Four women have coffee at the same place on the second Tuesday of the month. They have been doing it since 1974, when their children were in the same nursery class.
Nobody organizes it. Nobody confirms. If one of them doesn’t turn up, somebody calls the house.
Somewhere else there is a group chat named after a joke from a bachelorette party in 2019. It has four thousand messages in it. It has produced one dinner, which was cancelled.
It’s easy to assume that one generation was better at friendship and the other one is soft and distracted and on its phone.
The decline itself is real enough to measure. In 1990, 3% of Americans said they had no close friends at all.
By 2021, that share had quadrupled to 12%, and the group reporting ten or more close friends fell from a third of the country to 13%.
But the four women weren’t working harder at it. They were standing inside a set of conditions that made loyalty almost free. A street where everyone was outside anyway. A phone that cost money to use, so the calls were short and real. A fixed night that arrived whether anyone felt like it or not, and a set of favors that had to be asked for out loud because there was no other way to get to the airport.
Nearly all of those conditions have been removed. What’s left is intention, which turns out to be a much weaker adhesive than a shared driveway ever was.
The group chat is what people are now using to do work the driveway used to do for nothing.
1. They saw each other constantly without anyone planning it
The friendship didn’t start with a decision. It started with a shared fence, a bus stop, a shift, a pew, four houses on the same cul-de-sac.
You saw the person forty times before you had a single conversation worth remembering. By the time you liked them, you already knew what their car sounded like pulling in.
This is the oldest finding in the field. When Leon Festinger and his colleagues tracked friendship formation in postwar student housing at MIT, the strongest predictor of who became close turned out to be the physical arrangement of the buildings. Shared background and interests barely registered next to it.
Residents whose front doors faced the stairwell knew everybody. Residents a short walk apart stayed strangers for years.
Nobody living through it knew that was what was happening to them. They thought they had found their people. They had largely been assigned them, by a floor plan and a shift rota.
A group chat has no floor plan. Every person in it was chosen deliberately by somebody, and then kept there by ongoing effort from the same one or two people, forever.
2. The friendship lived at a place, not on a calendar
There was a bar, a diner, a bowling alley, a church hall, a barbershop, a union office, a spot in the park where the same six men played cards.
You didn’t arrange to see anyone. You went to the place, and the people were at the place, and if they weren’t you went home and went back Thursday.
A friendship with an address behaves completely differently from one that exists only as a plan. A plan requires someone to propose it, someone to accept it, and both people to still want it on the day.
Those addresses have been closing for two decades, and the closures compound. The coffee shop goes, then the bar it fed, then the reason to be on that street at all.
When the place goes, the friendship converts into a scheduling problem. Scheduling problems have a way of never quite resolving.
More Bolde Stories
3. Showing up unannounced was normal rather than an imposition
Somebody knocked at seven on a Wednesday with no warning and no reason. They came in, had a coffee, complained about a supervisor, left after forty minutes.
The house was not clean. Nobody apologized for the house.
This mattered more than it sounds, because it made contact almost free to initiate. There was no ask, no proposed date, no risk of the other person saying they were busy that week and then never proposing another one.
That practice has all but disappeared. In the mid-to-late 1970s, the average American entertained friends at home about fifteen times a year.
By the late 1990s, that figure had fallen to eight, a drop of roughly 45% in two decades, and it has kept falling since.
The modern equivalent of the pop-in is a message in the group chat. It costs nothing to send, which sounds like the same trade, except that a knock at the door produced a person in a kitchen and a message produces a thumbs up.
4. There was no cheap way to stay in touch, so the contact had to be real
Long distance was billed by the minute and rationed to Sunday evenings. You called, you talked, you got off the phone, and the whole thing had a beginning and an end.
There was no way to maintain a friendship at low effort. You either spoke to the person or you didn’t, and if you didn’t for two years, everybody understood what that meant.
Texting removed that fork in the road. It’s now possible to stay technically in contact with someone for a decade without ever hearing their voice. That feels like maintenance right up to the moment you need something from them.
People know this and avoid the call anyway, because they expect it to be awkward. Amit Kumar and Nicholas Epley had participants reconnect with an old friend by phone or by email and found the callers formed a stronger bond.
They also didn’t feel any more awkward than the typers did, and the call took about as long as writing and reading the email.
So the awkwardness people are bracing against is mostly imaginary. The cost of bracing against it is a friendship that never gets past logistics.
5. They belonged to something that met on a fixed night whether they felt like it or not
The league bowled Thursdays. The union met the first Monday. The choir rehearsed Wednesday, the Elks met monthly, the PTA had a calendar, and the calendar did not consult anybody’s energy levels.
Friendship was the byproduct. Nobody joined the league for companionship, and companionship is what the league quietly manufactured, which is roughly how it works for adults.
The crucial part was that attendance had stopped being a decision. Nobody polled the group about whether Thursday still worked for everyone. Thursday was Thursday, and the people who kept showing up became the people you knew.
Robert Putnam spent a career documenting what happened to that scaffolding, and the short version is that most of it came down inside one generation. The bowling alleys stayed busy. The leagues emptied.
What replaced the standing Thursday is a message asking whether anybody’s free next week. That puts the entire burden on whoever is willing to be the person who always asks, and there’s a limit to how long anyone stays willing.
It’s worth saying that the obligation cut both ways, and younger generations have good reasons for refusing some of it. A standing commitment you’ve outgrown is its own kind of trap.
More Bolde Stories
6. Boredom kept pushing them out of the house
Three channels. No delivery. A television that was small and in the wrong room, and a house that was cold in the parts nobody heated.
Staying in was a bad option, so people left. Leaving is most of what friendship requires.
The house has since been re-engineered to win that argument every time. Homes are bigger and more comfortable, the screens are enormous, essentially every restaurant is on a delivery app, and remote work has kept a large share of the population indoors on weekdays.
It shows up cleanly in the time diaries. Average time spent socializing has fallen from 45 minutes a day to 35 over the past twenty years, and among 15 to 24 year olds it has dropped from an hour to 35 minutes.
Thirty-five minutes is one coffee, once, if you don’t count the walk. Spread that across everyone you’d call a friend and it gets bleak fast.
7. They accumulated hours that nobody was counting
Forty minutes in a driveway. An hour on a porch. Two hours in a car going somewhere neither of them cared about. A whole Saturday helping with a deck that ended up taking three Saturdays.
Almost none of it would qualify as quality time. Volume was the active ingredient, and volume is what the whole thing runs on.
Jeffrey Hall at the University of Kansas tried to put a number on it, tracking people forming new friendships after a move or a school transition. He estimated it takes somewhere between 40 and 60 hours to become casual friends.
It takes 80 to 100 to become friends, and more than 200 to become close ones. Hours spent working alongside each other counted for noticeably less than hours spent hanging around.
Two hundred hours is five full working weeks. Nobody schedules that. It only accumulates when two people are repeatedly in the same place with nothing much to do, which describes 1972 and does not describe now.
A group chat generates almost none of it. You can be in one for six years and still be at forty hours.
8. They asked each other for help that actually cost something
Watch the kids Saturday. Drive me to the garage. Cover my shift. Come help me move a piano, which is a terrible thing to ask of anybody.
The asking wasn’t a sign the friendship was strong. The asking is what made it strong, because a person who has carried your furniture is invested in you in a way that a person who has liked your photo is not.
Modern life has removed most of the occasions. There are apps for the ride, the move, the childcare and the meal, and paying a stranger is easier than the small exposure of asking somebody you like.
That exposure is also based on a misread. Across six experiments, the psychologists Xuan Zhao and Nicholas Epley found that people needing help consistently underestimated how willing others were to give it, overestimated how put out the helper would feel, and underestimated how good helping would make that person feel.
Which means the favor you’re currently not asking for is, on average, one the other person would have been pleased to be asked for.
The group chat is full of people subtly declining to give each other that.
More Bolde Stories
9. A falling out had to be survived because there was nowhere else to go
Two men fell out over money in 1979. They were in the same league, the same parish, the same four streets, with the same eleven mutual friends.
There was no exit available. So they were cold for eight months, then civil, then fine, and by 1985 the falling out had become a story they told at parties.
A fifty-year friendship contains several stretches where one person was furious and no mechanism existed for making the other one disappear. Endurance was structural, not admirable.
Now the mechanism exists. Mute, drift, leave the chat, never say a word about it, which is why so many friendships end without either party being told they ended.
The decay is quick once contact stops. Robin Dunbar’s review of the research notes that failing to meet up often enough causes emotional closeness to decline measurably within a few months, though it takes roughly three years of no contact for someone to drop out of your network altogether.
A fair objection, and it comes from the same review. Dunbar reports the same layered patterns of contact whether people are meeting face to face, calling, texting or posting, so the group chat does real work. The problem is the size of the load it inherited.
What the group chat is actually being asked to do
It’s the last piece of infrastructure left standing.
The bowling league is gone, the coffee shop closed, the pop-in became rude, the favors got outsourced, and the fixed Thursday dissolved into a rolling negotiation about everyone’s availability. Every one of those was doing a job.
The chat inherited all of them at once. So the joke about not being able to keep one alive is aimed at the wrong target, because a system that used to have nine or ten moving parts now has one, and that one is a text thread.
None of which makes any of this optional. The Harvard Study of Adult Development has followed people since 1938, and its summary of eighty years of data is that how satisfied people were with their relationships at fifty predicted their physical health at eighty better than their cholesterol did.
That study has fair critics. The original cohort was small, male and overwhelmingly white, and the popular version of it gets flattened a lot. The direction of the finding has held up regardless.
The four women at the diner almost certainly did not try harder than anybody reading this. They had a second Tuesday, and the second Tuesday did the work for fifty years without anyone ever having to feel motivated.
Which is the only real instruction in any of this, and it has nothing to do with wanting it more. One recurring thing, at a fixed time, in an actual place, that arrives whether anybody feels like it or not.
Then stop asking the group chat to be a friendship. Let it be what it’s good at, which is telling four people where to meet.
