IdeaWell

Project ideas from Hacker News discussions.

X Is Bad, Though

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. Praise for X’s changes
Many users argue that the platform has improved under Musk: the algorithm surfaces more interesting content, censorship has dropped dramatically, and Community Notes provide a useful fact‑checking tool.

“Algorithm is much improved. Just way more interesting content reaches me now.” – howunfortunate
“WAY less censorship. Like an unbelievably reduced amount of censorship.” – howunfortunate
“Community notes are the best 'fact check' type implementation I've yet to see.” – howunfortunate

2. Criticism of X’s decline
Opponents claim that reduced moderation has led to a surge of hateful content, deepfakes, and authoritarian censorship, making the site hostile.

“You could power a datacenter with how fast Orwell is spinning in his grave, as Elon's Xitter approved 83% of censorship requests by authoritarian governments (compared to pre-Elon Twitter's 50%).” – notthemessiah
“Less moderation means people can say slurs, post deepfakes of children, and make the place an extremely hostile environment filled with racists, slanderers, bots, disinformation campaigns, and conspiracy nutjobs.” – notthemessiah

3. Indifference / contextualization (Twitter was already flawed)
A significant contingent notes that Twitter’s problems predate Musk; using X doesn’t meaningfully empower him, and moral posturing over platform choice is misplaced.

“I think it's worth noting that Elon didn't make Twitter bad. Twitter was already bad. Jack found himself accidentally in control of the world's message board…” – shadowgovt
“Me shitposting on Twitter … is not why Elon Musk holds the shitty views he does nor why he has a trillion dollars in net worth.” – wpm

🚀 Project Ideas

FeedGuard: User‑Controlled Content Filter Extension

Summary

  • A browser extension that lets X (Twitter) users define custom filters (keywords, regex, ML‑based toxicity scores) to hide or flag slurs, conspiracy theories, deepfakes, and other unwanted content, with optional community‑shared filter lists.
  • Core value proposition: puts moderation back in the hands of the user, reducing exposure to toxic material without relying on platform‑wide censorship or algorithmic bias.

Details

Key Value
Target Audience Power users of X who are frustrated by slurs, hate speech, deepfakes, and algorithmically amplified rage‑bait (as noted by notethemessiah, braden‑lk, and fivefifty).
Core Feature Configurable filter engine (keyword/regex + optional LLM classifier) that can hide, blur, or label tweets; supports import/export of filter lists and community‑submitted “safe‑list” feeds.
Tech Stack JavaScript/TypeScript, WebExtensions API, optional TensorFlow.js or ONNX Runtime for client‑side LLM scoring, IndexedDB for list storage.
Difficulty Medium
Monetization Hobby

Notes

  • HN commenters asked for ways to “block the blue” userscripts and to filter out “unhinged conspiracy theories” (notethemessiah, fivefifty). FeedGuard directly addresses those requests.
  • By enabling users to share and subscribe to community‑curated filter lists (similar to Community Notes but user‑driven), it creates a practical utility that could be discussed and improved openly on HN.

OmniFeed: Cross‑Platform Social Aggregator with Pluggable Algorithms

Summary

  • A desktop/web app that pulls public posts from X, Bluesky, Mastodon, and other federated networks into a single unified feed, letting users choose or mix ranking algorithms (chronological, relevance, LLM‑curated, custom).
  • Core value proposition: gives users control over what they see and where their attention goes, alleviating the need to stay on a single platform while still accessing interesting voices that have migrated elsewhere.

Details

Key Value
Target Audience Users who miss interesting voices on X but dislike supporting its owner (wpm, howunfortunate, ayaros) and those who want a neutral, algorithm‑transparent feed (ripley12, bashtoni).
Core Feature Unified feed connector with plug‑in ranking modules (e.g., chronological, engagement‑based, LLM‑topic filter) and per‑source moderation rules (hide replies, boost community notes).
Tech Stack React/Electron frontend, Node.js backend, ActivityPub and AT‑Protocol clients, optional HuggingFace Transformers for LLM ranking, SQLite for local caching.
Difficulty High
Monetization Revenue‑ready: Freemium (free basic aggregation; $5/mo for advanced algorithm plug‑ins and ad‑free experience).

Notes

  • Commenters lamented that “the interesting ones are figuring out what a lame idea it was to leave” (wpm) and praised Bluesky’s openness for custom feeds (ripley12). OmniFeed offers the best of both worlds.
  • The ability to swap algorithms on the fly would spark HN discussion about feed quality, echo chambers, and the trade‑offs between engagement‑driven vs. chronological feeds.

AdOffset: Transparent Ad Revenue Tracker & Offset Service

Summary

  • A browser extension that estimates the ad revenue generated by a user’s X usage (based on impressions tracked) and offers to offset that amount via donations to privacy‑focused or open‑social‑media projects, or to subscribe to an ad‑free tier that routes fees to those causes.
  • Core value proposition: lets users who block ads still understand and neutralize the financial support they inadvertently give to the platform, aligning usage with personal ethics.

Details

Key Value
Target Audience X users who block ads but feel uncomfortable indirectly funding Elon Musk (bashtoni, wpm, ShinyLeftPad).
Core Feature Real‑time impression counter, revenue estimate model (CPM‑based), one‑click donation to preselected EFF/OSMF‑style funds, optional premium ad‑free subscription that forwards fees to those funds.
Tech Stack JavaScript/TypeScript WebExtension, background service worker for network request sniffing, simple API to donation platforms (Stripe Connect or PayPal), local storage for stats.
Difficulty Low
Monetization Revenue‑ready: 15 % fee on processed donations (covers operation) or “Hobby” if donation‑only.

Notes

  • bashtoni pointed out that “by visiting X you are a product he's selling to advertisers” and wpm questioned how much value their visits bring; AdOffset makes that implicit revenue explicit.
  • Providing a clear way to offset or eliminate that financial link would resonate with the HN audience’s desire to avoid indirectly supporting objectionable owners while still participating in public discourse.

Read Later