An average adult today consumes the equivalent of several paperback books every twenty-four hours, and remembers almost none of it. It's not a brain failure: it's the only way it has to stay operational. Cognitive saturation (overload of the attentional and memory systems from excess input) is managed with two valves — forgetting and simplifying — and when both run at their limit, the blockage appears. The phone replaced silence; AI promised to lighten the load and is filling the gap with more load.
The figure that already falls short
Bohn and Short, in their report How Much Information? 2009 for the University of California, San Diego, calculated that an average adult in the United States already consumed back then around a hundred thousand words a day, adding up television, radio, press, web, email, messaging and video games. A hundred thousand words is, measured in paperback novels, about four books. Four books every twenty-four hours.
The figure is from 2009. Before TikTok, before the infinite scroll (endless scrolling designed so the screen never ends), before messaging as a default work tool, before the pandemic and remote work. Hilbert and López, in Science in 2011, took the accounting to global scale and the curve wasn't pointing toward leveling off.
Today that calculation falls short for anyone with a phone in their pocket and an office job. The honest figure is closer to two hundred thousand, and it's not made of text alone: it's fifteen-second video fragments, notifications, images that cross the retina without permission, forwarded audios, screenshots, half-read threads.
The brain has four seats, not four thousand
None of that is remembered. And it isn't remembered because it can't be.
Working memory (the cognitive space where the information being used this instant is manipulated) operates with a handful of simultaneous elements. George Miller spoke in 1956 of the "magical number seven, plus or minus two," and later research, with Nelson Cowan at the front, has lowered that figure to four elements for strictly concurrent tasks. Four. Not four thousand. Four.
Everything that comes in above that competes for a seat, and most of it doesn't get one. The system discards before the subject is aware. It discards while the subject thinks they're paying attention. When someone says "I forgot," in the vast majority of cases nothing was forgotten: it never got in. The word "forgetting" is misapplied to material that was never encoded.
Forgetting is not a defect
This discard function is the first survival mechanism under saturation. And it's function, not breakdown.
Daniel Schacter devoted The Seven Sins of Memory in 2001 to explaining why the seven classic defects of human recall —transience, absent-mindedness, blocking, misattribution, suggestibility, bias and persistence— aren't failures but the necessary flip side of properties that keep the subject operational. A memory that retained everything wouldn't free up capacity to think. Remembering everything and thinking are incompatible operations in a finite-sized brain.
Whoever doubts it has the case of Solomon Shereshevsky, described by Aleksandr Luria across three decades in The Mind of a Mnemonist: a man with near-perfect episodic memory, incapable of abstracting, of recognizing faces under changing light, of understanding a metaphor because each word evoked literal sensations that blocked the transfer. His memory was his disability.
The second valve. Simplifying
Forgetting is the first of the two valves. The second is simplification.
When discarding isn't enough, because the material that does get in is still more than can be handled deliberately, the system turns to heuristics (a fast mental rule that replaces reasoned analysis with a cheap shortcut). Kahneman christened them almost fifteen years ago, in Thinking, Fast and Slow, the two classic modes: the fast system, automatic, based on cognitive shortcuts, and the slow system, deliberative, costly in energy.
The adult brain runs in fast system the vast majority of the time. It decides what to read and what to skip, whom to trust and whom to doubt, not from reasoned analysis but from cheap rules: the source looks like the usual one, the headline reproduces already-established biases, whoever's signing seems likeable. Apply that filter to two hundred thousand words a day and you understand why the average human decision is the quality it is. It's not that people are stupid. It's that you can't make every decision with the slow system: it would come out expensive, it would come out late, and the organism would break.
Simplification, like forgetting, isn't a defect. It's an evolutionary solution to the problem of keeping the subject operational in an informationally hostile environment. The bad part is that the environment that designed the solution wasn't this one. The heuristic that served to tell friend from enemy in a village of two hundred people operates, without having been rewritten, in a timeline (the chronological line of posts a social network shows) where two hundred fragments go by in fifteen minutes. It errs at industrial rate, but the subject keeps deciding, because the alternative is not deciding, and not deciding costs too.
When both valves saturate at once
When both valves saturate, the third state appears, where a growing portion of the adult population lives: blockage.
The subject stops discarding well and stops simplifying well at the same time, and then there's no decision, there's paralysis, or absurd decision. The most cited study is Danziger, Levav and Avnaim-Pesso's, published in PNAS in 2011, on Israeli judges ruling on parole applications. The probability of a favorable ruling hovered around sixty-five percent at the start of the session and fell, over the course of the morning, to near zero just before the lunch break. After the break, it returned to sixty-five percent. The initial reading was that the decisive factor wasn't the file but the time of day and the distance to the last sandwich.
The study has been challenged. Weinshall-Margel and Shapard (2011), in the same journal, showed the case order wasn't random: prisoners without a lawyer and those from the same prison tended to cluster at the end of each block, just before the break. Glöckner (2016) added that favorable cases take longer to resolve than unfavorable ones, which produces, by pure statistical artifact, a pattern resembling the one observed without needing to invoke fatigue. The original authors re-ran their regressions controlling for those variables and the effect held, though reduced. The clean picture of the judge who denies releases because his glucose drops doesn't hold up as a sole example; what is documented in other contexts is decision fatigue (the progressive deterioration of decision quality as decisions accumulate over the day).
The office workday as a waterfall
Multiply that structure by a contemporary office workday. How many decisions between nine and one. How many back-to-back meetings. How many notifications in parallel. How many context switches.
Cal Newport has spent years repeating the figure, in Deep Work and after: the cognitive cost of each context switch isn't zero, and the sum turns the workday into a sequence of tasks all done halfway. What the subject believes is efficiency is, in cognitive terms, a waterfall of interrupted operations that never reach depth and never leave a consolidated trace. You remember almost nothing of an office workday because almost nothing gets processed. It gets executed on the fly and discarded.
Silence was the place where the brain consolidated
There's an additional detail, one of the ugliest. This saturation now needs something it didn't need before: to fill the gaps.
Silence was the place where the brain consolidated. Walking to work with no incoming stimulus, waiting in a queue without a screen, sitting on the sofa with nothing playing in the background: windows in which the system reorganized the day's material, trimmed the incidental, marked the important. The average twenty-first-century adult brain no longer has those windows. It has filled them all. Every minute without stimulus gets covered with a glance at the phone, a forwarded audio at double speed, a podcast while doing three other things.
An average urban adult's minutes of non-stimulus on a workday can be counted, if at all, on the fingers of one hand. The brain was designed to have several hours.
The phone that replaces silence
The phone hasn't replaced the computer, or the TV, or the book. It has replaced silence. The earlier screens occupied defined physical spaces; silence existed as a natural interruption between them. The phone occupies the interruptions too, and turns the workday into a single continuous session of input (information entering the cognitive system). As soon as the pressure drops, it rises again.
This explains much of the contemporary feeling of spending the day busy and having, at the end, not a single memory worth talking about. It's not early Alzheimer's. It's chronic saturation of a system that wasn't designed to operate without pauses.
AI promised to lighten and has added weight
And the last element appears, which is where the picture stops being merely bleak and turns comic.
The promise of generative AI (models that produce text, image or video from an instruction) repeated in every ad and every keynote (corporate product presentation), is to free up cognitive load. Let the models read for us, summarize, draft, decide, remember. The argument is elegant. The practice is the opposite.
Every tool that delivers cheap summaries lowers the cost of producing material, not of processing it. Producing a thousand-word text cost, five years ago, an hour of a human. Today it costs thirty seconds and a couple of cents. Since producing costs less, more gets produced. Since more gets produced, more reaches the receiver, who is still the same brain as before, with the same four working-memory seats. The gap was the problem, and the solution has filled it with more problem.
How many films you truly remember
It's worth posing a question with no alibi. How much of the information an adult consumes even pretends to be remembered. How many films, how many series, how many short videos, how many threads, how many podcasts, how many articles.
Run through your last week without cheating. Then do the hard exercise: how many of the films you've seen in your life could you describe today in enough detail for someone who hasn't seen them to understand the plot. The honest figure, for almost any reasonably film-loving person, is on the order of ten. Ten. Twenty at most. How many have you actually seen? Hundreds. Possibly thousands counting your childhood.
The consumption hasn't produced memory. It has produced the sensation of having seen, which isn't the same. The brain, faithful to its economy, has discarded almost everything and kept a handful of loose images you can no longer match to their film, a vague atmosphere from a few works, and the titles. The titles do stick, because they're useful for conversation and for fitting into the social network of cultural consumption. But they aren't memory. They're labels.
The folder isn't memory
The same happens with the books read halfway, the podcasts heard while driving, the articles saved in a folder never reopened, the screenshots piled by the thousand in a camera roll nobody's going to review. We archive in industrial quantities with the superstitious conviction that some of it, at some point, will become part of us. It won't. The brain already did its job in real time: it discarded it.
The folder saves nothing, because the pending operation was to process and consolidate, and that one can't be delegated. A saved note isn't a memory. A conversation with a model that summarizes a book for you isn't equivalent to having read the book: what mattered was the stretch of time when your brain had to chew it over slowly, and that stretch you haven't had.
The way out nobody sells
The picture is ugly and the way out isn't any of the ones the market sells.
It's not buying another note-taking app, or a personal knowledge-management system, or a productivity coach, or a model that promises to become your external memory. All those operations raise the incoming flow even more: each comes with its own load of learning, maintaining, reviewing and deciding.
The necessary operation, if it exists, is the opposite: to lower the flow. Reduce input. Tolerate silence, which is where the brain consolidates. Accept that remembering ten films instead of a thousand is what was planned, and that when you try to remember a thousand you end up remembering not a thousand but none. Most of the material that comes in isn't going to stay, and the sensible decision is not to let it in.
That clashes with the economic structure of almost everything we're sold. The attention economy is built to prevent it. Work organization is built to prevent it. Digital sociability is built to prevent it. That's why almost nobody does it, those who try do it halfway, and those who do it halfway end up where they started. Blockage and absurd decision are the default state of the contemporary urban adult's working afternoons, not the exception.
Who decides what you are
There's a dirty question worth posing without softening. If the system is discarding almost everything that comes in, and if what remains isn't what the subject chose but what their heuristic filtered, who decides what I am?
Memory builds the subject. Endel Tulving posed it when he separated episodic from semantic memory, the classic neuroscientists say it, common sense says it. If what my memory keeps isn't chosen by me but by an automatic valve triggered by saturation, my biographical continuity is being written by the filter.
The most likely answer, seeing what comes in and what stays, is that those of us who live in this environment are less and less the product of our decisions and more and more the product of the unnoticed discards of an exhausted system. And the unnoticed discards of an exhausted system produce lives that resemble one another more than their protagonists would admit.
Definiciones
Cognitive saturation. A state in which the flow of incoming information exceeds the capacity of the attentional system and working memory to process it, forcing the brain to discard or simplify massively and often unconsciously.
Working memory. A limited cognitive space, of about four simultaneous elements according to post-Miller research, where the information actively in use at a given instant is manipulated.
Heuristic. A fast mental rule that replaces reasoned analysis with a shortcut based on surface cues, enough to decide under pressure at the cost of high error rates.
Decision fatigue. The progressive deterioration of a subject's decision quality as decisions accumulate over the day, through exhaustion of the slow system's deliberative resources.
Infinite scroll. An interface mechanism that continuously loads new content without the user reaching the bottom of the page, designed to maximize time on the page.
Timeline. The chronological line of posts a social network shows a user, generally ordered or filtered by an algorithm.
Input. In this context, any information that enters the subject's cognitive system: text, sound, image, notification.
Generative AI. A family of artificial-intelligence models capable of producing text, image, audio or video from an instruction, rather than merely classifying or predicting.
Referencias
Bohn, R. & Short, J. (2009). How Much Information? 2009 Report on American Consumers. Global Information Industry Center, University of California, San Diego. Cited for the figure of a hundred thousand words a day consumed by an average adult.
Hilbert, M. & López, P. (2011). The World's Technological Capacity to Store, Communicate, and Compute Information. Science 332, 60–65. Reference for the global scale of information consumption and its rising curve.
Miller, G. A. (1956). The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information. Psychological Review 63, 81–97. Origin of the classic calculation of working-memory capacity.
Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences 24, 87–114. The downward revision of Miller's figure used in the article.
Schacter, D. L. (2001). The Seven Sins of Memory: How the Mind Forgets and Remembers. Houghton Mifflin. Basis of the argument on forgetting as adaptation and not as failure.
Luria, A. R. (1968). The Mind of a Mnemonist. The clinical case of Solomon Shereshevsky; used as an example of memory without discard as disability.
Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux. The framework of the fast and slow systems used in the section on simplification.
Danziger, S., Levav, J. & Avnaim-Pesso, L. (2011). Extraneous Factors in Judicial Decisions. PNAS 108, 6889–6892. The study on Israeli judges cited to illustrate decision fatigue under saturation.
Weinshall-Margel, K. & Shapard, J. (2011). Overlooked factors in the analysis of parole decisions. PNAS 108, E833. DOI: 10.1073/pnas.1110910108. A critique of the earlier study for the non-random order of cases.
Glöckner, A. (2016). The irrational hungry judge effect revisited: Simulations reveal that the magnitude of the effect is overestimated. Judgment and Decision Making 11, 601–610. A reanalysis of the Danziger effect by simulation; shows that part of the pattern is explained as a statistical artifact.
Newport, C. (2016). Deep Work: Rules for Focused Success in a Distracted World. Grand Central Publishing. Reference for the cognitive cost of constant context switching.
Tulving, E. (1972). Episodic and Semantic Memory. In Organization of Memory (Tulving & Donaldson, eds.), Academic Press. The classic distinction between episodic and semantic memory, evoked in the closing.
También te interesa
- The mind as limited space. Knowledge does take up room
- RAM memory vs human memory. Two different things with the same word
- States of the mind. No test measures the same person twice
No comments yet
No comments yet. Be the first.
Leave a comment