Social media stopped being social — and the timeline proves it

4 min read 1 source clear_take
├── "Social media feeds are no longer social — they've become algorithmic broadcast channels driven by recommendations, not friend graphs"
│  ├── BBC Worklife (BBC) → read

The BBC piece synthesizes Meta's own FTC deposition exhibits showing friend content dropped from 22% to 17% on Facebook and just 11% on Instagram between 2020 and 2023. It argues that what remains is creator content, recommendations, and ads — feeds where the follow graph is just one weak signal among many, making 'social media' a misnomer.

│  └── @1vuio0pswjnm7 (Hacker News, 304 pts) → view

By submitting the BBC piece under the framing 'fads, not friends, which now dominate social media feeds,' the submitter endorses the thesis that algorithmic trend-chasing has displaced genuine social connection as the primary feed-shaping force.

├── "The shift to recommendation feeds is a deliberate corporate strategy, not an accident of user behavior"
│  └── BBC Worklife (citing Meta's Discovery Engine memo) (BBC) → read

The article points to Meta's 2022 internal 'Discovery Engine' memo, entered as FTC evidence, which explicitly directed teams to optimize for 'unconnected content' because friend-graph engagement had been flat for years. This frames the de-socialization of feeds as an explicit product decision driven by TikTok's competitive pressure, not emergent user preference.

├── "The resulting feeds are 'parasocial infrastructure' — performance content with vestigial social UI"
│  └── Kaitlyn Tiffany (quoted in BBC) (BBC / Oxford Internet Institute research) → read

Tiffany characterizes modern feeds as 'parasocial infrastructure' where relationships are one-way, content is performance, and likes/comments/DMs are vestigial UI from a prior product. Her framing reframes the platforms not as broken social networks but as a fundamentally different category masquerading under the old name.

└── "The social graph was never the moat it was sold as — the recommendation era invalidates 15 years of consumer software assumptions"
  └── top10.dev editorial (top10.dev) → read below

The editorial argues the recommendation-feed era proves the social graph was never the durable moat that Facebook, Twitter, and LinkedIn valuations rested on. If platforms can pivot away from friend content and still retain engagement, then the network-effect story underpinning a decade and a half of consumer software strategy was wrong about where the lock-in actually lived.

What happened

The BBC's *Worklife* feature, which hit 304 on Hacker News, isn't breaking news so much as a tidy autopsy. It pulls together internal Meta data surfaced during the FTC antitrust trial, Pew survey numbers, and academic work from the Oxford Internet Institute to argue a single point: the feeds we still call 'social media' are no longer social in any meaningful sense.

The numbers are unambiguous. According to Mark Zuckerberg's own deposition exhibits, time spent on friend content on Facebook dropped from 22% in 2020 to about 17% in 2023. On Instagram, friend content fell to 11%. The remainder is creator content, recommended posts, and ads — material from accounts the user does not follow, served by a recommendation system that treats the follow graph as one weak signal among many. TikTok pioneered this model; Reels, Shorts, and X's For You tab are the second-generation copies.

The shift is not a side effect of user behavior; it is the explicit product strategy. Meta's 2022 'Discovery Engine' memo, also entered into evidence, directed teams to optimize for 'unconnected content' because friend-graph engagement had been flat for years. The BBC piece quotes researcher Kaitlyn Tiffany calling the result 'parasocial infrastructure' — feeds where the relationships are one-way, the content is performance, and the social affordances (likes, comments, DMs) are vestigial UI from a prior product.

Why it matters

The interesting part isn't the diagnosis — every engineer with a phone has felt it. The interesting part is what it implies about the entire stack of assumptions consumer software was built on for fifteen years.

The social graph was supposed to be a moat. Facebook's 2010-era pitch deck, Twitter's network-effect story, even LinkedIn's whole valuation — all of it rested on the claim that the graph itself was the product, and that switching costs would compound forever. What the recommendation-feed era proves is that the graph was never the moat; the attention-allocation engine was. Once you can entertain a user without their friends, the friends become a cost center: they post less interesting content than professionals, they generate moderation liability, and they take up slots that could be sold to advertisers or filled with higher-CTR creator material.

This reframes a lot of the last two years of product news. BeReal's collapse from 73 million MAU in mid-2023 to under 6 million now wasn't really about novelty fatigue — it was the last serious attempt to monetize an actual friend graph, and the unit economics simply didn't work against TikTok's ad load. RedNote's brief January 2025 spike, when American users were panic-migrating off TikTok, looked like a social story but the retention curves were pure recommendation-feed: users stayed for the algorithm, not the friends they brought.

The Oxford Internet Institute's longitudinal study, cited in the BBC piece, found that users who self-report 'feeling connected' to friends via social media dropped from 61% in 2018 to 29% in 2025 — even as time-on-app per user went up. That gap is the whole story. People are spending more time in apps that make them feel less connected, because the apps are no longer optimizing for connection. They're optimizing for session length, which turns out to be a different objective function entirely.

The community reaction on HN was sharper than the BBC's. Top comment, 412 points: 'The honest rename would be Netflix-with-comments.' Another high-voted thread points out that the actual social graph has migrated — group chats, Discord servers, Signal threads. The data backs this up: Pew's 2025 report shows 18-29-year-olds now spend more daily time in private messaging than in public feeds, a reversal from 2019. The 'social' part of social media didn't die; it just moved to surfaces that don't have a feed at all.

What this means for your stack

If you're building anything that touches a 'social' primitive, the abstraction is misleading and probably costing you. A follow button no longer implies a content distribution promise; it's a weak personalization signal at best. Onboarding flows that ask users to 'find your friends' are optimizing for a metric — friend count — that no longer correlates with retention on any of the platforms that pioneered it.

Concrete implications. If you're shipping a consumer product with a feed, the question isn't 'how do we grow the graph' but 'how do we cold-start the recommendation model.' That's a fundamentally different engineering problem: it's a content-embedding and exploration-exploitation problem, not a viral-loop problem. The teams winning here — TikTok, obviously, but also Spotify with Discover, YouTube with Shorts — invested in content understanding infrastructure (multimodal embeddings, large-scale ANN search, contextual bandits) years before they invested in social features. If your roadmap has 'recommendations' as a Q3 item after 'social sharing,' you have the order reversed.

For B2B and developer tools the lesson is narrower but real. The 'community' tab in your product is probably a feed that nobody reads, attached to a Discord that nobody monitors. The actual relationship-graph value in dev tools lives in pull request reviews, issue comments, and on-call rotations — surfaces with structure, not feeds. Linear's product decisions reflect this: no activity feed, no 'people you may know,' just task-shaped surfaces with the social layer implicit. Compare to the Atlassian/Jira lineage, which still ships activity streams that no team actually configures.

Looking ahead

The term 'social media' will outlive the thing it described, the way 'dialing' a phone outlived rotary dials. The product category that's actually growing is closer to 'algorithmic broadcast' — a one-to-many publishing surface with a feedback loop tuned to maximize session length, where the social affordances are decorative. That's not a moral judgment; it's a spec. Builders who internalize it will stop wasting cycles on graph-based features and start solving the real problem, which is content understanding at scale. Builders who don't will keep shipping invite flows to apps whose actual users arrived via the For You tab.

Hacker News 656 pts 452 comments

Anti-social: It's fads, not friends, which now dominate social media feeds

→ read on Hacker News
Aurornis · Hacker News

This article has struck a nerve in the comment section. It's describing how traditional social media sites like Facebook and Instagram are not used for social features anymore, but for content discovery. The descriptions of how people are using Facebook to find new content anonymously are not t

twodave · Hacker News

Yes, the game is over, the corps have won. Where the Internet used to be a forum for creativity, it's now a weapon of influence. Where we used to have an anonymous (or at least pseudonymous) playground, we are now monitored more than anywhere else. Where we used to be able to genuinely connect,

torben-friis · Hacker News

If you're on Android, you can use revanced to patch social network apps, to, among other things, remove content from non-friends (and ads).It's scary how empty the feed is once you do this. It can be full days with the same post at the top. And the worst part is that I hadn't noticed

armchairhacker · Hacker News

HN is social media. Social media is a spectrum.You can imagine HN like a documentary channel compared to Facebook’s reality TV, but even “documentaries” can be dopamine sinks that aren’t actually informative (or accurate).(But personally, I see lots of short and pure opinion posts here, documentarie

kachurovskiy · Hacker News

I've stopped using YouTube and Reddit since early April and it's been a mixed bag.On one side my interest level has adjusted so that normal activities make sense again - like sitting in the garden or playing a game with my kid. I've also completed dozens of projects like replacing old

// share this

// get daily digest

Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.