Shifts in U.S. Social Media Use, 2020–2024: Decline, Fragmentation, Polarization (2025)

Original Article ↗ Hacker News Discussion ↗

Perceived accuracy of the findings

  • Many commenters say the description of a “smaller, sharper, louder” online public sphere feels intuitively right, including on HN: vocal minorities dominate while the broad middle mostly watches or leaves.
  • Several report personal exits or drastic “diets” from social media that improved their well-being.
  • There’s broad agreement that overt political posting is now disproportionately done by the angriest or most partisan users.

Methodology and data skepticism

  • Multiple comments argue the paper’s usage trends conflict with other surveys, especially around YouTube, which other data sources show as still growing.
  • Some suspect the study’s interpretation of “social media” (excluding chat apps like Discord) misses major shifts in behavior.
  • Others point out apparent AI-written code in the project’s repo and AI-detector flags on the text, raising doubts about rigor, though AI detectors themselves are called “snake oil.”

Polarization, centrism, and partisanship

  • Commenters debate “partisan” vs “independent” vs “centrist,” noting these are not equivalents and that one can be independent but ideologically extreme or centrist yet fiercely loyal to a party.
  • Some criticize “moderate” norms and civility policies as protecting the status quo; others argue some conflicts (e.g., over basic rights) are not amenable to “both-sides” compromise.
  • Several emphasize that online polarization partly reflects real-world, structural conflicts, especially in U.S. politics.

Migration to private / semi-private spaces

  • Many say the real social activity has moved to group texts, iMessage, WhatsApp, Discord, and small private servers, which are invisible to studies of big public platforms.
  • These spaces are seen as closer to old forums or instant messaging, but with problems: poor searchability, “lost media” risk, and less public discoverability.

Decline of “old internet” and loss of value

  • Strong nostalgia for an era when the internet felt exploratory, less monetized, and less politicized.
  • Several argue social media once delivered real value (staying in touch, organizing, niche communities) but has decayed into ads, ragebait, and low-value content.
  • Some still find Facebook groups useful for hobbies or local communities, but feeds are widely described as “dumpster fires.”

Algorithms, monetization, and enshittification

  • A recurring view: advertising and growth incentives are the core drivers of enshittification and polarization, not “human nature” alone.
  • Algorithms are described as optimizing for engagement (often anger), not user happiness; doomscrolling and ragebait are seen as predictable outcomes.
  • Others push back that algorithms mainly reflect aggregate user behavior; society is “in a prison of its own design.”
  • Several link the shift from “social networking” (connecting people) to “social media” (content to consume) to this monetization logic.

Bots, AI, and “slop”

  • Commenters report Twitter/X feeling overrun by bots, fake videos, and engagement manipulation, making it unusable for real-time news.
  • Some fear AI-generated “slop” will accelerate content overload and user fatigue, hastening social media’s decline.
  • There’s concern that personalized AI assistants may become the next vector for subtle opinion-shaping and polarization.

Youth attitudes and changing norms

  • Multiple anecdotes from parents and instructors: many teens and college students view public social media as toxic and prefer private group chats.
  • Compared to the Facebook-everyone era, there is no longer a single “default” platform for college life.
  • Some compare social media’s reputation trajectory to smoking: ubiquitous in one generation, seen as unhealthy and uncool in the next.

Broader societal and political reflections

  • Some argue social media mainly amplifies existing economic and political grievances; others think it increasingly shapes them through feedback loops with politicians and media.
  • Views diverge on root cause: social media design vs economic austerity/inequality vs human tendencies to seek low-effort, emotionally charged content.
  • A few see the contraction of public platforms and return to smaller, ephemeral spaces as healthy; others worry about loss of searchable, durable communal knowledge.