Bouncer: Block "crypto", "rage politics", and more from your X feed using AI

Overall value of X vs. walking away

• Many argue the healthiest solution is to quit X altogether; some say they already did and don’t miss it.
• Others stay because of strong network effects, niche communities (e.g., VTubers, anime, technical circles), and X’s role as a real‑time news source.
• Some describe curate-or-leave tradeoffs: even with a well‑tuned feed, a small amount of upsetting content can ruin the experience.

Filtering approaches: AI vs simple tools

• The project is welcomed by people who want hobby/tech-only feeds and to remove “crypto” and “rage politics.”
• Skeptics say regex, word mute lists, and curated blocklists get you “90% there” without LLM cost, latency, or privacy concerns.
• X’s built‑in mute words, keyword/emoji blocking, and “Following-only + reverse chronological” are frequently cited as effective.
• Some propose local models plus userscripts (e.g., Greasemonkey) to classify and hide content client-side, avoiding external APIs.

Risks of filtering “rage politics”

• Several worry that filtering on vague labels like “rage politics” will produce misinforming bubbles: users see reactions but not causes, or underestimate genuine public anger.
• Others say being misinformed or underexposed to outrage on X is acceptable or even desirable for mental health.
• It’s noted that social media outrage is often driven by the “terminally online,” not necessarily representative of populations.

Algorithmic feeds, curation, and addiction

• Many blame algorithmic “For You” feeds for surfacing ragebait; they prefer Following-only feeds or other platforms with simple chronological timelines.
• Some insist that if you keep seeing topics you dislike, it’s partly due to your engagement; “block, scroll past, and move on” is advocated.
• Analogies are made to addiction (e.g., smoking): people may need tools to keep using X even if quitting might be healthier.

Broader applicability and moderation debates

• Multiple commenters want similar AI-based filters for Reddit and Hacker News to hide recurring political or low‑quality topics.
• There is support for per-user filtering instead of global flagging, which currently shapes what everyone sees.
• Complaints about Reddit’s and X’s moderation, bots, and influence operations lead some to favor personal filtering over platform governance.