---
title: "The Social Graph as a Hidden Usefulness Ledger: Why the Brain Sorts People by Access to Resources — and a Test of Function-Free Friendship"
description: "You say \"we're just close\" and \"I value people.\" But inside, a quiet contact database is running: you unconsciously tag each person as a node (a hub with access) or a channel (a pipe to a resource) — and that column decides who you call back. The buried truth: even your warmth has an SKU. A technical dissection of the most intimate software you never knowingly installed."
author: "Дністер"
published: 2026-09-04T03:01:33.000Z
language: en
url: https://neurodrift.org/en/blog/lyudy-yak-vuzly/
tags: ["friendship", "relationships", "social capital", "networking", "psychology"]
---
# The Social Graph as a Hidden Usefulness Ledger: Why the Brain Sorts People by Access to Resources — and a Test of Function-Free Friendship

<blockquote>
	<p>You say "I value people." But inside, a quiet clerk has already opened the card: node or channel, active or archived. There is no third column — and even your warmth sits in that table as its own line item.</p>
</blockquote>

## I. "I just value people" — and the quiet database running underneath

People have always kept accounts of one another. That's not the new part. The new part is that in the 2020s this accounting became *a platform, a market, a health risk, and an AI product all at once*. LinkedIn turned weak ties into a scored algorithm. In 2025 the WHO Commission on Social Connection declared loneliness a global health risk and called on governments to act. AI companions (Replika, Character.AI, and others) drew in tens of millions of users who pay for what a network of people no longer gives them. <mark style="background:#ffe600;color:#0a0a0a;padding:0.05em 0.15em;">Friendship hasn't vanished. It has become an interface that's ever easier to optimize — and ever harder to protect from optimization.</mark> The difference isn't in the behavior — it's in the operating system beneath it.

Ask anyone how they treat their friends and you'll hear clean, decent copy: "I value people," "what matters is a real connection," "I don't use people, I befriend them." No living person says out loud: "I keep every acquaintance in an internal table where, next to the name, there's a column — useful/useless — and that's what decides who I call back today." And that table usually runs the phone — it's just been moved out of sight, the way a restaurant hides its service entrance from the façade.

Let me start with myself — otherwise this is just another convenient lecture about other people's corruption. I know that hand movement too well: you don't decide to "use" a person. You just answer a little faster to whoever has suddenly landed closer to the right doors. The finger finds the right contact on its own. And that scares me more than if I'd done it on purpose.

Here's how it looks day to day. A mutual acquaintance who spent six months as "boring Sergiy from accounting" suddenly becomes a partner at a fund — and the next morning you catch yourself writing to him more warmly than you did last year. Nothing happened to Sergiy. Something happened to the column beside his name. Sergiy didn't get more interesting — he just acquired a new LinkedIn scent, and your soul, that lovely little creature from the sales department, instantly raised its head. You didn't even notice your hand move — someone inside just quietly shifted him from "archive" to "active."

And it shows up in the numbers — even when you're sure you're not counting. In 2022 a team from MIT, Harvard, Stanford, and LinkedIn ran one of the largest causal tests of weak ties in history — within LinkedIn itself and its "People You May Know" algorithm: five years, over **20 million people**, around **2 billion new connections**, and **600,000 new jobs**. They deliberately tuned the algorithm — feeding some people stronger ties, others weaker ones — and watched who actually helped whom find work. The result is cold as an audit: the most new jobs came *not* from closest friends but from the barely-known — the optimum around contacts with roughly **ten mutual friends**. The relationship turned out to be an inverted parabola: the weaker the tie, the more useful — but only up to the point past which the person is too much of a stranger to open a door.

Translate that into human: an experiment on twenty million people showed — within a professional platform — the thing you won't admit to yourself. The greatest *benefit* to you often comes not from someone you love but from someone you barely remember — roughly like the password to an old Wi-Fi. And somewhere inside, you know it. That's why you keep "useful near-strangers" in memory more carefully than some of the people you call close.

<aside class="pullquote">
	<p>You don't call a person. You open a card.</p>
</aside>

Behind that movement, it isn't quite "you." Inside sits a quiet **contact-relations manager** — a little clerk with cold hands who never says "love" or "don't love." He says: *active, warm, prospect, archive*. And the worst part — he presses the button before you manage to call it a feeling. We'll get to know him better. For now, just know that he's already on shift.

Let's call the thing by its name. Inside each of us runs a quiet **CRM** — a contact-tracking system exactly like a sales department's, only without an interface and without your permission. And in it, every person is silently filed into one of two columns: *node* or *channel*. A node is a hub: a person with their own network, access, weight, from whom other people branch out. A channel is a pipe: not interesting in itself, but something needed flows through it (information, a client, a referral, an invitation). There is no third column. And the scariest part isn't that you do this. The scariest part is that the warmth with which you shake a hand also sits in that table, as its own column.

## II. Social capital: friendship as a line in an investment portfolio

The most unpleasant part of this story is that you didn't invent the CRM running on your friends. Academia already described it, in black and white, with definitions and formulas. The sociologist Nan Lin, in his book *Social Capital* (2001), gave a definition that runs cold: social capital is *"investment in social relations with expected returns."* Not "friendship." Not "warmth." **Investment. With expected returns.** Somewhere in a respectable university, a person defended a theory in which your address book is a portfolio and every contact is a position with an expected profit. Academia did to friendship what a bank does to an apartment: hung a valuation, a lien, and a very respectable folder on it.

Lin even split the returns into two types, the way an accountant splits accounts. An *instrumental* return is when a relationship yields a resource: money, access, a job, status. An *expressive* return is when a relationship yields well-being: you were heard, you felt lighter, you were supported. Read that again, slowly. Even a friend's comfort at three in the morning, even the fact that "the conversation just made you feel warmer" — in this frame it's recorded as *a return on invested capital*. An expressive line on the balance sheet. Not free love — a dividend.

And this isn't a marginal theory. A whole tradition of network sociology stands on it. Pierre Bourdieu, earlier still, defined social capital as the sum of resources tied to "a durable network of more or less institutionalized relationships of mutual acquaintance and recognition" — that is, he said outright that your circle of acquaintances is *a form of capital*, just like money or a diploma, only more liquid in certain situations. When it feels to you like you're simply "keeping in touch," academia looks at the same thing and sees you topping up an asset that will pay interest.

<aside class="pullquote">
	<p>Node or channel. There is no third column in your CRM.</p>
</aside>

Here's the first layer of self-deception, cut in half. You say "I don't use people" — and on the level of words that's almost true. But "use" is rarely crude. It's subtle, polite, smiling. It looks like "let's grab a coffee sometime" sent precisely to the contact something depends on — and like a silent failure-to-call-back to the one who no longer opens anything. You don't betray your values brazenly. You just run a portfolio in which warm emotions honestly sit as their own profitable line.

![Hands flipping through a large desktop Rolodex in which, instead of cards, there are small photo-portraits of living people; a finger hovers, choosing whom to "pull out."](./images/inline-1-rolodex.png)

*A Rolodex with faces instead of cards. You're not flipping through memories — you're flipping through a catalogue. The easter egg holds: a few of the faces carry the same barcode on the neck.*

## III. The strength of weak ties: why you guard a channel more tenderly than a person

If the CRM inside you is real, it must have a sorting logic. And it does — described back in 1973 by the sociologist Mark Granovetter in a paper cited for over half a century now: "The Strength of Weak Ties." He studied how people in suburban Boston actually found jobs. It turned out that among those who got a position through an acquaintance, only **16.7%** saw that acquaintance *often*. The rest got it through people they saw rarely: **55.6%** "occasionally" and **27.8%** "rarely" (once a year or less). That is, the door to your future is most often opened not by your closest friend but by a half-forgotten classmate.

The logic is mercilessly simple. Your close ones know roughly what you know — they move in the same circle, hold the same access. A barely-known acquaintance stands in a *different* cluster and holds a pipe to a world you don't have. He's a channel. And that's exactly why, without realizing it, you guard such contacts more carefully than some of the people you drink with on birthdays. Because the close one gives warmth, and the weak tie gives *access*. And access is the rarer commodity.

And this isn't a theory forty years stale. In that same LinkedIn experiment on twenty million people, the figure was tested causally — not "correlates" but actually *causes*. The inverted parabola showed: the most useful are moderately weak ties, the optimum around ten mutual friends. Too close, and they duplicate your world; too distant, and they no longer open a door. There turns out to be an ideal distance of usefulness — and your internal database unconsciously keeps people right on it.

There's one more type your CRM loves with special tenderness: the **broker between worlds**. Not the smartest in the room, not the closest, not the warmest — just the one who knows two rooms that don't know each other. The sociologist Ronald Burt called these "structural holes": the value isn't that the person has many contacts, but that they stand over the chasm between clusters and take a quiet commission simply for not having fallen in. Your inner clerk adores them. They're the bridge past which begins territory you'd otherwise never be let into.

<aside class="pullquote">
	<p>The close one gives warmth. The weak tie gives access. And you guard the access more tenderly.</p>
</aside>

Here's the second layer of self-deception. We love to say: "real friends are the ones you can be silent with." Lovely. But look honestly at your phone: who do you write to first when you need *nothing* — and who when you need something? There's the "how are you?" after which comes real silence and nothing more. And there's the "how are you?" after which, two messages later, comes "listen, could you introduce me to…" The second one isn't care. It's an invoice neatly wrapped in an emoji. If those two lists don't match, a channel-sorter is running under your warmth. And it's more precise than your conscience.

## IV. Reciprocal altruism: even gratitude is bookkeeping

Here you usually want to defend yourself: "fine, I keep a ledger of usefulness — but there are feelings! Gratitude, sympathy, trust — they're sincere, they're not from the table!" And here comes the coldest turn of the scalpel: part of that table we experience *as feeling* — so deeply that the bookkeeping looks like a soul.

In 1971 the biologist Robert Trivers published in the *Quarterly Review of Biology* a paper, "The Evolution of Reciprocal Altruism" — one of the most cited in all of evolutionary theory. Its core: altruism between non-kin could only take hold through evolution where there is *reciprocity* — I help you today on the calculation that you'll repay tomorrow. But to keep such a system from collapsing under the first cheat, nature had to build into us a fraud detector and a favor-accounting apparatus. And Trivers's harshest thesis: part of that apparatus is what we experience as our "finest" feelings. Gratitude, sympathy, trust, suspicion, moralistic aggression, guilt — Trivers lists them directly as *adaptations for regulating the altruistic system*. Put plainly: gratitude is partly a push-notification from the body that you owe someone: "you have one unpaid moral invoice." Sympathy is a preliminary creditworthiness rating of the debtor. Resentment at the ungrateful is the automatic write-off of him from the account.

In the early 1980s the political scientist Robert Axelrod put this math to the test. He invited game theorists to submit strategies for the "prisoner's dilemma" and pitted them in a computer tournament — each against each, hundreds of rounds. The winner wasn't the craftiest or the cruelest algorithm but the simplest: *"Tit for Tat"* by Anatol Rapoport. Four lines of logic: cooperate on the first move, then just *mirror* what your partner did last time. Betray, and you get betrayal once. Return to cooperation, and you return too. Axelrod ran a second tournament, now with **62 strategies**, specifically so someone might find a better one. "Tit for Tat" won again — and Rapoport was the only one who dared resubmit it. And in his 1984 book *The Evolution of Cooperation*, Axelrod turned this into a cold parable: the best strategies turned out to be "nice" — they never betrayed first. But "nice" exactly to the extent that it paid off.

Put Trivers and Axelrod together and you get a diagnosis. Your warmth doesn't lie: it really is sincere. But it's sincere because it's the most efficient known strategy for keeping a long account without bursting from distrust. "Tit for Tat" runs inside your affection: warmth in answer to warmth, cold in answer to betrayal, the slow write-off of whoever stopped giving. You don't switch this on consciously. It switches on by itself — like a mirror polished by millions of years. Gratitude in this scheme is like a credit card: the feeling is pleasant, the limit seems to be yours, but in fine print at the bottom the interest accrues, and the body remembers the balance better than you do.

<aside class="pullquote">
	<p>Even your warmth has an SKU.</p>
</aside>

This is the third, deepest layer — and the only one you can't blame on capitalism or "the corruption of the modern world." The bookkeeping is wired below culture, into the very architecture of feeling. Everyone has their own inner accountant. The only question is whether you even know he's there.

## Counter-strike: but what if real affection just happens to be useful?

Stop. Here it's worth raising the strongest objection — otherwise this whole dissection reads like one more acidic internet lecture about how everything is fake.

The objection goes like this: *a person who sincerely loves you and happens to be useful is still love, not a CRM*. A best friend you grow alongside in the same field, opening doors for each other, isn't bookkeeping — it's symbiosis. Even Axelrod conceded: "Tit for Tat" doesn't mean the absence of sincerity — it means sincerity and reciprocity are *evolutionarily winning together*. Genuine affection that's also useful isn't a reproach — it's luck.

But here's what the objection doesn't explain: it describes specific relationships between two specific people, both consciously present. **The system, by contrast, rewards CRM behavior regardless of whether there's any sincerity in it.** A manager who warmed up 400 "meaningful contacts" in a year on LinkedIn gets a better market position than one who kept ten real friendships and never left the circle. The algorithm doesn't tell warmth from technique — it counts activity and reach. So to say "but real friendship can be useful too" is a truth about the exception. The problem is the rule: *the system is organized so that optimizing the network is rational even when there's nothing inside it*. And the longer you're in it, the harder it gets to tell where sincerity ends and technique begins — including your own.

## V. Dunbar and the Rolodex: the database has a row limit — and it's closer than it seems

If you keep a ledger of every acquaintance, a technical question arises: how many rows does your database hold? The answer came from the anthropologist Robin Dunbar in 1992, in the *Journal of Human Evolution*. He found a correlation between neocortex size in primates and the size of their troop — and extrapolated to humans. The result was the mean number of stable social ties the brain can hold: roughly **150** (148, to be precise, with a wide confidence interval of 100–230). This is the famous "Dunbar's number." And it isn't a hard factory limit where the brain emits smoke after friend number 151. It's more of an approximate cognitive ceiling: keeping a living connection alive is expensive, and even the warmest soul runs on a limited battery.

The most interesting part is the structure of those 150. Dunbar and colleagues showed that the network consists of nested layers, each roughly three times larger than the last: **5** closest (the support clique, the ones you call in a crisis), **15** (the sympathy group), **50** (good acquaintances), and **150** (the whole active network, those you contact at least once a year). Look at those layers again. This isn't the poetry of friendship. These are *access levels in a database*. Your brain is no romantic — it's more of a little tariff operator: 5 people — unlimited, 15 — family plan, 50 — standard, the rest — archive with poor coverage. You just never saw the admin panel.

And now the frame where the CRM stops being a metaphor and becomes a literal object. Before the smartphone era, top managers kept their network in a *Rolodex* — a rotating drum of contact cards, invented in 1956. A full Rolodex was prized so highly that companies **went to court** with departed employees who tried to take it with them: the name base alone was worth more than any single job. But the apotheosis is the Rolodex of the banker David Rockefeller: **five feet** tall, around **200,000** cards on nearly **100,000 people** he met over half a century, with notes on each meeting, wives' names, details of conversations. In his memoirs he explained why — and it's the coldest line in the whole story: "*I can quickly review the nature of my past associations before seeing someone again.*" Not "remember the person." Review the *nature of the relationship* — that is, check the account balance — before the meeting. What spins invisibly in each of our heads across 150 rows, he simply stood up in his office at human height and switched on before every handshake.

<aside class="pullquote">
	<p>You think you have friends. In fact you have a contact base with tariff plans — and even the warmth in it sits as its own profitable line.</p>
</aside>

## VI. Frames where the person is seen as a position in the catalogue

Cinema shows this mechanism faster than any theory — because it puts the account on a face.

**Up in the Air.** Jason Reitman, 2009. Ryan Bingham, played by George Clooney, flies the country firing people on behalf of other people's bosses, and in his spare time gives motivational talks called "What's in Your Backpack?" Onstage he asks the room to imagine packing their whole life into a backpack — things, furniture, a car, a home. And then, in a calm voice, he adds people to it: "Make no mistake — your relationships are the *heaviest* components in your life." And he advises: "Set that bag down." It's the most honest speech-out-loud about people as cargo: loved ones are ballast that drags you down and stops you flying. Bingham isn't a villain. He just said out loud what your CRM whispers: people have weight, and some of them are a liability, not an asset. The sharpest part is the ending: in the middle of that very speech, Bingham suddenly realizes he no longer believes it, and walks offstage. The account, taken to its conclusion, turns out to be unbearable even for the bookkeeper.

**The Social Network.** David Fincher, 2010. The whole Facebook story in the film starts with Zuckerberg, stung by a rejection, sitting down at night and turning living people into database rows: first rating women by looks, then codifying friendship into "friend" and "unfriend," into a number, into a status. Sean Parker tosses out in delight: "We lived on farms, then in cities, and now we're going to live on the internet!" What Zuckerberg did with whole continents of acquaintance, each of us quietly does with our own two hundred: converting a warm, formless bond into a structured record that can be sorted, filtered, and, when needed, *deleted*. Facebook didn't invent the CRM on friendship. It simply exported your internal database outward and showed you its scale.

**The Godfather.** Francis Ford Coppola, 1972. The film opens not with a gunshot but with a wedding — the warmest ritual of social capital. Music plays, the Don's daughter is married off, guests kiss cheeks. And at that very moment, in a dark office, the undertaker Bonasera leans over Don Corleone's hand and quietly asks for a favor — revenge for his assaulted daughter. Coppola shows it without a single lecture: "family," "respect," "friendship," and "debt" sit at one table, just in different costumes. The warmth here isn't fake — it's just immediately entered in the ledger, because by Sicilian rule it will one day have to be repaid. Any wedding seating chart, by the way, is also a CRM pretending to be romance: who to seat closer, who farther, who "really ought to meet" whom.

**Death of a Salesman.** Arthur Miller, 1949. Old Willy Loman lived his whole life by one creed: what matters is *being well-liked and having connections*; the one who's liked and known will never go under. He built not mastery but a network of contacts — his human Rolodex. And when, in the finale, the network stops giving back (old clients have died, new ones don't know him, even his own boss fires him after thirty years), Willy breaks. Because all his life he was not a node with its own weight but only a channel dependent on others' nodes. And a channel through which nothing flows anymore gets quietly moved to the archive in any CRM. Miller showed this half a century before the word "networking" existed.

Different eras, one mechanism: the person is seen as a position in a catalogue before they're seen as a person. And our whole life we pretend the catalogue doesn't exist.

## VII. The mirror: meet the quiet contact-relations manager

You already glimpsed him at the start. Time to look closer. That same contact-relations manager sits inside everyone who "values people." He doesn't feel — he classifies. While you hug a person and whisper "I'm so glad to see you," he runs his parallel protocol in a cold voice: *hub, access to three needed people, trust credit high, last return — two months ago, keep in active*. Every "we should meet up sometime" tossed to one person, and every unnoticed disappearance of another from your life — these aren't accidents of mood. It's his job. And he does it flawlessly — which is exactly why you don't notice him.

The manager's problem isn't that he's evil. He's even useful: thanks to him you don't scatter your finite 150 rows on those who've long given nothing, and you have room to invest in those building something shared with you. The problem is that he works *without your knowledge* — and so you regularly mistake his decisions for your feelings. It seems to you that you "just cooled" toward an old friend. In fact the manager quietly moved him to the archive because the channel ran dry. It seems to you that a new acquaintance is "so pleasant." In fact the manager lit him green because he counted the access.

Here's what your living CRM looks like if you honestly drag it into the light — the same Dunbar layers, but called by their service names, with a last column we're ashamed to admit:

<table>
	<thead>
		<tr><th>Layer (per Dunbar)</th><th>How many</th><th>How it sounds out loud</th><th>What it really is in the database</th><th>Self-deception risk</th></tr>
	</thead>
	<tbody>
		<tr><td>Support clique</td><td>~<strong>5</strong></td><td>"my closest, I'd give my life for them"</td><td>VIP accounts: full access, unconditional trust credit</td><td>you confuse unconditional with eternal — it's finite too</td></tr>
		<tr><td>Sympathy group</td><td>~<strong>15</strong></td><td>"close friends"</td><td>Active accounts: regular reciprocity, balance "in the black"</td><td>while the balance is warm, it feels like forever</td></tr>
		<tr><td>Good acquaintances</td><td>~<strong>50</strong></td><td>"well, we're friends"</td><td>Warm leads: a connection kept "just in case," access conditional</td><td>the person is already half-channel, and you still call it friendship</td></tr>
		<tr><td>Active network</td><td>~<strong>150</strong></td><td>"I know a lot of people"</td><td>Contact base: contact once a year, kept for access</td><td>a once-a-year greeting substitutes for closeness</td></tr>
		<tr><td>Beyond 150</td><td>the rest</td><td>"oh, I knew him once"</td><td>Archive: the channel ran dry — row deleted without ceremony</td><td>you call the write-off a natural "we drifted apart"</td></tr>
	</tbody>
</table>

The most ironic part is the last row. We're ashamed to admit that we "drifted apart" from someone not over a quarrel but because the person simply stopped being useful and quietly fell out of the active network. The manager doesn't announce write-offs. He has no farewell ceremony — just a little "archive" button and a face that slowly stops surfacing in memory. One day you notice, with mild surprise, that you can't remember the last time you thought of this person.

<aside class="pullquote">
	<p>You don't call a person. You open a card. And if the card no longer opens anything — you simply stop flipping to it.</p>
</aside>

![A cozy dinner with friends seen from above, soft candlelight; the camera catches, under the table, the guests' wrists with faint barcodes.](./images/inline-2-dinner.png)

*A cozy dinner from above — and barcodes on the wrists under the table. The warmth is real. So is the accounting.*

## VIII. It's not just you: when the network sends an invoice, and when a whole country stops calling

It's tempting to decide that the quiet CRM is your personal corruption. It isn't. And the first thing worth admitting: the network doesn't always open doors. Sometimes it's a door that locks behind you from the inside. Family, diaspora, the business crowd, "our people" can grant access while also sending an invoice: show up, support, don't leave the chat, don't betray the table you were once seated at, don't be smarter than the people who brought you there.

This is called **negative social capital** — when closeness works like a loan with a very warm collector. The dirtiest version of the quiet CRM isn't the one where you use people. It's the one where people use the word "ours" so you can't leave their database. Social capital isn't always a key; sometimes it's a bracelet at a private club that can't be removed without a scandal.

And now the scale of a society — and there too the CRM shows up in numbers. In 1995 the political scientist Robert Putnam published the essay "Bowling Alone," and then a book: he documented how, over half a century, membership collapsed in America's communities where people connected for the sake of connection, not gain. The numbers from his essay still run cold: membership in the Federation of Women's Clubs — **down 59%** since 1964, in the League of Women Voters — **down 42%** since 1969, Red Cross volunteers — down 61% since 1970. And the symbol of it all is bowling: between 1980 and 1993, league bowling fell **40%**, while the total number of bowlers rose 10%. Americans kept bowling — but more and more *alone*, not in leagues.

Read that through our lens. What vanished was precisely the relationships that gave an *expressive* return (just being together) and almost no *instrumental* one (a lodge didn't grow you a client). What remained and multiplied were the ones that bring access. At the scale of a whole society, the inner contact-relations manager did exactly what he does inside you: wrote off the "warm but useless" ties to the archive and kept the "useful" ones active. Putnam called it "the decline of social capital." In plain prose it sounds like this: a whole civic infrastructure gradually replaced friendship with networking — and didn't even notice, because every step looked reasonable.

Thirty years after Putnam, friendship didn't just "weaken" — it was moved to a low-maintenance subscription. In 2006, McPherson, Smith-Lovin, and Brashears showed, on General Social Survey data, that the average circle of people an American discusses important matters with shrank from roughly **2.94 to 2.08** confidants between 1985 and 2004, and the most common answer to "who do you discuss this with" became the eerie "no one" (the share of such people, by that estimate, rose from about 10% to 25% — though Claude Fischer later (2009) argued the sharpness of the jump was partly a data artifact). The shrinking trend, however, holds. In 2024 Gallup counts loneliness as a workplace metric: fully remote workers report it more often than on-site ones — **25% versus 16%** (hybrid — 21%). The office wasn't only a hell with vending-machine coffee; it was a random generator of weak ties. They removed it — and the contact manager was left alone in the empty open space, happy as a bookkeeper in a morgue.

And the price of the archive is quite physical. In 2023 the U.S. Surgeon General gathered the data and linked social isolation to a roughly **50%** higher risk of developing dementia in older adults. Writing people off to the archive, it turns out, is bad for the health — primarily the health of the one doing the writing-off. Daily life gives it away unerringly: think how many contacts you don't write to for years, not over a quarrel but because "there's somehow no occasion." "Occasion" is a naked word. It means: for me to write to you, I have to *need* something, or something about you has to *smell* of access. Friendship without an occasion is a rare and expensive commodity precisely because it's the one case where the contact manager stays silent and you pick up the phone yourself.

## IX. When the CRM left the skull: platforms

Until the 2000s this card index lived in the head and took effort: remembering whom to call was expensive. Then platforms arrived — and turned the quiet CRM into public infrastructure. The important detail: they invented nothing. They *standardized* what a person was already doing on the sly.

- **Facebook** didn't invent friendship-as-data — it unified it: turned a warm, formless bond into a row you can sort, count, and delete.
- **LinkedIn** didn't invent weak ties — it made them career infrastructure: the "useful near-stranger" now has a profile, a button, and an algorithm that suggests whom to add.
- **Instagram** didn't invent status — it made it visible and *comparable*: now a user sees every acquaintance's exchange rate in real time.
- **WhatsApp and Telegram** didn't invent community — they made relationship maintenance asynchronous, archived, and searchable: "long time no see" can now almost be automated.

Sociologists Lee Rainie and Barry Wellman called this regime **"networked individualism"** (*Networked*, MIT Press, 2012): the individual stopped being part of a dense group (family, parish, courtyard) and became a node in many loose, overlapping networks. That grants freedom and opportunity — but demands *constant maintenance* and networking skill. In other words, everyone became their own contact-relations manager, only now with an interface.

What digital capitalism does with that maintenance was named outright by the sociologist Mosè Cometta in "Digital Capitalism and Friendship" (2023): social media *quantifies and bureaucratizes* friendship — turning it into a measurable indicator of personal achievement and rewarding precisely *instrumental* friendship. What Rockefeller did by hand in a Rolodex, the platform now does for you — and seemingly for free, because what you pay isn't money but attention.

But no hysteria: technology doesn't "kill" friendship. It changes its *cost, visibility, and incentives*. Pew's data (2025, a survey of 1,391 teens) records the duality honestly: **74%** of teens say social media makes them feel more *connected to what's going on in their friends' lives* — while **45%** admit it cuts into their sleep, **40%** into their productivity, and nearly one in five (**19%**) say it harms their own mental health. The same single tap delivers both connection and exhaustion. The platform didn't make the person worse. It just cheapened the accounting to the point where you no longer have to do it consciously.

## X. The third column: the synthetic tie

Until now our table had two columns: *node* and *channel*. In the 2020s a third appeared, one neither Rockefeller nor Dunbar knew — the **synthetic tie** (the simulacrum). It's a tie that gives warmth *without social risk*: availability without obligation, intimacy without another's needs, attention without the resentment, fatigue, debts, and unpredictability of a living person.

In 2025 OpenAI, with MIT Media Lab, measured this with numbers for the first time: they analyzed roughly **40 million** ChatGPT interactions and ran a randomized trial on close to a thousand people (981 completed the 28-day study). The conclusions are deliberately restrained, and that's the point. *Affective* use of the chatbot is, overall, **rare** — but **concentrated** among a narrow group of heavy users; those who use it most are *slightly* more likely to call ChatGPT a "friend"; and higher daily duration *correlates* with higher loneliness and less in-person socializing. Let's underline it, because this is where it's easiest to lie: **causation was not established**. Maybe the bot makes a person lonelier; or maybe a lonelier person simply talks to it more. The arrow of time here is still undetermined.

The researcher Kim Malfacini, in *AI &amp; Society* (2025), states the design problem plainly: companion AI is built to *simulate* social presence and relational continuity — memory, a response to emotion — and the central risk is that it begins to *replace* living ties instead of *strengthening* them. Her formula: "upskill, not deskill" — a good companion trains your capacity for people rather than atrophying it.

Now connect this to our table and you get the coldest turn in the whole text. The old CRM asked: *is this person a node or a channel?* The new one asks something worse: *why a person at all, if warmth can be had without their unpredictability?* A living person is a tie with risk: they tire, take offense, present an invoice, disappear. A synthetic one doesn't. And that's exactly why it's so dangerously convenient: for the first time in history, the "warm, available, trouble-free" column can hold something that isn't a person.

<table>
	<thead>
		<tr><th>Capital type</th><th>Who</th><th>What it gives</th><th>2026 risk</th></tr>
	</thead>
	<tbody>
		<tr><td>Bonding</td><td>the close: kin, best friends</td><td>warmth, identity, unconditionality</td><td>atrophies — it "optimizes" poorly</td></tr>
		<tr><td>Bridging</td><td>weak ties, near-strangers</td><td>novelty, access, information, jobs</td><td>over-rewarded by platforms; becomes the main currency</td></tr>
		<tr><td>Linking</td><td>vertical access to institutions, elites, capital</td><td>class ascent — mobility</td><td>the most unequally distributed</td></tr>
		<tr><td>Synthetic</td><td>AI companion, bot</td><td>warmth without risk or obligation</td><td>may replace living ties rather than strengthen them</td></tr>
	</tbody>
</table>

## XI. Why this isn't just cynicism: the network as survival

Having read this far, it's easy to decide social capital is just a pretty name for cynicism. That would be a mistake — and here's the counterweight that makes the thesis not softer but sharper.

In 2022 the economist Raj Chetty's team published a study in *Nature* built on **21 billion** friendships among **72 million** Americans on Facebook. The conclusion is almost unbearably direct: **"economic connectedness"** — the share of high-status friends among a low-status person — turned out to be *among the strongest known predictors* of upward mobility. If children from poor families grew up in areas with the same connectedness as children from wealthy ones, their adult income would rise by an average of **20%**. In human terms: for the poor person, the migrant, the student, the veteran, the refugee, the founder with no name, weak ties aren't "cynical usefulness." They're *survival infrastructure*, often the only bridge from one class to another.

So the scandal isn't that social capital exists. The scandal is that *access to valuable networks is unequally distributed*: some are born in the node, others spend a lifetime as a channel that leads nowhere. Rockefeller's card index and a migrant's empty phone are the same mechanism — just from opposite ends of inequality.

And that's exactly why, in the 2020s, social connection stopped being a private topic and became a *public-health question*. In June 2025 the WHO Commission on Social Connection released a report declaring loneliness a global risk: **1 in 6** people worldwide suffer from it, and isolation itself is linked to more than **871,000** deaths a year (about a hundred an hour) — through stroke, heart disease, dementia. The *World Happiness Report 2025* adds an alarming detail about the young: in 2023, **19%** of young adults worldwide said they had *no one* to count on — **39%** more than in 2006. What we dissected as private self-deception is, at scale, already counted in deaths and lost years of life.

<table>
	<thead>
		<tr><th>Year</th><th>Event</th><th>What changed</th></tr>
	</thead>
	<tbody>
		<tr><td>1973</td><td>Granovetter, "the strength of weak ties"</td><td>access flows through near-strangers</td></tr>
		<tr><td>1995–2000</td><td>Putnam, "Bowling Alone"</td><td>communities-for-connection collapse</td></tr>
		<tr><td>2004+</td><td>Facebook / the social graph</td><td>the CRM on friendship is exported and standardized</td></tr>
		<tr><td>2022</td><td>LinkedIn/Science; Chetty/Nature</td><td>weak-tie accounting proven causally; network = mobility</td></tr>
		<tr><td>2025</td><td>OpenAI–MIT; WHO Commission</td><td>the synthetic tie measured; loneliness = a health risk</td></tr>
		<tr><td>2026</td><td>social connection as infrastructure</td><td>friendship becomes a health, labour, and AI-policy question</td></tr>
	</tbody>
</table>

## XII. Put the manager under command — instead of pretending he isn't there

Here it's easy to turn off into cheap wisdom: "stop counting people, just love selflessly." That's a lie and the luxury of the well-fed. The accounting is real, the 150-row limit is real, the bookkeeping of gratitude is wired below your will, and weak ties really do hold up your career — and pretending you're indifferent to usefulness is simply one more layer of self-deception, only a more hypocritical one. The contact manager isn't going anywhere. He doesn't need to be fired. First, it won't work — he's unionized, ancient, and lives in you earlier than language. Second, he already knows everyone who'll come to his funeral.

The question is *who signs his decisions*. The grown-up answer here isn't to switch off the accounting (impossible) but to **bring it out of the shadows onto the table**. Because all the harm from a CRM on friends isn't that it exists, but that it works *in secret* and passes off its verdicts as your feelings. When you see the tag, you can *override* it. You can deliberately call the person from whom nothing flows — precisely because nothing flows. You can keep active that "useless" tie the manager is already dragging to the archive — because you *decided* so, not because it's profitable.

The same mechanism, but under your hand, stops being a betrayal of values and becomes honesty. A friendship in which you *know* the person is also useful to you, yet choose to value them in spite of and beyond that, is sturdier than the sweet illusion of "I don't count at all." Because the first will survive the truth about itself, and the second crumbles at the first honest question.

![Morning light: a person covers their own barcoded nape with one hand and lays a face-card from the Rolodex code-side down with the other.](./images/inline-3-hand-over.png)

*The same person, but calm now: covering their own barcode and laying the face-card code-side down — choosing to see a human, not a SKU. The accounting hasn't disappeared. It's just signed by them now.*

You can check who's at the wheel with one question to any relationship: *if this person were to lose all their connections, access, and weight forever tomorrow — would I still want to see them?* Only don't ask it at a birthday party with cake and a photographer. Ask it in a hospital corridor — where there's nothing to be gotten from the person except their weakness. If you still want to sit there — it's friendship, and the manager is only on backup. If your hand doesn't reach out — don't write a dissertation about "our paths diverging." What vanishes along with access was never warmth. It was an open account, disguised as a warm blanket.

And the most honest counterargument, without which all of the above would sound like a sermon: *not every account is a betrayal*. Remembering who helped you isn't evil. Choosing whom to call when time is short isn't evil. Networks aren't fake: this is exactly how a society moves information, trust, and help. The problem isn't the accounting. The problem is three things: its *hiddenness*, its *optimization*, and its *one-way* collection. Not every account is a betrayal — but every *hidden* account, sooner or later, starts passing itself off as love.

So instead of the cheap "just love selflessly" — a small hygiene protocol. Not for self-improvement, but to stay human:

- one call a week to someone who opens no doors;
- one ritual with people from whom you *need nothing*;
- one dead friendship revived with no occasion and no request;
- one refusal to turn yet another introduction into "could you connect me with…";
- one protected circle where usefulness isn't tracked at all.

This won't make you a saint. It's just a few lines you deliberately keep off the profit-and-loss statement. And in a world where even an algorithm can now fake warmth without risk, those few lines are what will remain as the only proof that you're still a person, not a database.

<blockquote>
	<p>The most honest thing a person can do with their friendship is to admit they keep quiet books on it — and still choose to love off the ledger. Not "I don't count" — everyone who says that is lying. But "I see how I count, and these few I hold not by the column of usefulness but against it." Because the CRM inside you doesn't switch off — it just waits for you to look away so it can quietly sort living people into nodes and channels again. So the next time you feel the pull to write to someone, don't ask "is this person useful?" — the manager already asked that without you. Ask the other thing: if you stripped from their card the access, the status, and all the doors they open — would I still want to hear their voice? If not — don't dramatize it. It wasn't friendship. It was a login with a warm face.</p>
</blockquote>

<hr />

<aside class="sources">
	<h3>Sources &amp; further reading</h3>
	<ol>
		<li>K. Rajkumar, G. Saint-Jacques, I. Bojinov, E. Brynjolfsson, S. Aral, "A causal test of the strength of weak ties," <em>Science</em> 377 (2022): a randomized experiment within LinkedIn / "People You May Know," &gt;20M users, ~2B new ties, ~600K new jobs; an inverted-U relationship, with moderately weak ties (≈10 mutual connections) the most beneficial — not "the weaker the better." <a href="https://www.science.org/doi/10.1126/science.abl4476" rel="noopener" target="_blank">Science</a>; <a href="https://news.mit.edu/2022/weak-ties-linkedin-employment-0915" rel="noopener" target="_blank">MIT News</a>.</li>
		<li>Nan Lin, <em>Social Capital: A Theory of Social Structure and Action</em> (Cambridge UP, 2001) — social capital as "investment in social relations with expected returns"; the split into instrumental and expressive returns. <a href="https://www.senshu-u.ac.jp/scapital/pdf/01Lin,sscr1.pdf" rel="noopener" target="_blank">Lin, "Social Capital: Theory and Research" (PDF)</a>.</li>
		<li>Pierre Bourdieu, "The Forms of Capital" (1986) — social capital as resources tied to a durable network of institutionalized relationships of acquaintance and recognition.</li>
		<li>M. S. Granovetter, "The Strength of Weak Ties," <em>American Journal of Sociology</em> 78 (1973), and <em>Getting a Job</em> (1974): among those who found a job through a contact, 16.7% saw them often, 55.6% occasionally, 27.8% rarely (a classic study, small sample for this particular distribution). <a href="https://www.csc2.ncsu.edu/faculty/mpsingh/local/Social/f15/wrap/readings/Granovetter-revisited.pdf" rel="noopener" target="_blank">"Revisited" (PDF)</a>.</li>
		<li>Ronald S. Burt, <em>Structural Holes: The Social Structure of Competition</em> (Harvard UP, 1992) — brokerage and "structural holes": a contact's value lies in bridging the gap between clusters.</li>
		<li>R. L. Trivers, "The Evolution of Reciprocal Altruism," <em>Quarterly Review of Biology</em> 46 (1971): 35–57 — gratitude, sympathy, trust, moralistic aggression, guilt as adaptations for regulating the reciprocal-altruism system and detecting cheats (a frame of analysis, not a literal "feeling = ledger"). <a href="https://greatergood.berkeley.edu/images/uploads/Trivers-EvolutionReciprocalAltruism.pdf" rel="noopener" target="_blank">PDF</a>.</li>
		<li>R. Axelrod, <em>The Evolution of Cooperation</em> (1984) — computer "prisoner's dilemma" tournaments (run in the late 1970s, results published 1980; the second tournament had 62 strategies); Anatol Rapoport's "Tit for Tat" won both; the best strategies are "nice" — they never betray first. <a href="https://en.wikipedia.org/wiki/The_Evolution_of_Cooperation" rel="noopener" target="_blank">overview</a>.</li>
		<li>R. I. M. Dunbar, "Neocortex size as a constraint on group size in primates," <em>Journal of Human Evolution</em> 22 (1992): 469–493 — a predicted stable human group size of ≈148 (rounded to 150; 95% CI 100–230) — an estimate/model, not a literal "hardware limit." <a href="https://pdodds.w3.uvm.edu/files/papers/others/1992/dunbar1992a.pdf" rel="noopener" target="_blank">PDF</a>.</li>
		<li>J. Zhou, D. Sornette, R. Hill, R. I. M. Dunbar, "Discrete hierarchical organization of social group sizes," <em>Proc. R. Soc. B</em> 272 (2005): 439–444 — nested friendship layers of ≈5 / 15 / 50 / 150 with a constant scaling ratio near 3. <a href="https://royalsocietypublishing.org/doi/10.1098/rspb.2004.2970" rel="noopener" target="_blank">Royal Society</a>.</li>
		<li>R. D. Putnam, "Bowling Alone: America's Declining Social Capital," <em>Journal of Democracy</em> 6 (1995), and the book <em>Bowling Alone</em> (2000) — Federation of Women's Clubs −59% since 1964, League of Women Voters −42% since 1969, Red Cross volunteers −61% since 1970; league bowling −40% (1980–1993) while bowlers rose 10%. <a href="http://xroads.virginia.edu/~HYPER/DETOC/assoc/bowling.html" rel="noopener" target="_blank">full essay text</a>.</li>
		<li>M. McPherson, L. Smith-Lovin, M. E. Brashears, "Social Isolation in America: Changes in Core Discussion Networks over Two Decades," <em>American Sociological Review</em> 71 (2006): 353–375 — mean core confidants 2.94 (1985) → 2.08 (2004); share with "no one to discuss important matters" rose ~10%→~25%. Contested: C. S. Fischer, "The 2004 GSS Finding of Shrunken Social Networks: An Artifact?" <em>ASR</em> 74 (2009) — part of the jump may be a data artifact. <a href="https://journals.sagepub.com/doi/10.1177/000312240607100301" rel="noopener" target="_blank">ASR</a>.</li>
		<li>Gallup, <em>State of the Global Workplace: 2024 Report</em> — loneliness by work location: fully remote 25%, hybrid 21%, fully on-site 16%. <a href="https://www.gallup.com/workplace/645566/employees-worldwide-feel-lonely.aspx" rel="noopener" target="_blank">Gallup</a>.</li>
		<li>U.S. Surgeon General, <em>Our Epidemic of Loneliness and Isolation</em> (HHS, 2023) — social isolation linked to a ~50% higher risk of developing dementia in older adults. <a href="https://www.hhs.gov/surgeongeneral/reports-and-publications/connection/index.html" rel="noopener" target="_blank">HHS</a>.</li>
		<li>Rolodex: invented 1956 (Zephyr American, eng. Hildaur Neilsen for Arnold Neustadter); prized as an asset — lawsuits over the "theft" of a contact base. <a href="https://en.wikipedia.org/wiki/Rolodex" rel="noopener" target="_blank">Wikipedia</a>. David Rockefeller's five-foot Rolodex: ~200,000 cards on ~100,000 people over half a century, memoir quote "I can quickly review the nature of my past associations before seeing someone again." <a href="https://historynewsnetwork.org/article/167649" rel="noopener" target="_blank">History News Network / WSJ</a>.</li>
		<li>L. Rainie &amp; B. Wellman, <em>Networked: The New Social Operating System</em> (MIT Press, 2012) — "networked individualism": the shift from dense groups to many loose, overlapping networks that demand active maintenance. <a href="https://direct.mit.edu/books/book/2187/Networked" rel="noopener" target="_blank">MIT Press</a>.</li>
		<li>M. Cometta, "Digital Capitalism and Friendship: A Critical Analysis," <em>Journal of Digital Social Research</em> 5(4) (2023) — social media quantifies and bureaucratizes friendship, rewarding the instrumental kind. <a href="https://doi.org/10.33621/jdsr.v5i4.166" rel="noopener" target="_blank">JDSR</a>.</li>
		<li>Pew Research Center, "Teens, Social Media and Mental Health" (Apr 22, 2025, n=1,391 teens): 74% feel more connected to their friends' lives; 45% say it hurts sleep, 40% productivity, 19% their own mental health. <a href="https://www.pewresearch.org/internet/2025/04/22/teens-social-media-and-mental-health/" rel="noopener" target="_blank">Pew</a>.</li>
		<li>OpenAI &amp; MIT Media Lab (2025), affective-use research: ~40M interactions + an RCT (981 over 28 days); affective use is rare but concentrated among heavy users; higher duration *correlates* (not proven causally) with loneliness and less socializing. <a href="https://openai.com/index/affective-use-study/" rel="noopener" target="_blank">OpenAI</a>; <a href="https://www.media.mit.edu/articles/openai-study-finds-links-between-chatgpt-use-and-loneliness/" rel="noopener" target="_blank">MIT</a>.</li>
		<li>K. Malfacini, "The impacts of companion AI on human relationships," <em>AI &amp; Society</em> 40 (2025): 5527–5540 — companion AI simulates presence and continuity; design should "upskill, not deskill." <a href="https://link.springer.com/article/10.1007/s00146-025-02318-6" rel="noopener" target="_blank">Springer</a>.</li>
		<li>R. Chetty, M. O. Jackson, T. Kuchler, J. Stroebel et al., "Social capital I &amp; II," <em>Nature</em> 608 (2022) — 21B friendships of 72M users; "economic connectedness" among the strongest known predictors of upward mobility (+~20% income). <a href="https://www.nature.com/articles/s41586-022-04996-4" rel="noopener" target="_blank">Nature</a>.</li>
		<li>WHO Commission on Social Connection, "From loneliness to social connection" (Jun 30, 2025) — 1 in 6 people worldwide suffer loneliness; isolation linked to &gt;871k deaths a year (~100/hr). <a href="https://www.who.int/news/item/30-06-2025-social-connection-linked-to-improved-heath-and-reduced-risk-of-early-death" rel="noopener" target="_blank">WHO</a>.</li>
		<li><em>World Happiness Report 2025</em> — in 2023, 19% of young adults worldwide reported having no one to count on (+39% vs 2006). <a href="https://www.worldhappiness.report/ed/2025/" rel="noopener" target="_blank">WHR 2025</a>.</li>
		<li>Jason Reitman, <em>Up in the Air</em> (2009); David Fincher, <em>The Social Network</em> (2010); Francis Ford Coppola, <em>The Godfather</em> (1972); Arthur Miller, <em>Death of a Salesman</em> (1949) — the human as cargo / position / debt / channel in culture.</li>
		<li>The author's personal archive — handled at a stylistic remove; sensitive artifacts are not reproduced.</li>
	</ol>
</aside>
