Why social networks stopped being social

Picture the last half hour you spent on a social app. How much of it was someone you know, saying something to you? And how much was a stranger’s video, chosen for you by a recommendation system, that you watched and forgot?

If the answer is “mostly the second”, you’re not imagining it. The big social networks have quietly changed what they’re for. This post looks at what changed, what research says about the thing we lost, and why we think the answer is a table rather than a better feed.

From friends’ posts to strangers’ videos

For years, a social network meant a feed of posts from people you knew. That’s no longer how the largest ones are used.

In the U.S. Federal Trade Commission’s antitrust case against Meta, the court’s ruling cites Meta’s own internal data from January 2025. It found that Americans spend only 17% of their time on Facebook viewing content from their friends, and 7% on Instagram. The judge’s summary of what replaced it is blunt: “For the most part, short videos posted by strangers and recommended by AI.”

The trend is still moving. According to the same ruling, in the two years to January 2025 the share of Facebook time spent on friends’ content fell by almost a quarter, while time on Reels more than doubled. On Instagram, the share spent on friends’ content fell by more than a third, and time on Reels more than tripled.

Two caveats matter. These are figures for users in the United States, not Europe. And the ruling went Meta’s way: the court concluded that Facebook and Instagram now compete with apps like TikTok and YouTube. That’s rather the point. By the court’s own account, the apps we still call social networks now largely work like entertainment channels.

Where the talking went

People didn’t stop talking to their friends. They moved. The same ruling describes how, as smartphones spread, people increasingly kept up with friends in group chats, noting that “texting a medium-sized group of close friends, it turns out, was often more appealing than broadcasting updates to 1,000 acquaintances.”

We think that’s a healthy instinct. Group chats are where a lot of real conversation happens now. But they have one limit: they’re made of people you already know. A group chat rarely introduces you to a nurse from Lyon or a game designer from Tallinn with an opinion you’d never have heard otherwise.

So the old social network split in two. Talking moved to private chats with familiar people. The open, public part became a place to watch rather than to talk. What got lost in between is the conversation with a stranger.

What research says about talking to strangers

It’s tempting to think we lost it because we didn’t want it. Psychology research suggests we’re simply bad at predicting how good it will be.

In a set of experiments published in 2014, researchers at the University of Chicago asked commuters on trains and buses to either talk to a stranger near them, keep to themselves, or commute as usual. The commuters who talked reported a more positive experience, and no less productive a one, than those who sat in solitude. Other commuters, asked to predict the outcome, expected the opposite. The authors trace part of the mistake to underestimating how interested other people are in connecting.

A later series of experiments looked at depth. People expected deep conversations with strangers to be more awkward than they turned out to be, and those conversations created more connectedness and happiness than participants predicted. The researchers found that people underestimate how much strangers care about what they have to say, and that this keeps conversations shallower than they need to be.

Put those findings next to the feed data and a picture emerges. The apps got better at holding our attention, and our own expectations talk us out of the conversations that might have been better for us.

Feeds and tables are built for different things

This isn’t an argument that feeds are evil. A feed is good at what it’s built for: showing you a lot of things quickly, and learning what keeps you watching. But being watched isn’t the same as being heard.

A conversation needs different things:

• A small group. You can only really talk with a handful of people at once.
• A shared moment. Everyone is there at the same time, and a reply comes back while the thought is still warm.
• Someone who can answer. A stranger’s video can’t ask you a follow-up question.
• A reason to listen. Talking with someone who sees the world differently only works if both of you expect to be heard.

None of these are things a recommendation algorithm is trying to optimise, and we don’t think a better algorithm fixes that. It’s a different shape of product.

What we’re trying instead

spareachair is our attempt at that different shape. Instead of a feed, it runs scheduled online Events, such as an Evening, a Lunch break or a Morning coffee. You sit at a small table of 3–6 verified adults, talk for a round of 15 to 30 minutes, depending on the format, then the tables reshuffle and you meet new people. Each table is deliberately mixed, so you get conversations with people who see the world differently.

It starts in text, and voice and video are only ever your choice: mic and camera are off by default at every new table, and nothing is recorded. Everyone is 18+ and ID-verified before joining an Event. You can read the details in how an Event works, on our safety page and in the FAQ. We also walk through a whole Event, round by round, in how an Event works, and how we keep it safe.

The idea came from a real evening: a comics premiere in a pub, over four hours, and about 30 strangers who thought and lived differently.

The takeaway

The social networks didn’t stop being social because people stopped wanting to talk. They changed what they show, and our own instincts make it easy to settle for watching. The research suggests that the conversation you’re hesitant to start is often better than you expect.

spareachair is free, for verified adults (18+) in the EU/EEA. Get the app from the links on our home page.

Sources

• Federal Trade Commission v. Meta Platforms, Inc., Civil Action No. 20-3590 (JEB), Memorandum Opinion. U.S. District Court for the District of Columbia, 2025. https://www.ftc.gov/system/files/ftc_gov/pdf/MemorandumofOpinion.pdf
• Epley, N. & Schroeder, J., “Mistakenly Seeking Solitude”. Journal of Experimental Psychology: General 143(5), American Psychological Association, 2014. https://doi.org/10.1037/a0037323
• Kardas, M., Kumar, A. & Epley, N., “Overly Shallow?: Miscalibrated Expectations Create a Barrier to Deeper Conversation”. Journal of Personality and Social Psychology 122(3), 367–398, American Psychological Association, 2022. https://doi.org/10.1037/pspa0000281