HN Debrief

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

  • Social Media
  • Media
  • Advertising
  • Consumer Internet
  • Policy

The BBC article says the social part of social media has been hollowed out. Facebook, Instagram, TikTok, and similar apps increasingly show professionally made or trend-driven content from strangers because that keeps people scrolling and keeps ad revenue flowing. Personal updates from friends have thinned out, partly because people now share more in WhatsApp groups, Discord servers, and other private spaces, and partly because the platforms themselves bury friend activity behind recommendation engines. The comments largely agreed that this shift happened years ago, not recently, and framed it less as a cultural quirk than a business model outcome. Once feeds optimize for engagement, outrage, fear, novelty, and repetition win. Several people compared today’s feeds to cable news or tabloids with better targeting, more precise reinforcement, and far less friction.

If you still treat Facebook, Instagram, or X as relationship tools, you are already using the wrong layer of the internet. For products, marketing, or community work, assume public feeds are attention markets and move real trust-building into direct, opt-in, or chronological channels.

Discussion mood

Mostly negative and resigned. People see mainstream feeds as manipulative attention machines driven by ad-funded engagement, with a smaller undercurrent of nostalgia for early Facebook and cautious optimism about private groups, RSS, and stripped-down tools that remove recommendations.

Key insights

  1. 01

    Removing recommendations reveals an empty network

    Cutting out non-friend content with tools like ReVanced or using Instagram's hidden Following view exposes how little genuine social posting is left. That changes the diagnosis. The problem is not only that algorithms crowd out friends. Friends mostly stopped posting because the platforms trained everyone to consume creators, not share with each other.

    Measure how much real social graph activity is left before investing in any strategy built on public social feeds. If a filtered friend-only feed is nearly empty, move community and retention efforts to messaging, email, events, or niche groups instead of trying to revive the main feed.

      Attribution:
    • Ajedi32 #1
    • al_borland #1
    • testudovictoria #1
    • andix #1
  2. 02

    Engagement economics force the drift to slop

    The strongest explanation was brutally simple. A public company cannot build endless growth on baby photos and life updates, so it turns the product into a generic entertainment channel. Creators then adapt to the same pressure with negative framing, clickbait thumbnails, and constant novelty because the algorithm rewards what keeps people watching, not what strengthens relationships.

    Treat feed quality as an incentive design problem, not a moderation problem. If your product or growth plan depends on engagement-ranked discovery, expect sensational content and professionalized posting to outcompete genuine social use.

      Attribution:
    • jandrese #1 #2
    • tantalor #1
    • lenerdenator #1
  3. 03

    Users are building their own anti-feed layer

    A lot of the practical value came from people describing ad hoc countermeasures. They use ReVanced, Morphe, Unhook, UnTrap, uBlock filters, hidden chronological feeds, disabled watch history, and browser-only access to strip out Shorts, recommendations, and autoplay. The pattern matters more than any one tool. People still want the utility of these services, but only after amputating the recommendation engine.

    There is demand for products that preserve utility while blocking discovery mechanics. If you build on top of existing platforms, give users hard controls over recommendations, autoplay, and feed source rather than soft wellness nudges.

      Attribution:
    • pwndByDeath #1
    • zbikowski #1
    • mattbruv #1
    • nunez #1
    • Anthony-G #1
    • captainclam #1
  4. 04

    Early Facebook worked because it mirrored real relationships

    The nostalgia was not just sentimentality. Several commenters pinned a real functional difference between early Facebook and today's feeds. In the mid-2000s, the network mostly reflected people you actually knew, events you might attend, and photos tied to offline life. Messenger and status updates strengthened weak ties. Once popularity ranking and mixed content feeds took over around 2009 to 2011, the product stopped organizing social life and started interrupting it.

    If you want a social product to support real relationships, constrain network scope and default to chronological delivery. The more the feed is allowed to optimize across the whole network, the less useful it becomes for actual coordination and memory.

      Attribution:
    • Insanity #1
    • semitones #1
    • dbspin #1
    • pryelluw #1
    • LtWorf #1
  5. 05

    The older media critique still fits perfectly

    People pulled in Neil Postman's Amusing Ourselves to Death and Ben Bagdikian's The Information Machines to argue that the current mess is less a shocking break than a more efficient version of old media pathologies. The novelty is not sensationalism itself. It is precision targeting, personalization, and endless repetition delivered by software instead of editors and broadcasters.

    Do not frame this as a totally new class of harm that needs totally new instincts. Older media critiques about format, incentives, and concentration still apply, and they can guide both product decisions and policy arguments.

      Attribution:
    • spking #1
    • randusername #1
    • PaulHoule #1
  6. 06

    This is not just a technical problem

    Several commenters rejected the idea that smarter builders can simply code their way out. They pointed to the fact that the same industry talent built these systems in the first place and that ad-tech incentives swamp good intentions. Mastodon came up as evidence that different structures can produce healthier behavior, but the deeper point was that governance, liability, and business models matter more than interface tweaks.

    When evaluating any alternative platform, look past feature lists to ownership, revenue model, and moderation structure. A healthier feed needs institutional constraints, not just better UX.

      Attribution:
    • zerobees #1
    • 2OEH8eoCRo0 #1
    • intended #1
    • sph #1
    • vitally3643 #1

Against the grain

  1. 01

    Not everyone’s feed has collapsed into strangers

    A minority said the article overstated the shift because their personal Instagram or Facebook feed still mostly shows people they know, especially if they keep a small network and use the app lightly. That does not refute the broader trend, but it does show that usage pattern and account hygiene still matter. Smaller circles and intentional habits can preserve some of the old social function.

    Segment your assumptions by user behavior. Heavy users and passive scrollers see a very different product from people with tight networks and deliberate follow lists, so product and policy conclusions should not treat them as identical.

      Attribution:
    • icepush #1
    • Timpanzee #1
    • kayo_20211030 #1
    • hunglee2 #1
  2. 02

    Social media was never that social

    Some pushed back on the nostalgia entirely. In this view, the old promise was always overstated because even early feeds mostly offered shallow awareness, status signaling, and weak-tie maintenance rather than deep social connection. The current turn toward trends and entertainment is an intensification, not a betrayal of some golden age.

    Be careful not to build strategy around a mythologized past. If you are trying to create healthier online interaction, aim for concrete functions like coordination, small-group sharing, or direct conversation instead of trying to recreate a vague 'social feed' ideal.

      Attribution:
    • Insanity #1
    • SirFatty #1
    • seydor #1
    • phendrenad2 #1
  3. 03

    Manipulated feeds predate modern recommender systems

    A few comments argued the fixation on recent algorithms hides a longer history. Search ads, scam funnels, sockpuppet campaigns, and forum manipulation all existed well before TikTok-style recommendation engines. Personalized ranking made the problem worse, but the internet did not suddenly become weaponized in the last decade.

    If you regulate only modern recommendation systems, older forms of manipulation will keep thriving. Threat models should include astroturfing, paid placement, bot amplification, and moderation abuse, not just personalized video feeds.

      Attribution:
    • hedora #1
    • intended #1

In plain english

Discord ↩
A chat platform built around servers and channels, often used by communities, gaming groups, and private groups.
Mastodon ↩
A decentralized social networking platform made up of independently run servers, usually showing chronological feeds instead of algorithmic ones.
Morphe ↩
A tool mentioned in the comments as an alternative app patching system for changing social media app behavior.
ReVanced ↩
A community-made Android patching tool used to modify apps such as YouTube or social apps to remove ads or unwanted interface features.
RSS ↩
Really Simple Syndication, a standard feed format that lets users subscribe to website updates in a reader without visiting each site directly.
uBlock filters ↩
Custom rules used in uBlock to remove or alter specific interface elements on a webpage.
Unhook ↩
A browser extension that removes recommendation features and other distracting elements from YouTube.
UnTrap ↩
A tool or extension for YouTube that disables distracting features like Shorts and recommendations.
WhatsApp ↩
A messaging app widely used for private one-to-one and group chats.

Reference links

Books and long-form critiques

Tools to strip feeds and recommendations

Historical and background references

Evidence of bots and manipulation