By Thursday, you realize you have barely spoken to anyone all week.

Your meetings happened through a screen. You ate lunch over your laptop. The only person you saw in real life was the delivery driver, and even that exchange lasted less than a minute.

Then Saturday arrives. A friend invites you to dinner. You think about getting dressed, crossing the city, and sustaining conversation for three hours. You cancel because you don’t have the social battery.

This is the loneliness paradox of remote work: we miss people while becoming less willing to tolerate the effort of being around them.

It is easy to describe this as hypocrisy or introversion. I think it points to something more structural.

Remote work did not eliminate connection. It privatized the work of creating it.

Connection used to be ambient

The office gave us a baseline of social contact without requiring us to produce it.

We greeted someone in the elevator. We overheard a conversation. We asked a minor question instead of scheduling a call. We spent five unremarkable minutes next to a colleague while waiting for coffee.

None of these encounters felt important enough to arrange deliberately. Collectively, they created familiarity, weak ties, and the feeling of participating in a world beyond ourselves.

Once work moved home, most of that disappeared. What remained were interactions that required intention: send the message, compare calendars, choose a place, cross the city, arrive in the right mood, and sustain attention.

This changed the cost structure of social life. We did not necessarily start wanting people less. We lost many of the forms of contact that made being around them inexpensive.

A 2026 study published in Science analyzed five nationally representative U.S. surveys covering 588,322 respondents. After the pandemic, people in remote-capable occupations spent about 1.1 additional waking hours alone per workday relative to people in occupations that could not move online. They did not substantially compensate by socializing more after work. The increase in extreme solitude was especially pronounced among people living alone.

Those findings do not mean every remote worker is lonely. Someone with a partner, children, neighbors, and an active community may welcome solitude during the day. Someone living alone in a new city may experience the same working arrangement very differently.

The larger point is that connection is not produced by desire alone. It is partly an infrastructural outcome. Certain environments make repeated encounters likely. Others leave us to manufacture those encounters ourselves.

The rising cost of seeing people

Loneliness and social exhaustion look contradictory only if we treat all contact as interchangeable.

They are not.

A casual conversation with a familiar colleague asks little of us. A dinner, networking event, or full day in an office asks for travel, preparation, attention, and often performance. A person can be deprived of the first kind of contact while feeling overwhelmed by the second.

Most office interactions were not deep. That was partly their value.

A close friendship asks something of us. A weak tie sits between intimacy and anonymity: the person who knows enough to ask how the presentation went but not enough to require a complete account of our life. These relationships rarely appear on a calendar. They accumulate through repetition.

Remote work is good at preserving relationships with a reason to exist. Teams still meet because a project requires it. Close friends still make plans because the relationship matters. Remote work is much worse at producing encounters that have no obvious purpose.

It also gives us more control over our environment. At home, we can regulate noise, temperature, interruptions, appearance, silence, and how much of ourselves other people can see. That control is one of remote work’s real benefits. It can also make ordinary human unpredictability feel unusually expensive.

In an office, social life included hundreds of tiny negotiations. Someone interrupted while we were concentrating. A colleague arrived in a bad mood. A conversation lasted longer than we wanted. We had to recover from awkwardness, accommodate different working styles, and accept that not every moment would be optimized for our comfort.

Much of this was annoying. Some of it was genuinely harmful. But it also gave us daily practice in the friction of coexistence.

This does not mean remote work has weakened everyone’s “social muscle.” For caregivers, disabled and neurodivergent workers, or people escaping discriminatory and overstimulating workplaces, greater control may expand life rather than narrow it. Avoiding a bad office is not a psychological disorder.

The relevant question is not whether comfort is bad. It is what our preferred arrangement enables us to do and what it gradually makes harder.

Maintaining relationships is not the same as building them

Remote work offers an unequal social bargain.

Established professionals often built their networks, reputations, and workplace fluency in person. They can work remotely while drawing on relationships accumulated over many years.

Younger workers are asked to operate through digital channels before acquiring the relationships and tacit knowledge that make those channels effective. The same arrangement that allows a senior employee to maintain social capital can leave an early-career employee struggling to build it.

This may help explain why the youngest workers do not fit the assumption that digital natives naturally want fully remote lives. In Gallup’s May 2025 U.S. survey, only 23 percent of remote-capable Gen Z employees preferred fully remote work, while 71 percent preferred hybrid work. Twenty-seven percent reported feeling lonely for much of the previous day, compared with 10 percent of Baby Boomers.

The numbers do not establish that remote work caused the difference. Younger adults were already more likely to report loneliness. Housing, income, relationship status, and career stage all matter.

But “Gen Z lacks social skills” is too shallow an explanation. Millennials and Gen Z did not experience a gradual evolution in workplace norms. They experienced a rupture. Many entered adulthood or began their careers while the institutions that once generated routine contact were disappearing.

They may not want the old office back. They may want access to what older workers built there.

Remote work is also only one part of a wider social transformation. Social media turned communication into a constant but fragmented stream. Delivery apps reduced the need to enter public space. Housing costs pushed people away from friends and city centers. Dating apps reorganized intimacy around a marketplace of options. Cities lost third places where people could stay without buying much. The pandemic disrupted habits, relationships, and trust.

Even “social battery” may describe exhaustion more than social decline. A long commute, an open-plan office, or mandatory workplace performance can consume energy people might otherwise give to relationships they actually value. For some workers, staying home creates more capacity for family, friends, and neighbors not less.

Remote work is therefore unlikely to be the sole cause of the loneliness paradox. But it may amplify it by removing one of the last institutions that placed adults in repeated, unplanned contact with the same people.

Flexibility needs coordination

None of this makes a five-day office mandate the obvious answer.

Remote work returned hours previously lost to commuting. It gave people more autonomy and allowed many to concentrate without the noise and performance of an office. It made work accessible to people whose health, location, or caregiving responsibilities had excluded them from certain jobs.

These are not minor conveniences. They changed who could participate in work and what a sustainable working life could look like.

The evidence also resists simplistic return-to-office arguments. In a randomized trial of 1,612 Trip.com employees, a hybrid schedule with two days at home improved job satisfaction and reduced attrition by one-third without damaging performance or promotion outcomes over the following two years.

Remote work can improve satisfaction and retention while increasing isolation. Those findings are not contradictory. They measure different dimensions of life.

The more useful tension is not remote versus office work. It is individual flexibility versus collective coordination.

When everyone chooses independently where and when to work, each person gains freedom, but no one can count on anyone else being present. An organization can preserve autonomy while losing the repetition through which relationships form.

An office full of people is not necessarily socially useful. Neither is commuting to sit on video calls. What matters is coordinated presence: seeing the same people often enough, under conditions that justify being together.

That might mean team days organized around work that is genuinely better in person rather than arbitrary attendance quotas. It might mean informal mentorship, recurring coworking groups, neighborhood institutions, or digital spaces designed for casual contact instead of turning every interaction into a meeting.

The responsibility cannot fall entirely on the lonely individual. Telling people to schedule more dinners misunderstands the problem. When connection becomes another task to manage, the people most depleted by isolation are also the people asked to perform the most labor to escape it.

We do not need to abandon flexibility. We need to rebuild deliberately what once happened without planning: weak ties, repeated encounters, low-stakes conversation, and predictable moments of shared presence.

Autonomy works best when it rests on social infrastructure. The next phase of flexible work should begin building that infrastructure again.

What forms of unplanned connection have you lost since working remotely and have you found anything that genuinely replaces them?