Project ideas from Hacker News discussions.

The revolt of the reader

📝 Discussion Summary (Click to expand)

1. AI‑generated “slop” is everywhere and irritating readers
Many commenters complain that low‑quality, LLM‑written content floods Hacker News and makes it hard to find worthwhile material.
- “AI slop really needs to be auto‑flagged. It’s been like 20% of front page links I click on lately.”jonstewart
- “A lot of similar pieces have not considered a post is both the first and final work of a thing: opinions, and experience, and less of much in cited facts; with AI, that there even was a revision pass at all.”turtleyacht

2. Detection tools (e.g., Pangram) are used but their reliability is contested
Users rely on AI‑detectors to flag suspect posts, yet they worry about false positives/negatives and the tools’ sensitivity.
- “I liked your piece, and agree with almost all of it, but I'm surprised by your faith in the accuracy of Pangram at detecting AI writing.”nkurz
- “My experience is using Pangram quite often with lots of writing of all flavors (including a bunch of known origin). … over 176 posts spanning 22 years, all 176 … are 100% human.”bcantrill

3. AI undermines the human writing process and the value of authentic effort
Several writers argue that relying on LLMs bypasses the cognitive work of writing, breaking the writer‑reader contract and degrading personal understanding.
- “to use an LLM to write is to void the social contract between writer and reader: we readers shouldn’t be expected to labor to understand a sentence that the writer themselves didn’t work to create.”forgeties79
- “using an LLM is robbing you of the process of writing, a process that is crucial to developing and understanding your own ideas.”ofjcihen


🚀 Project Ideas

Generating project ideas…

HN AI Scan – Browser Extension for Real‑Time AI‑Generated Content Labeling

Summary

  • A browser extension that adds AI‑likelihood scores (using Pangram and optional open‑source detectors) to each Hacker News post and comment, letting users quickly filter out low‑quality AI slop.
  • Core value: saves readers time by surfacing likely AI‑generated content without manual inspection, while keeping detection costs low via local caching and optional BYOK for Pangram.

Details

Key Value
Target Audience Hacker News readers who want to curate their feed and avoid AI‑written noise
Core Feature Inline badge showing AI probability (0‑100%) on each HN item, with hover tooltip explaining the score
Tech Stack JavaScript/TypeScript, WebExtensions API, optional Rust/Wasm for local detection, IndexedDB for caching
Difficulty Medium
Monetization Hobby

Notes

  • HN commenters explicitly asked for “a browser extension to label HN posts with Pangram results” (duhhhhh1212) and wanted a way to “scan all words … and colour them more and more transparent as the likelihood of llm prose is increased” (pavo-etc).
  • Provides a practical utility that can be discussed in future HN threads about content quality and could inspire similar tools for other forums.

OpenDetect – Self‑Hosted AI Text Detection API Using Open‑Source Models

Summary

  • An open‑source, self‑hostable micro‑service that exposes a simple HTTP API for detecting AI‑generated text using models like DetectGPT or fine‑tuned classifiers, reducing reliance on paid third‑party detectors.
  • Core value: gives developers and communities a cheap, privacy‑preserving way to integrate AI detection into extensions, moderation bots, or editorial workflows.

Details

Key Value
Target Audience Developers, moderators, and community builders who need affordable AI detection
Core Feature REST endpoint /detect returning AI probability and highlighted spans; supports batch requests
Tech Stack Python (FastAPI), HuggingFace Transformers, ONNX Runtime for CPU inference, Docker
Difficulty Medium
Monetization Revenue-ready: Usage‑based pricing (e.g., $0.0005 per 1K characters) if offered as a hosted service; otherwise Hobby for self‑hosted

Notes

  • Commenters complained about Pangram’s cost (“using Pangram for it ends up being quite expensive” – demibabs) and wanted alternatives like BYOK or cheaper detection.
  • A self‑hosted solution addresses privacy concerns and enables community‑run instances, fostering discussion on detection accuracy and false‑positive trade‑offs.

SlopFlag – Community Moderation Bot for Hacker News that Auto‑Flags Suspected AI‑Generated Posts

Summary

  • A moderation bot that watches the HN new‑submissions feed, runs AI‑detection (via OpenDetect or Pangram) on each post/comment, and automatically flags likely AI‑generated content for review by human moderators or the community.
  • Core value: reduces the burden on readers and moderators by surfacing potential slop early, improving overall signal‑to‑noise on the site.

Details

Key Value
Target Audience HN moderators, power users, and community stewards
Core Feature Automatic flagging + comment with detection score; optional auto‑collapse of high‑confidence AI posts
Tech Stack Node.js (or Go) for polling HN Firebase API, Python detection service, PostgreSQL for flag storage, deployed on Fly.io or similar
Difficulty High
Monetization Hobby

Notes

  • Users expressed frustration that “AI comments are auto‑flagged, but not posts” (mitxela) and wished for “AI slop really needs to be auto‑flagged” (jonstewart).
  • The bot could spark meta‑discussions on moderation policies, detection reliability, and the balance between automation and human judgment on HN.

Read Later