Show HN: Filter out engagement bait and politics on your X/Twitter feed

Original Article ↗ Hacker News Discussion ↗

Overview of the tool and reactions

  • Browser extension uses an LLM (via Groq) to hide engagement bait and political content from X/Twitter feeds, mainly “For You”.
  • Several commenters praise the concept and speed, and see it as an early example of AI-powered personal feed curation.
  • Some want more granular controls (e.g., filter all posts about specific public figures, or certain topics) and real-time “mood” sliders.
  • Others note that even humans struggle to define what counts as “bait,” so perfect automated classification is unrealistic.

Existing platform controls and alternatives

  • Many argue that turning off “For You,” using only the “Following” tab, disabling retweets, and using lists already remove most low-quality content.
  • Third-party extensions (e.g., “control panel” tools) are mentioned as effective for cleaning feeds without AI.
  • Some use lists like an RSS reader, organized by topic, to minimize algorithmic influence.

Debate on staying vs leaving X/Twitter

  • A vocal group says the healthiest solution is to leave entirely, deactivate accounts, and switch to RSS, blogs, Mastodon, Bluesky, etc.
  • Others stay for network effects, real-time news, expert communities, or to share projects, while acknowledging rising toxicity and owner-driven enshittification.
  • Some see X as no worse than legacy media and valuable if one can filter well; others highlight overconfidence in personal “discernment.”

Algorithmic feeds, propaganda, and truthfulness

  • Several note that algorithms amplify outrage, politics, and war propaganda; balanced, nuanced content gets little traction.
  • There is concern about misinformation around conflicts, and skepticism that community fact-checking mechanisms can work in polarized situations.
  • Some argue platforms and their incentives—not just users—drive the most toxic patterns.

AI/LLMs as filters: promise and concerns

  • Commenters see a broader future where AI filters overwhelming information streams, including social feeds, local news, and long videos.
  • Others criticize this as wasteful: AI will both generate low-value content and then be used to summarize/filter it.
  • Doubts are raised about LLMs inferring intent (e.g., whether something is deliberately engagement bait) and about reliance on remote APIs vs. local models.