Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. GPT-6 Sol and Luna(openai.com)
    332comments
  2. Claude Opus 5.5(anthropic.com)
    595comments
  3. OpenAI GPT–6 Astra breaks Enigma message that has resisted solution since 2005(cryptocellar.org)
    330comments
  4. Claude Opus 5.5 Intelligence, Performance and Price Analysis (Max)(artificialanalysis.ai)
    43comments
  5. WordPress: Unauthenticated path traversal leading to conditional RCE(github.com/wordpress)
    43comments
  6. Pentagon: Palantir AI Overreliance Led to Strike Killing 123 Iranian Children(gizmodo.com)
    43comments
  7. Unreal Agent(unreallabs.ai)
    11comments
  8. There's a high chance of devices being sold with GrapheneOS preinstalled in 2027(grapheneos.social)
    69comments
  9. OpenAI is well positioned to fast-follow Jev(arcturus-labs.com)
    156comments
  10. SAML: A Fractal of Bad Design(trailofbits.com)
    1comments
  11. 16-bit Intel 8088 chip (c. 1985)(allpoetry.com)
    10comments
  12. Launch HN: Coverage Cat (YC S22) – Umbrella insurance via your personal agent(coveragecat.com)
    20comments
  13. Writing Rust code that's fast by asking agents to make the code faster(minimaxir.com)
    38comments
  14. George Lucas Returns to Earth, Bearing Gifts(commonedge.org)
    9comments
  15. Apple has added persistent 'ads' to iOS, and it's driving users crazy(techradar.com)
    358comments
  16. Solitaire Alone Together(solitairealonetogether.com)
    25comments
  17. Show HN: Drop – A rootless Linux sandbox with gVisor support(droprun.sh)
    43comments
  18. Show HN: AI·rete·RAG – a Rete rule engine decides, RAG explains why(ai-rete-rag.com)
    2comments
  19. Can gzip be a language model?(nathan.rs)
    135comments
  20. A study of sequence weighting at scale(janestreet.com)
    discuss
  21. One Minute Park(oneminutepark.tv)
    5comments
  22. AMD's random number generator can't generate a 0?(flatassembler.net)
    169comments
  23. Explaining to business people why building software is still hard(manager.dev)
    1comments
  24. Truman World(trumanworld.live)
    31comments
  25. The Economics of Open-Weight Inference(ornn.com)
    25comments
  26. Relativistic raytracing(publish.obsidian.md)
    5comments
  27. Show HN: InstinctFlash – Run 5B world-action models in real time on Jetson Thor(github.com/general-instinct)
    2comments
  28. Training a model to identify AI-generated web content from structure alone(arxiv.org)
    2comments
  29. MUNI Heritage Weekend in San Francisco(lawrence.lu)
    28comments
  30. Side-stepping the Secretary Problem, unwittingly(evalapply.org)
    2comments

Arch Linux propose changing compression method from xz to zstd

18 pointsby 7y agolists.archlinux.org
3 comments
7y agoHN ↗

"No-brainer" depends on who you ask. It seems like a good idea and the tech seems solid... but Zstandard is made by a guy at Facebook and seems to be owned by Facebook as well, since it's hosted on their github: https://github.com/facebook/zstd

On the other hand, according to https://facebook.github.io/zstd it seems that zstd is part of the Linux kernel now, so if you're all-in with Linux then it can't really be any worse. Personally the Facebook angle doesn't bother me all that much because privacy has been completely lost now (and anyone who honestly thinks they can defeat the might of all these corporations and governments is deluding themselves), so I'm all for the change. The code is open after all.

7y agoHN ↗

The increase in package size would worry me. The download and storage size of packages is nearly always more important than the decompression speed, since the package can be decompressed in parallel with the download.

Surely xz could be modified to produce the same bitstream on single core machines?