A 'friendship app' connected me with five strangers. We actually hit it off.

On one of the coldest nights in New York this year, I bundled up and went to a trendy tourist trap to meet up with five people who also felt that the pull to make new friends was stronger than the windchill. I didn't know anything about my matches aside from a text I'd received that told me that two of them also like reading, that several enjoyed the same irreverent TV shows I do, like "Arrested Development" and "Broad City," and that one had bungee jumped in New Zealand.

We were meeting as a group curated by the app 222, which asks users to fill out questionnaires with info like their birth order among their siblings, the economic class they grew up in, and how spontaneous they are. I was delighted to see I was paired with other people in their late 20s and early 30s, so I wouldn't have to be the aging millennial at the Gen Z table. (There was a collective sigh of relief on this point.) Despite the freezing cold, everyone showed up — no doubt given extra motivation by the app's strict no-show policy that not only involves added fees but kicks out people who don't cancel with notice.

The app, acting as a sort of virtual host, gave us prompts to get the conversation going, such as "If you could pick another name for yourself, what would it be?". I didn't find them particularly enlightening. We read through some but decided we didn't need generic prompts to chat. The people who had bought into this experience (by paying a small fee to 222 and by opening themselves up to strangers) were friendly and eager to meet someone new.

Founded in 2021, 222 is one of several apps that launched in recent years with the aim of fighting loneliness and finding people real, in-the-flesh friends. Bumble has for years had a friendship setting for swiping. There's also Clockout, Timeleft, and Pie. Timeleft says it has brought more than 400,000 people together in more than 170 cities. In 2023, Bumble estimated that about 15% of its monthly active users were using the friends setting. While social media algorithms are often blamed for entrapping people and isolating them as they spend more of their time on screens than with friends, apps like 222 say their algorithms can pair us up with our new besties or potential romantic partners, sussing out similarities and determining compatibility. A cottage industry of apps is raising millions of dollars on the idea that IRL is the next big thing online.

The apps feel like the next iteration of online dating, which normalized the virtual meet-cute. But they're having a moment as more young people feel burned out by the churn of commodified romance. Finding friendship is lower stakes than finding true love, but that's also a part of the barrier to it. There are fewer norms for pursuing a potential friend: How soon after meeting do I text them? How do I know if they're interested as well? Am I being too needy? Does this person think I'm a total freak? The anxiety spiral can keep any of us from texting someone to hang out.

The point of 222 is to get past that. "If we can use technology to be an invisible facilitator to just nudge people gently in the direction of spending more time with each other, then I think we've truly created more connection in the world," Keyan Kazemian, the 25-year-old cofounder and CEO of 222, tells me. "We've created a chance encounter and potentially a relationship that didn't exist before."

Before the pandemic there was a rise in the celebration of canceling plans, as showcased by viral social media posts and several articles unpacking the psychology behind the act. It felt so good, many people said, to bail and binge Netflix instead of shuttling around to after-work catch-up cocktails with friends. "I'm just not feeling it" became a justification for ditching someone over text on short notice and choosing to rot on the couch instead. Leaving friends on read for days became the norm. People prioritized individualism and the idea of self-care over community and friendships.

The backlash from the self-involved era is in full swing. Turns out, people want friends. They want to practice the lost art of hanging out, central to '90s sitcoms like "Friends" and "Seinfeld." But as we've increasingly lost our connection to "third places," like religious centers, local coffee shops, and community centers, and pivoted toward hybrid or remote work, finding our people has become so much harder. Former US surgeon general, Dr. Vivek Murthy called loneliness an epidemic in 2023, highlighting the serious physical health risks and financial costs of isolation. In a 2024 poll conducted by Morning Consult for the American Psychiatric Association, 30% of adults said they felt lonely once a week, and 10% said they felt lonely daily. Among 18- to 34-year-olds, 30% said they felt lonely each day or several times a week. The definition of loneliness for the poll wasn't just having a slow week or lacking a romantic partner but the much darker dread of "feeling like you do not have meaningful or close relationships or a sense of belonging."

When you look at the increasing ways people are trying to study and understand the science of making friends, the idea of using algorithms doesn't seem so far-fetched. A 2018 study from Jeffrey Hall, a professor of communication studies at the University of Kansas, found that people had to spend about 50 hours together to become casual friends and 90 hours to reach general "friend" status. Another study he conducted in 2023 suggested that reaching out to friends for even quick quality interactions daily could help lower stress and increase happiness.

How much of our compatibility is innate, and how much have we built into each other by spending thousands upon thousands of hours together?

With apps like 222 and Timeleft, the algorithms may matter less than the buy-in: The apps gather people who want to make new friends and then push them to follow up, avoiding some of the awkwardness of pursuing something unreciprocated. "Whether or not the specific magic sauce of the matching is necessary is an open question. It might work, it might not work, it may not matter," Hall tells me. "What really matters is people are caring enough to sign up and show up, and showing up is what we do for our friends." I tried another matched dinner on Timeleft, which tries to match compatible people for Wednesday-night dinners around the US, but half of our group ditched, which made it harder to have the full group experience.

There's also Pie, which brings people together around shared interests — like book clubs, dinner clubs, and dancing — in Chicago, Austin, and San Francisco. The company pays local creators to host events, leveraging the concept of the gig economy and local social groups as an engine for new connections. Andy Dunn, who founded Pie after his run as the cofounder and CEO of the menswear brand Bonobos, says people have to keep seeing each other to form those strong ties. For now, Pie brings people together for recurring events, but it's working on ways to better pair up people at those events who are likely to get along. Dunn says he thinks the answer to ending loneliness isn't technology itself but rather people who decide to show up and put themselves out there. "I don't think the algorithm can do it," he said. "I think we can do it."

The idea is catching on. Pie raised $11.5 million in a Series A round last year, and 222 raised $3.6 million so far. "While technology has offered us so much, including a new dimension of connectivity via social networks, it has had many unintended consequences that have led in part to, ironically, less connectivity in real life," Kirsten Green, the founder and managing partner of Forerunner Ventures who invested in Pie, tells me in an email. "We are entering what we see as the next phase of digital adoption where we better learn to live with this — not by being consumed by it, but by thriving because of it."

I'm thinking about the gaggle of people I've collected as my closest friends over the past 20 years. Two I met through Craigslist and Facebook Marketplace as random roommates who also needed someone to split the rent with. Another was my last-minute prom date, and the fourth I picked out of my 300 or so seventh-grade peers because she wore a Panic! At The Disco T-shirt. Each was brought into my life by repeated proximity — we shared classrooms or kitchens and living rooms and experiences for so long that we eventually became inseparable.

I recently asked one of them: Do you think we're best friends because we can see the subtleties of the world in the same way, or do we see the world the same way because we've been hanging out since 2007? How much of our compatibility is innate, and how much have we built into each other by spending thousands upon thousands of hours together?

Hall says friendships are likely forged at both ends. You discover that you have a similar worldview or sense of humor in certain situations. Over time, your friends also shape you: They can influence your values, your exercise routine, your eating and drinking habits. Similarity alone is not enough. We could read the same book but have completely different interpretations. We might both be dog people, but that doesn't mean I'll like you and your dog. But my random roommates and I, who look different on paper and were brought together over nothing more than similar budgets in the same city, can look at a social situation and walk away with the same vibes and opinions. Would an algorithm match me with these people for dinner? If it did, would I even recognize their potential to be the people who'll stand by me during my most important life events? There's no way to know.

A day or so after my 222 dinner, the app asked me to rate my dinner companions and provide more info on why I did or didn't like someone. It said that would help boost my chance of getting better matches going forward. After filling out the questionnaire, I found that someone I had positively reviewed said the same about me, and the app nudged us to hang out again — taking the awkwardness of pursuing a new pal out of my hands. It may matter less what the algorithm sees in the two of us and more that we're open to keep trying to find our people together.