Hacker News Confidence

Less annoying hacker news. (What is this?)

This site scrapes Hacker News once every 30 minutes, then sorts according to this formula, with upvotes as positive signals and comments as negative ones. In my experience, the best articles to read on HN are the ones with a high upvote-to-discussion ratio. Mostly because controversial pieces tend to produce a disproportionate number of comments compared to upvotes.


  1. A Multimodal Automated Interpretability Agent 55 upvotes, 3 comments
  2. Anyone can access deleted and private repository data on GitHub 871 upvotes, 175 comments
  3. A Multimodal Dataset with One Trillion Tokens 99 upvotes, 16 comments
  4. When eyesight fades and climbing provides comfort 14 upvotes, 0 comments
  5. InteractiVenn – Interactive Venn Diagrams 67 upvotes, 10 comments
  6. Bubble Sort: An archaeological algorithmic analysis (2003) [pdf] 37 upvotes, 6 comments
  7. The Puzzle of How Large-Scale Order Emerges in Complex Systems 19 upvotes, 2 comments
  8. Dungeons and Dragons taught me how to write alt text 81 upvotes, 22 comments
  9. Rediscovering Transaction Processing from History and First Principles 46 upvotes, 11 comments
  10. Solving the out-of-context chunk problem for RAG 192 upvotes, 68 comments
  11. Show HN: NoteTech – Create personal automations by writing notes 20 upvotes, 3 comments
  12. Biological Circuit Design by Caltech 67 upvotes, 20 comments
  13. Space-filling curves, constructively 42 upvotes, 11 comments
  14. The Origin of Emacs in 1976 207 upvotes, 89 comments
  15. You got a null result. Will anyone publish it? 214 upvotes, 114 comments
  16. Britain is running out of babies 9 upvotes, 1 comments
  17. Hiding Linux Processes with Bind Mounts 73 upvotes, 33 comments
  18. Pnut: A C to POSIX shell compiler you can trust 165 upvotes, 102 comments
  19. Taking my diabetes treatment into my own hands 374 upvotes, 266 comments
  20. Large Enough – Mistral AI 543 upvotes, 427 comments
  21. Phish-friendly domain registry ".top" put on notice 130 upvotes, 89 comments
  22. Ask HN: Am I crazy or is Android development awful? 32 upvotes, 17 comments
  23. AI models collapse when trained on recursively generated data 188 upvotes, 150 comments
  24. "Doors" in Solaris: Lightweight RPC Using File Descriptors (1996) 110 upvotes, 86 comments
  25. Preliminary Post Incident Review 117 upvotes, 136 comments
  26. Physicists may now have a way to make element 120 101 upvotes, 117 comments
  27. How Olympics officials try to catch “motor doping” 221 upvotes, 453 comments
  28. Alexa is in millions of households and Amazon is losing billions 196 upvotes, 413 comments
  29. Why Discover is no American Express 80 upvotes, 157 comments
  30. Join YC's Ruby Meetup on August 13th at Our New SF Headquarters 0 upvotes, 0 comments