Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Hacking OpenAI(hacktron.ai ↗)
    47comments
  2. Waymo in Singapore(waymo.com ↗)
    22comments
  3. Astra for Law(openai.com ↗)
    450comments
  4. Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint(prismml.com ↗)
    108comments
  5. Bend – A language that blocks AI mistakes via proof, on CPU and GPU(bend-lang.com ↗)
    191comments
  6. Pre-Greek: The lost language hidden within Ancient Greek(linguisticdiscovery.com ↗)
    5comments
  7. Hister: A private search engine for the pages you visit and the files you keep(github.com/asciimoo ↗)
    142comments
  8. Qwen 3.8 Omni Flash(qwen.ai ↗)
    25comments
  9. Wax motor(wikipedia.org ↗)
    59comments
  10. Shapelearn Qwen 3.8 27B (13.1 GB VRAM)(byteshape.com ↗)
    2comments
  11. Fujitsu launches made-in-Japan next-generation CPU FUJITSU-MONAKA(global.fujitsu ↗)
    207comments
  12. Apple detectives solved mystery of ancient tree and rewrote the history of fruit(scientificamerican.com ↗)
    1comments
  13. The Scourge of x86 Emulation(fex-emu.com ↗)
    discuss
  14. Telstra outage: The night a network decided the year was 2006(netnod.se ↗)
    12comments
  15. Ask A Monk – A digital wilderness for thoughts with no immediate answer(askamonk.online ↗)
    12comments
  16. How to Write with an LLM(sockpuppet.org ↗)
    62comments
  17. Flet 1.0 – Build cross-platform apps in Python(flet.dev ↗)
    39comments
  18. Code Scans(devin.ai ↗)
    3comments
  19. Diplodocus, Long Thought Exclusively American, Turns Up in Spain(sci.news ↗)
    28comments
  20. The most important product decision is what you don't build(liamnugent.me ↗)
    25comments
  21. How Uber Protects Against Retry Storms(uber.com ↗)
    33comments
  22. Khipu (Quipu) Field Guide(khipufieldguide.com ↗)
    discuss
  23. CrowdSec Source Code Leak(crowdsec.net ↗)
    42comments
  24. Why I didn’t sign the Fields medallists’ letter(gowers.wordpress.com ↗)
    335comments
  25. How do we prevent mathemathics from devolving into the Medieval Era of secrecy?(mathoverflow.net ↗)
    79comments
  26. Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data(arxiv.org ↗)
    38comments
  27. I Put Nam A2-Lite Inside an iRig HD X(playtaurus.com ↗)
    6comments
  28. Show HN: Snapdrop: Instantly share files between devices. No setup, no signup(snapdrop.me ↗)
    29comments
  29. The American Religion of Self-Storage Facilities(newyorker.com ↗)
    365comments
  30. Rate limits on GitLab.com are changing(about.gitlab.com ↗)
    113comments

Autopilot: Workload Autoscaling at Google

29 pointsby 6y agodl.acm.org
8 comments
6y agoHN ↗

Autopilot is still used today by first party Microsoft teams. Even some azure teams are using AP now.

6y agoHN ↗

yes. last I heard it's alive and kicking.

6y agoHN ↗

Have you considered the possibility that internally Google's Autopilot was available even before Microsoft's paper?

6y agoHN ↗

It's possible but I highly doubt it. We're talking 10+ years (2007) since Microsoft made it public via the research paper.

6y agoHN ↗

Autopilot is the name of the vertical scaling system inside Google used for borg. It's been there as long as I have, which is 2012.

I imagine the team that named it didn't need to read a research paper from MSFT to understand the benefits of dynamically scaling machine resources to match load.

6y agoHN ↗

I just skimmed the abstract and had an impression that this looks more similar to Borg, rather than Google's autopilot which solves a very particular problem of machine resource utilization within Borg. And yes, Borg has been there more than 15 years.