Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Hacking OpenAI(hacktron.ai ↗)
    122comments
  2. Jemalloc 5.4.0(github.com/jemalloc ↗)
    17comments
  3. Astra for Law(openai.com ↗)
    488comments
  4. The scourge of x86 emulation(fex-emu.com ↗)
    8comments
  5. Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint(prismml.com ↗)
    118comments
  6. Bend – A language that blocks AI mistakes via proof, on CPU and GPU(bend-lang.com ↗)
    205comments
  7. Qwen 3.8 Omni Flash(qwen.ai ↗)
    59comments
  8. Pre-Greek: The lost language hidden within Ancient Greek(linguisticdiscovery.com ↗)
    24comments
  9. Hister: A private search engine for the pages you visit and the files you keep(github.com/asciimoo ↗)
    158comments
  10. Wax motor(wikipedia.org ↗)
    63comments
  11. Fujitsu launches made-in-Japan next-generation CPU FUJITSU-MONAKA(global.fujitsu ↗)
    220comments
  12. Shapelearn Qwen 3.8 27B (13.1 GB VRAM)(byteshape.com ↗)
    5comments
  13. When the fractional part of a float fixes your shader(crocidb.com ↗)
    discuss
  14. How to Write with an LLM(sockpuppet.org ↗)
    81comments
  15. Telstra outage: The night a network decided the year was 2006(netnod.se ↗)
    21comments
  16. Flet 1.0 – Build cross-platform apps in Python(flet.dev ↗)
    51comments
  17. Apple detectives solved mystery of ancient tree and rewrote the history of fruit(scientificamerican.com ↗)
    4comments
  18. Ask A Monk – A digital wilderness for thoughts with no immediate answer(askamonk.online ↗)
    20comments
  19. Diplodocus, Long Thought Exclusively American, Turns Up in Spain(sci.news ↗)
    36comments
  20. The most important product decision is what you don't build(liamnugent.me ↗)
    32comments
  21. Fixing an NZXT Signal 4K30 part 2: the green/pink video bug(downtowndougbrown.com ↗)
    8comments
  22. CrowdSec Source Code Leak(crowdsec.net ↗)
    44comments
  23. Waymo in Singapore(waymo.com ↗)
    106comments
  24. How do we prevent mathemathics from devolving into the Medieval Era of secrecy?(mathoverflow.net ↗)
    94comments
  25. Why I didn’t sign the Fields medallists’ letter(gowers.wordpress.com ↗)
    353comments
  26. How Uber Protects Against Retry Storms(uber.com ↗)
    35comments
  27. Show HN: Snapdrop: Instantly share files between devices. No setup, no signup(snapdrop.me ↗)
    34comments
  28. Khipu (Quipu) Field Guide(khipufieldguide.com ↗)
    discuss
  29. Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data(arxiv.org ↗)
    39comments
  30. Code Scans(devin.ai ↗)
    4comments

Ask HN: Two unlabeled downvote buttons rather than one

4 pointsby 5y ago
2 comments
Hear me out. Regardless of the rules, everyone has their own personal reasons for downvoting a post. With a single bit, there's no way to discover the hidden rules that people use to express their reaction to a post.

Labeling the downvote buttons (e.g. "makes me feel sad|bad|mad") has been suggested in the past, but we all know that the expressed semantics of the buttons will be ignored by the community.

Suppose, however, that HN adds one additional unlabeled downvote button.

Faced with two unlabelled downvote buttons, would you click both? When would you click one but not the other? Would one turn into the quality downvote while the other turned into the feelings downvote?

5y agoHN ↗

Would one turn into the quality downvote while the other turned into the feelings downvote?

No, they wouldn't. The problem is that if considered intrapersonal, different buttons could get different qualities, but when considered interpersonal they probably wouldn't.

There are different possible ways to fill multiple downvote buttons with a sense. For example, it could be "feelings downvote" and "quality downvote". Or "quality downvote" and "feelings downvote" from left to right. Or it could be "slight downvote" and "strong downvote" and "slight+strong downvote" which is even stronger. Different people will use buttons differently, so in average it would be the same mess. Probably if a PhD dug into the raw data, she could make sense from it by categorizing downvoters and summing up their votes with normalizing coefficients, but I'm not sure.

Lichess recently changed their puzzles and forced me to vote for each puzzle. I really do not know which puzzle is good and which is not. I struggled to find a way to vote, and finally came to a simple strategy "upvote each puzzle", because I do not know any better. I'm sorry I did it, I wish to help lichess, but do not know how to. When faced with two downvote buttons I probably would struggle to make sense of it, then I'd find some simple strategy which I could verbalize and stick to it. It would be improbable, that the strategy would be "feelings" vs "quality", because how could I separate my feelings from a perceived quality?

5y agoHN ↗

It just doesn't matter that much. No one (well, save literally a single person) can even see vote totals. Votes do one thing—make comments more or less prominent within a thread.

I see voting as a sort of collective curation mechanism, and I vote accordingly. If I see a insightful comment that I think more people should read, I upvote it. If I see a dumb comment I think less people should read, I downvote it. Assuming others agree with me, the insightful comments end up higher in the thread, and the stupid ones become faded and easy to ignore.