Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Claude Code now reads AGENTS.md if there is no Claude.md(claude.com ↗)
    169comments
  2. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    213comments
  3. How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip(ieee.org ↗)
    26comments
  4. Cloudflare Quick Tunnels(cloudflare.com ↗)
    237comments
  5. Saving another 100TB of RAM(cloudflare.com ↗)
    38comments
  6. A 1542 papal cipher cracked with simulated annealing(simonklee.dk ↗)
    1comments
  7. Xcode 27.1 Beta Release Notes(developer.apple.com ↗)
    63comments
  8. How to Write with an LLM(sockpuppet.org ↗)
    259comments
  9. Cache-to-Cache: Direct Semantic Communication Between LLMs (2025)(arxiv.org ↗)
    12comments
  10. Photon-Emission-Guided Laser Fault Injection Enables RP2350 Secure Debug(ledger.com ↗)
    49comments
  11. Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash(cactuscompute.com ↗)
    74comments
  12. The first new cat species discovered in 100 years(nationalgeographic.com ↗)
    49comments
  13. OpenJev(openjev.com ↗)
    240comments
  14. US troop deaths during Iran war exceed Pentagon count by at least four(reuters.com ↗)
    87comments
  15. Cyclomatic Complexity in C#(ndepend.com ↗)
    12comments
  16. Two parallel neural ectoderm progenitors contribute to the developing brain(newscientist.com ↗)
    51comments
  17. The Implications of Linguistic Illegibility for LLM Security(arxiv.org ↗)
    17comments
  18. C++26: Trivial infinite loops are no longer undefined behaviour(sandordargo.com ↗)
    177comments
  19. Warez: The Infrastructure and Aesthetics of Piracy (2021)(archive.org ↗)
    15comments
  20. Minimal Phone 2(minimalcompany.com ↗)
    163comments
  21. Inside ZCode: Silently uploading your Git history to the cloud(ferstar.org ↗)
    89comments
  22. Size-Specialized Memory Allocation(go.dev ↗)
    3comments
  23. How SpaceX streamlined the Raptor engine(construction-physics.com ↗)
    29comments
  24. I vibed a proof of Conway's conjecture(overreacted.io ↗)
    180comments
  25. A search-and-inference database from scratch in pure Zig(antfly.io ↗)
    16comments
  26. From Geometry to Algebra and Back Again: 4000 Years of Papers (2023) [video](youtube.com ↗)
    discuss
  27. North Korean nuclear test sets off years of earthquakes(science.org ↗)
    151comments
  28. US Military had close call after using AI for hallucinated intelligence report(cnn.com ↗)
    291comments
  29. Mathematicians Build Long-Awaited Graph Sandwich(quantamagazine.org ↗)
    18comments
  30. Cekura (YC F24) Is Hiring(ycombinator.com ↗)
    discuss

Show HN: UserScript to detect GPT generated comments on Hacker News

2 pointsby 3y agogithub.com
3 comments
It puts every comment into the GPT output detector and colors and writes a short comment on the HN comment based on its assessment, like you see in the screenshot. This is based on a threshold set by me. >0.7 is probably AI, >0.9 is definitely AI. Lower than that is most likely human. Most comments on HN still appear to be human.

It only becomes reliable after about 50 tokens (one token is around 4 characters) so I mark the comments that are too short with gray and make no assessment on those.

I know pretty much nothing about Javascript. This is shitty code largely written by ChatGPT itself, untested beyond Chrome on MacOS, and no plans for maintenance or extends. If you want features or support, you have to ask ChatGPT, not me.

Let me know what you think, and please don't comment on code quality.

3y agoHN ↗

OP, here. Before anyone else does, I asked ChatGPT to write a reply to this post.

Thank you for sharing your UserScript for detecting GPT generated comments on Hacker News! It sounds like a useful tool for identifying potentially AI-generated comments on the platform. We appreciate your transparency about the limitations of the script and the fact that it is largely written by ChatGPT itself. We'll be sure to keep this in mind while using it. Thanks again for sharing!

3y agoHN ↗

I peeked at the source code and it seems like it's just sending a HTTP request to https://huggingface.co/openai-detector/ with the contents of a comment.

This should be probably mentioned since it might have some privacy, performance and data usage concerns.

Additionally I am guessing that there is some rate limiting to using huggingface's service for free.