Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Pruning LLMs Like a Physicist: Block Removal as an Ising Optimization Problem(huggingface.co)
    discuss
  2. Continual learning on a 28B Learner 1.0 model, new architecture(learnerlabs.ai)
    1comments
  3. What's going on with millennium maths problems [video](youtube.com)
    discuss
  4. Ralph(danluu.com)
    discuss
  5. Porting Doom's game logic and renderer to SQL(cedardb.com)
    discuss
  6. Tech elites launch college alternative, raises $42M(nypost.com)
    discuss
  7. Rpi – a composable LLM agent SDK in Rust(github.com/bigfish1913)
    discuss
  8. Editors of higher education journal coauthor editorial with fake references(retractionwatch.com)
    discuss
  9. God Help Us, Let's Try to Learn About Mechanistic Interpretability Techniques(astralcodexten.com)
    discuss
  10. Vigil: Keeps your Mac awake while your coding agents work(vigilmac.com)
    discuss
  11. Failures shown to be the harness, not the model (blog post)(herlein.com)
    discuss
  12. Eiderdown: A Family Tradition – Trailer [video](youtube.com)
    discuss
  13. Open source tries to govern AI code(freenode.net)
    discuss
  14. Explore page for early internet websites(straw.page)
    discuss
  15. Managers Are Catalysts(laughingmeme.org)
    discuss
  16. Show HN: An API that tells you what people love about a product(bb-product-api-docs.web.app)
    discuss
  17. Tokens Too Cheap to Meter(jyn.dev)
    discuss
  18. I deleted slow, expensive LLM turns in coding agents with Jev(tawfeks.github.io)
    discuss
  19. Should I use Cloudflare for my SaaS, its so annoying
    discuss
  20. M&Ms and Skittles sorting machine(willempennings.nl)
    discuss
  21. PayPal partners with Muse for agentic checkout(twitter.com/alexandr_wang)
    discuss
  22. OpenCode Reloaded(anoma.ly)
    discuss
  23. Binomial Coefficient Coincidences(johncarlosbaez.wordpress.com)
    discuss
  24. Show HN: Beachball – A Beachball that lives in your macOS Dock(beachball.click)
    discuss
  25. Creator of Windows Task Manager has vibe coded a version that runs on Mac/Linux(engadget.com)
    discuss
  26. I Asked 100 Agents to Hack Me(sshh.io)
    discuss
  27. Writing Rust code that's fast by asking agents to make the code faster(minimaxir.com)
    discuss
  28. How much of the "AI is going rogue" talks is just a publicity stunt?
    2comments
  29. How to Get from AI-Assisted to AI Native(oreilly.com)
    discuss
  30. Open-Source Business-as-Code Platform to Orchestrate Agents(github.com/littlehorse-enterprises)
    discuss

Ask HN: Is AI use not allowed in projects which is shared here to show?

2 pointsby 56m ago
6 comments
Dear members, I saw someone show HN project got flagged and people were commenting it is probably because it looked AI slop. I'm curious, is AI use not allowed in projects which is shown here ? I think, everyone is using AI in projects these days, isn't it ?
47m agoHN ↗

Why would I spend my time reviewing something that the “author” didn’t even spend time to create? I think there is a huge difference between AI assisted and AI slop

31m agoHN ↗

Curious when slop becomes, not slop. Revenue? Profit? At what point do detractors throw in the towel? Honest question.

For example DHH shipped an entire vibe coded Linux distribution. Should we all roll of eyes at that?

Weird times.

16m agoHN ↗

I think it becomes not slop when it is useful or unique. The projects that just lazily reimplement something that has been done 10 times before are generally the least impressive

30m agoHN ↗

Projects using Ai are fine

What's not is (a) submitting low quality Ai output, stay in the loop (b) using Ai to post or comment on HN, no bots

16m agoHN ↗

The purpose of Show HN is to "show something you've made that other people can play with. HN users can try it out, give you feedback, and ask questions in the thread" (https://news.ycombinator.com/showhn.html)

That isn't possible with many AI generated projects because OP didn't write the code, probably doesn't understand the code and can't discuss the code or their process in creating it (because said process doesn't exist.) There is little if anything of substance to discuss or play with to satisfy intellectual curiosity, and in many cases the projects themselves wouldn't pass muster if written by a human, they're just vibe-coded slop.

There have been exceptions which have done well (ex:https://news.ycombinator.com/item?id=49550375) but in that case AI was used in a specific technical context by someone knowledgeable who definitely wasn't "vibe coding" anything. But again - if a human had done all of that work, it would still have been a worthwhile post.

Show HN can't simply be a place where people post slop and everyone claps at how cool the latest models are. It can't simply be advertising for LLMs.