The digital third place: why group chats and anonymous rooms replaced the public feed
Public feeds still have record audiences but nobody's talking on them anymore — the conversation moved into group chats, DMs and small rooms, and the door policy of those new spaces will shape who feels connected and who gets left out.
Start chatting now →The people who run the feeds have stopped pretending
When the head of Instagram writes "that feed is dead," it stops being a hot take and becomes the industry's official position. In a year-end memo dissected by Om Malik on 1 January 2026, Adam Mosseri admitted that "people stopped sharing personal moments to feed years ago" and that "the primary way people share now is in DMs: blurry photos and shaky videos of daily experiences." Sociable life online has been quietly relocating into group chats, DMs and small rooms — what a lot of researchers now call the digital third place, borrowing Ray Oldenburg's old term for the pub, the barbershop, the café: the low-stakes spot that isn't home and isn't work.
The numbers behind the vibe are stark. Charts shown during Meta's FTC trial had friend-posted content falling from 22% to 17% of Facebook time and from 11% to 7% on Instagram between 2023 and 2025. Mosseri had already conceded around 2023–24 that teens spend more time in DMs than in Stories, and more in Stories than in the feed. Meta's own figures show Reels reshared over 2 billion times a day — mostly into DMs. The public layer has become a vending machine that feeds the private one.
What a digital third place actually is
Oldenburg's 1989 book The Great Good Place set out the conditions: neutral ground, a levelling of status, conversation as the main activity, regulars mixed with newcomers, low-stakes playfulness. Public feeds violated nearly all of them. They replaced conversation with broadcast, levelling with follower counts, and neutrality with what researchers Marwick and boyd called context collapse — performing to your boss, your ex and ten thousand strangers at once.
The psychology of why that failed is well documented. A 2024 meta-analysis in the Journal of Computer-Mediated Communication — 141 studies, 897 effect sizes, roughly 145,000 participants — found most links between social media and wellbeing negligible, with one standout: active, interactive use predicted perceived online social support at r = .34, roughly triple the effect of passive scrolling. Connection comes from reciprocal back-and-forth with a bounded audience, not from audience size. Group chats are, structurally, machines for exactly that. A January 2025 interview study (Kim, Klein-Balajee, Kelly and Hiniker) found Discord had independently rebuilt Oldenburg's architecture — 21 design elements mapping onto third-place traits, with pseudonymity as the leveller and persistent "regulars" providing repeated low-pressure contact. Discord's scale suggests demand: roughly 656 million registered users and about 850 million messages a day in 2025.
Is public social media actually dying? Depends what you count
Here the panel of evidence genuinely splits, and it's worth keeping the disagreement visible. Pew's November 2025 survey of over 5,000 US adults shows every major platform growing or flat since 2021: TikTok up from 21% to 37% of adults, Instagram from 40% to 50%, Reddit from 18% to 26%. Consumption is at record highs. What died is peer-to-peer posting. The sceptical reading: the platforms didn't die, they completed their conversion into broadcast television — a small creator class performs, everyone else watches.
But that reframing arguably strengthens the retreat story rather than refuting it. Watching isn't connecting. Passive consumption is precisely the behaviour that studies from Bournemouth University (2023) and the EU Joint Research Centre (2024) linked to loneliness and psychological distress. A stadium with record attendance is still not a town square.
The catch: the new third place has a guest list
The most uncomfortable finding in the whole debate is distributional. A physical third place supplied weak ties and serendipitous entry — Sandstrom and Dunn's research showed even minimal interactions with acquaintances boost belonging. A group chat, by contrast, requires you to already have five friends. The socially rich have converted their social capital into invite-only rooms; people without that capital are left scrolling the very feeds most associated with distress.
One behavioural researcher on this question put it bluntly: expect loneliness inequality, not average loneliness, to be the metric that moves by 2030. Even the sceptics co-sign that prediction. If the internet's middle layer — public, weak-tie, stranger-tolerant space — is the thing genuinely disappearing, the people who most needed a door they could just walk through are the ones losing it.
Privacy you feel versus privacy you have
There is a second catch. What users bought in the retreat was social privacy — freedom from context collapse, from the algorithmic audience, from screenshot-ready performance. Around 58% of Gen Z say they speak more freely in group chats. What they did not necessarily get is institutional privacy. In May 2026, Meta removed end-to-end encryption from Instagram DMs, days before Take It Down Act enforcement began. Discord announced mandatory age verification in February 2026 — face scans or government ID — months after a third-party vendor breach exposed roughly 70,000 users' IDs and selfies.
The honest takeaway: a private-feeling room and a private room are different products, and almost every mainstream platform sells the first. Marketers already treat DM shares as premium signals — nearly half rate them above likes. The living room has a landlord, and the landlord can usually read the room.
The anonymous-room graveyard, and what it teaches
Anonymous spaces have their own cautionary history. Secret shut down in 2015 after a roughly $100 million valuation. Yik Yak lost 76% of its usage in 2016 and sold its IP for $1 million — a quarter of one percent of its $400 million peak — and by March 2024 the UNC university system moved to ban Yik Yak, Fizz, Sidechat and Whisper. The lesson is not that people dislike anonymity. It's that unmoderated intimacy at scale curdles into harassment. The anonymous rooms that survive are the ones that pair anonymity with actual human moderation and visible norms — regulars, house rules, someone who can cut you off, like the bartender in Oldenburg's tavern.
Where that leaves the rest of us
Synthesis, then. The retreat from public feeds is real and rational: people moved to where the connection mechanics actually work. But the group chat is a gated third place, and the gate matters. What's missing is the walk-in room — somewhere a stranger can enter without an invite, an ID scan, or an existing social graph.
That's the niche small, moderated, anonymous chat spaces try to fill. Shush is one attempt: a free, text-only site with a live Lobby and user-created rooms, no signup, 18+. It isn't a private bunker — messages may be reviewed by an admin for safety, which is the tavern-bartender model rather than the sealed-envelope one — but it is a door you can walk through tonight without knowing anyone first.
And a plain-spoken footnote, because this topic touches loneliness: if you're struggling rather than just bored, talk to a human trained for it — call or text 988 in the US, 116 123 (Samaritans) in the UK, or find local services at findahelpline.com.
The public feed isn't coming back as a social space; even its architects say so. The question for the next few years isn't whether the internet rebuilds its third places — it's whether the doors stay open to people who arrive alone.