself in real time to keep you watching. The moral panic was recycled; the feedback loop is new.

What about educational short-form content?
Better than nothing, and not the same as learning. The format can point you at things — I have found books and ideas that way. What it cannot do is the part where understanding gets built, which requires sustained attention on one thing and is intrinsically slow. Treat the clip as a menu, not a meal. The trouble starts when reading the menu starts to feel like eating.

Do I have to give it up?
No, and the framing that says you do is part of why people bounce off this whole conversation. Nobody sensible tells you to never eat a biscuit. The question is proportion, and specifically whether the effortful things still have room. If a long book, a real conversation, or a slow evening still happens sometimes, your diet is probably fine. If they have quietly disappeared, that is worth noticing — and the noticing is most of the work.

What stayed with me from the audit was not the hours. It was the third column, and how often the honest answer to "why did I start" was simply that I did not want to be alone with the day yet.