Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. LensVLM: Compressing long context as images, expanding only relevant pages(huggingface.co)
    discuss
  2. Ask HN: Do you offload your coding to AI, or keep your skills sharp?
    discuss
  3. Can internal model transparency tame the AI race?(karthiktadepalli.com)
    discuss
  4. The meta-harness for coding agents(soloterm.com)
    discuss
  5. Show HN: Jev-mice – mouse colony simulation using Jev and deterministic engine(carsonsweet.com)
    discuss
  6. Discover Solo: The Ultimate AI-Integrated Control Panel(master.dev)
    discuss
  7. Steam Frame Teardown(ifixit.com)
    discuss
  8. SpaceX Starshield(wikipedia.org)
    discuss
  9. Mexican Navy's tall ship Cuauhtémoc will be at the Embarcadero all week(kqed.org)
    discuss
  10. A Worthwhile Trade(nejm.org)
    discuss
  11. Loopjacking in A2A Implementations: Hijacking Human-in-the-Loop Approvals(adithyanak.com)
    discuss
  12. How AI Helps and Hampers Junior Workers(nominalnews.com)
    discuss
  13. Jackrong/Qwopus3.8-27B-Flash-V2(huggingface.co)
    discuss
  14. Nemotron 3 Diarization(huggingface.co)
    discuss
  15. CKKS – Encryption and Decryption(jeremykun.com)
    discuss
  16. What AI pays the people who train it(dpdns.org)
    discuss
  17. Cloud Agents Are Inevitable AI Prisons(normanponte.io)
    discuss
  18. Ask HN: Why so many dead comments lately?
    discuss
  19. Illinois, California governors fast-track AI safety measures with EOs(transparencycoalition.ai)
    discuss
  20. Gvnr(gvnr.io)
    discuss
  21. HySparse2: Hybrid Sparse Attention with Two-Level KV Sharing(arxiv.org)
    discuss
  22. The Vigilante Who's Taking Down Recruitment Scammers, One Fake Job Ad at a Time(wsj.com)
    1comments
  23. Pretraining data, not verifiability, is why LLMs are good at math (and coding)(lesswrong.com)
    discuss
  24. Microsoft Publisher will no longer be supported after October 2026(support.microsoft.com)
    1comments
  25. When AI makes challenging last-minute requests no human client cares about(twitter.com/lauren_wilford)
    discuss
  26. Facing Congressional Scrutiny, Flock Details New Search Guardrails(wsj.com)
    discuss
  27. A quadruped robot designed to complete a marathon on a single battery charge(nature.com)
    discuss
  28. Swap, ZRAM, Zswap and Hibernate on NixOS(matthewbrunelle.com)
    discuss
  29. Claude discovers a novel enzyme system with CRISPR-like repeats(anthropic.com)
    discuss
  30. Introducing HLE-Diamond – Humanity's Last Exam(lastexam.ai)
    discuss

The web being slow is a choice (and it's a stupid choice)

1 pointsby 1h agoalbanbrooke.com
2 comments
1h agoHN ↗

Choice, or incompetence? Or maybe structural incompetence: those who decide what's being built only care about maximising profit, those who know how to build stopped caring because... why would they?

When whoever is in power finds it fine to build such terrible products full of ads, the only thing worth doing is the minimal amount of work required to get your salary, if you ask me.

1h agoHN ↗

U.S. download speeds increased 174% from 2020 to 2024. 5 Gbps is almost 90,000x the bandwidth I had in 1996. So why are we’re still dealing with slow, bloated websites?

Yes, because the web is more bloated now than it ever was, but also because you're conflating latency and throughput. If the server takes multiple seconds to send the first bit of data, no amount of bandwidth will ever get you within the 0.1-second deadline.

Conflating latency and throughput is a surprisingly common error, and at my last job processing sensor data in real time, I had to make a huge effort to convince other people that latency was the performance metric that mattered, much more than throughput.