Instagram's head says engagement falls by half without the algorithm

Original Article ↗ Hacker News Discussion ↗

Instagram’s chief says user engagement drops by up to 50% and “satisfaction scores fall off a cliff” when its recommendation algorithm is turned off, prompting broader questions about what “engagement” really measures. Many commenters argue that lower usage would be a societal benefit, likening algorithmic feeds to addictive products such as tobacco or slot machines, and calling for regulation, transparency, and better defaults like chronological, friends-only feeds. Others note that recommendation systems can be useful when user-controlled, but criticize Instagram and similar platforms for optimizing primarily for ad revenue rather than user wellbeing.

Overall reaction to “engagement falls by half”

  • Many commenters see a 50% drop in engagement as positive for users (less addiction, less time wasted).
  • Others describe this as exposing how dependent the platform is on “toxic” engagement-boosting features.
  • A few note this still implies Instagram would remain very large and viable even without the algorithm.

Algorithms, addiction, and harm analogies

  • Engagement is repeatedly equated with addiction rather than value.
  • Algorithms and infinite scroll are compared to:
    • Cigarettes and nicotine (engineered addictiveness, tobacco-industry playbook).
    • Slot machines and casino design (bright, optimized reinforcement).
    • “Digital sugar” and drug suppliers.
  • Some argue social media used to be net-positive, and design choices turned it into something harmful.

User experience and desired features

  • Strong demand for:
    • Chronological, friends-only or follow-only feeds.
    • A clear marker when a personal feed ends, with optional “discovery” after that.
  • Complaints about:
    • Spammy brand and business posts overwhelming feeds.
    • Poor search, intrusive verification, and media that interrupts audio.
  • Several users appreciate friction-inducing features (e.g., ad breaks, bad mobile web UX) because they help them stop scrolling.
  • One commenter notes it’s unclear whether users can truly disable the algorithm at all.

Discovery, alternatives, and algorithms’ upside

  • Some value algorithms for discovering new creators and content, when “tamed” to their interests.
  • Others reply this “liking” is itself a trained response to an addictive system.
  • Alternative platforms (Bluesky, HN, federated models) are discussed as less addictive or more user-controlled, though discovery trade-offs remain.

Regulation, taxation, and design constraints

  • Proposals include:
    • Regulating or banning engagement-optimizing recommender systems.
    • Treating them like addictive additives in food or drugs.
    • Mandating options to disable infinite scroll or recommendation feeds by default.
    • Taxing infinite scrolling or short-form video to fund education.
  • Some argue only regulation and education can counter “giant addiction machines”; others stress underlying user demand and personal responsibility.