More photos are being taken than at any point in history, and fewer of them are being posted. Media researcher Ruby Justice-Thelot has a name for the reason, and it explains why a dozen small apps are suddenly trying to rebuild social media from the ground up.
The reason is “connection bloat.” Speaking on the Media Studies podcast, Justice-Thelot described what happens to a profile over years: “People encounter something I call connection bloat over time. They accumulate more and more and more and more followers and following.” Nobody deletes anyone, the audience silently grows, and posting stops feeling casual. “People are taking more photos than ever, but they’re not posting them,” she said. “In part because there’s this reticence to post. It feels like such a commitment.”
Her answers to that anxiety are not new features on Instagram. They are a portfolio of separate, smaller platforms — each designed to do one of Instagram’s jobs, and only one — and each with a number attached that hints at whether the model can survive.
The four jobs Instagram does, and the startups picking one each
Justice-Thelot’s framework is a 2×2. One axis is whether you are encountering daily life or taste; the other is whether the relationship is existing or new. Instagram, she argues, sits at the center doing all four jobs at once.
“Instagram does all four,” she said. “Keep up with your people. Keep up with your taste. Discover new things. And then it’s also to discover the lives of ordinary people.”
That breadth is exactly what opens the door for challengers. “The premise and the promise of a lot of these new platforms is that they will kind of own one of the corners of this quadrant and do it better and in a more functionally convergent way,” she said. The table below maps who is taking which corner, with the figures she cited.
| Platform | Corner of the 2×2 | Core mechanic | Users (as stated) | Monetization |
|---|---|---|---|---|
| Retro | Friends / existing life | Weekly photo journals, private likes, friend list, group albums | 1.7M active | Premium, ad-free subscription |
| Perfectly Imperfect | Taste expansion | Personalized profiles of what people are into; celebrity participants | ~500K (larger Instagram following) | Member-supported, no ads, some partnerships |
| Posting | Ordinary life | Chronological feed, no algorithm | ~40K unique users | User/member support |
| Minish | Ordinary life | Random daily one-minute capture prompt; archive over time | ~40K monthly participants | Subscription; printed books and posters |
| Corner | Taste / places | Save and organize places, follow recommendations, import links | ~50K community members | Free app with a shop |
| Locket | Friends | Photos arrive via home-screen widget | ~10M DAU (per a Google Cloud case study) | Not specified |
| Yope | Friends | Private groups, lock-screen delivery, AI memory recaps, streaks | 14M (company claim for 2026) | Free + in-app purchases |
| Cosmos | Taste / visual references | Visual reference archive with attribution recovery, following feed | 2M | Premium; Series A funded |
| Arena | Taste / images | Channels of images, links, PDFs, text | 40K MAU, 20K paying | Subscription |
The spread is the point. Some of these want scale — Cosmos has raised a Series A and is on a venture path, and Justice-Thelot said she would like to have its team on the show. Others do not. “The goal for something like Retro is not to get huge,” she said. “The goal is to be sustainable, is for it to be positive for its users.”
The same self-imposed ceiling shows up at Arena, whose team she names as Chris Sherron, Charles Broskoski, Daniel Pianetti, and Denny Zucconi — founders she describes as openly comfortable at roughly 40,000 monthly active members and 20,000 paying, with no pressure to grow.
Snapchat’s real lesson: you don’t need the big network, you need the right one
If the alternative platforms sound too small to matter, Justice-Thelot has a counterexample that has been hiding in plain sight for a decade: Snapchat.
Her field observation is casual and striking. When she asks young seatmates on planes why they use Snapchat, the answer is that their college or high school friends are there. The friend group is platform-based, so leaving means losing the group — regardless of how many other users the app has globally.
“You don’t need the large network. You need the right network,” she said.
That directly challenges the orthodoxy that consumer social apps must grow as fast as possible to build sticky network effects, the bet Instagram made for a decade — more users, more creators, more influencers, more entertainment. The takeaway for anyone building or funding the next wave: retention can come from a small, correct social graph rather than a large one.
The web is closing, and Reddit is the tell
The fragmentation is happening against a second, quieter shift: the open web is closing. Reddit was reportedly moving toward a sign-in-required model, which Justice-Thelot reads as part of a broader move “from the open web to a more closed web.”
She ties it to the “Dark Forest” theory she and collaborators at the Dark Forest Collective have been writing about for nearly five years, naming Joshua Citarella, Nadia Asparouhova, and Yancey Strickler as associated thinkers. Borrowed from Liu Cixin’s The Three-Body Problem, the idea is that visibility creates danger, so actors hide: as social media exposure invites retaliation and cancellation, people retreat into separate realms with their own norms.
The commercial stakes for Reddit are unusually literal. Its threads are a major source of training and answer data for large language models — think “best dishwasher” queries — and that content was previously visible both to scrapers and to casual visitors who might see an ad. A sign-in wall changes who gets access, and therefore who captures the value.
Justice-Thelot also flags pushback against TikTok in Spain and the wave of school-district phone bans and national social media bans. She has been persuaded by a co-host, Taylor, to oppose banning children, partly because such measures end up controlling what is posted, not just who posts. She describes herself as “by and large a strong believer in freedom of speech both in the real world and on the internet.”
Why the glasses matter: a bet that undivided attention is worth more
The hardware thread is brief but it frames the whole episode. Justice-Thelot’s conjecture is that platforms are moving into glasses because attention has been diluted, and immersive hardware restores its value.
Her evidence is household arithmetic. The number of electronic devices per US household rose from roughly one television in the 1980s, to two in the 1990s, to three in the 2000s, to perhaps eight today. Attention is now split six to eight ways, so the advertising value of any single screen’s eyeballs has fallen. A headset or a pair of glasses, by contrast, captures all of it.
She frames this as vertical integration in service of selling higher-value attention: if an eyeball on the iPad after 7pm is worth a third — because it competes with the TV and the phone — the eyeball inside Vision Pro or Spectacles is worth full price. She explicitly calls this “pure conjecture.”
The trigger was Snap’s renewed Spectacles push, which she says included a campaign featuring Robert Downey Jr. and a social media blitz in late October 2026. She admits she has not tried them and that they “looked a little big.” She calls Apple Vision Pro “in many ways a brilliant product because it is a product of total immersion.”
She also mounts an unexpected defense of targeted advertising, arguing the pre-digital alternative was worse: Sunday golf coverage served pharmaceuticals aimed at people in their 60s, which she calls “very violent to be exposed to something so irrelevant.” By contrast, Instagram surfaces a knife holder she actually wants for her kitchen, and ad networks connect small makers in Spain or China to buyers in Vietnam or Brussels — “a marvel and a feat of modern technology.”
AI is splitting the last shared cultural experience
Two AI product moves anchor the personalization shift. Instagram now lets users describe an effect in natural language instead of picking from a fixed list — prompt-based creation. And Spotify has an AI remix-and-covers tool, reported by the Financial Times on October 30, 2026, targeted at the first quarter of 2027 around the Grammys, priced at roughly $4–5 more per month.
Justice-Thelot’s lens: the old Instagram dog-face filter was a “digital monocultural phenomenon” — Kylie Jenner and your high-school friends used the same effect. Prompt-based effects fragment that shared experience. The same logic applies to hardware and interfaces more broadly: “my Instagram is the same as your Instagram and my iPhone is same as your iPhone,” one of the few common experiences left.
On Spotify she holds two views at once. Sampling is a creative lineage she grew up on — J Dilla’s Donuts, A-Trak’s turntablism, The Avalanches’ Since I Left You built entirely from samples, and their new record incorporating internet and video-game sounds. Democratizing remix extends that tradition. But songs are “one of our last bastion of shared cultural understanding,” and she notes that the No. 1 song for 20 weeks — “Choosin’ Texas” — was unknown to many people close to her as she traveled. Letting everyone remix it erodes the common ground further. She concedes a jury remix of “Choosin’ Texas” already exists.
Connection bloat has a second cause: we scorn ordinary life
The posting-anxiety diagnosis has two halves. The first is connection bloat. The second is contempt for the mundane. “We’ve started to scorn our own mundane lives,” she said. “We think that like going to the cobbler is not post-worthy.”
“What I love about a lot of these platforms is that they’re not trying to capture the most exciting moments of one’s life,” she said. She illustrates with John Williams’s novel Stoner, which she summarizes as “about a man who dies” and which she rereads for the beauty of a simple life.
The new platforms counter the anxiety two ways: smaller audiences, following the logic of Finstas and close-friends lists, and constraints on when and how often you post — Retro’s weekly ritual, BeReal’s simultaneous daily post, Minish’s random daily alert. Constraint makes posting meaningful instead of a free-for-all.
She draws an explicit parallel to religion. Religare means to connect; église derives from the Greek ecclesia, meaning community. BeReal’s synchronized daily post resembles praying together; Retro’s weekly post resembles going to mass.
The small-audience paradox: freedom needs fewer eyes
If the constraint thesis sounds like a marketing line, Justice-Thelot’s own explanation is blunter. Freedom, she argues, is often a function of a small audience.
“Sometimes in order to be truly unhinged, in order to be truly free, you need a smaller audience,” she said.
And the mechanism that makes a platform feel authentic is the same one that makes it small: “The way to get to authenticity is to disintermediate sometimes the personal profile to the taste profile.” People already do this instinctively — a potter starts a second Instagram so pottery does not mix with colleagues and news. Arena’s and Cosmos’s archives of liked images act as a separate self from the person eating at restaurants and playing games with friends.
She connects this to her 2015 project on the self-archive: the archive is the closest thing to true nature in the digital realm, and Cosmos leans directly into that idea. Her practical advice for readers: treat these platforms as complements to Instagram, not replacements, and “decorticate” — separate functions across platforms to suit your own needs. She is explicit that she likes Instagram and finds it “an amazing platform,” citing Caleb Heron’s videos as an example of what it does well.
What to watch
The through-line worth tracking is whether the closed-web shift and the personalization shift reinforce each other. A sign-in Reddit and a Dark Forest internet both reduce what is publicly visible, while prompt-based effects and AI remixes reduce what is commonly shared — the same fragmentation seen from two directions.
For the market, the open questions are concrete. Will Reddit’s sign-in model attract users or repel them? Will Spain’s TikTok pushback and school phone bans produce measurable effects? Will Spotify’s remix tool ship on its first-quarter 2027 target? And will the small platforms stay small by choice, or get pulled toward scale by their funding — the difference between a Retro and a Cosmos?
Justice-Thelot’s own framing is that these are trade-offs rather than verdicts: personalization gains customization and loses shared context, and smaller audiences gain freedom and lose reach. For investors, the signal is that the definition of a winning consumer social company may be quietly changing — from a company that captures the most attention to one that asks for less of it, and charges its members instead of its advertisers.