Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. I built non-autoregressive decision models with RL a year ago(convaiinnovations.com ↗)
    179comments
  2. AI-generated posters don’t have to be horrible(john.hartnup.uk ↗)
    572comments
  3. Btrfs/ZFS/bcachefs under workloads classic benchmarks skip(bartosz.fenski.pl ↗)
    11comments
  4. Human brain is two separate organs, Stanford Medicine-led research finds(stanford.edu ↗)
    201comments
  5. A graphical desktop for the ZX Spectrum(github.com/mindbox77 ↗)
    88comments
  6. Tin: full-text search for Postgres(planetscale.com ↗)
    57comments
  7. The Secret Life of Circuits(coredump.cx ↗)
    60comments
  8. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    603comments
  9. Black Holes or Black Hole Stars? Astronomers Spar over 'Little Red Dots'(quantamagazine.org ↗)
    27comments
  10. Supabase (YC S20) Is Hiring for OrioleDB(supabase.link ↗)
    discuss
  11. New evidence for hidden chambers beyond Tutankhamun's tomb(nature.com ↗)
    14comments
  12. Suzanne Ciani's Buchla Cookbook(echo.orpheusinstituut.be ↗)
    1comments
  13. Show HN: CUA-S1 – A System One Model for Computer Use(github.com/trycua ↗)
    discuss
  14. GPT-6 Astra Solves a WWI German Radio Cipher(prinzai.com ↗)
    149comments
  15. San Francisco Onion Futures Company(onionfutures.com ↗)
    134comments
  16. Almost Never Use AI to Write Anything Substantive(erichgrunewald.substack.com ↗)
    44comments
  17. Cloudflare Quick Tunnels(cloudflare.com ↗)
    303comments
  18. How to Write with an LLM(sockpuppet.org ↗)
    366comments
  19. If math is more than proof, we need to better celebrate the rest of it(terrytao.wordpress.com ↗)
    214comments
  20. What Zig felt like, coming from Rust(besok.github.io ↗)
    150comments
  21. You can run Git on object storage if you re-make packfiles(tigrisdata.com ↗)
    30comments
  22. Saving another 100TB of RAM(cloudflare.com ↗)
    97comments
  23. How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip(ieee.org ↗)
    124comments
  24. Communication by means of modulated Johnson noise(pnas.org ↗)
    23comments
  25. SDCC – Small Device C Compiler(sourceforge.net ↗)
    27comments
  26. Asking Authors About Their Own Papers(medium.com/tmlrorg ↗)
    51comments
  27. Ray Ozzie and the Optimism of Being Early(reproof.app ↗)
    22comments
  28. Science Is Open Software(jepedersen.dk ↗)
    60comments
  29. The first new cat species discovered in 100 years(nationalgeographic.com ↗)
    130comments
  30. Why building a Rust LSP is hard(rust-glancer.github.io ↗)
    55comments

Google says its Gemini AI model hacked three other companies

5 pointsby 10h agotheguardian.com
4 comments
10h agoHN ↗

"The hacks occurred during a cybersecurity evaluation by AI-security firm Irregular."

Irregular is looking good for the Silicon Valley PR Firm of the Year award.

8h agoHN ↗

"We are also, in fact, cool kids. Please pay attention!"

5h agoHN ↗

“These events highlight the importance of training powerful AI models to act responsibly,” Adkins, the Google spokesperson, said.

No. These events highlight the importance of training powerful "AI" models' operators to act responsibly.

I'm not holding my breath.

2h agoHN ↗

What I really want to know is, did Google go hire Irregular after the other hacks?