Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. How Fast You Age During 40s May Predict Memory Lapses in Your 50s (nautil.us)
    —discuss
  2. Tyndall Effect (wikipedia.org)
    —discuss
  3. Ask HN: Allow agents access to cloud files with least privilege
    —discuss
  4. How JWT Authentication Works (medium.com/koukyosyumei)
    —discuss
  5. US Military disables ad trackers amid concerns over troops' safety (taskandpurpose.com)
    1comments
  6. Getting Banned from NT for Apple Silicon Project for Booting Windows on Apple M6 (twitter.com/acelogic_)
    —discuss
  7. Current ways of developing AI agents contribute to degradation of oversight (arxiv.org)
    —discuss
  8. postmarketOS Rebrand: Nura (nura.eco)
    —discuss
  9. EU urges UK to raise tariffs on Chinese cars to avoid 'made in Europe' barriers (ft.com)
    1comments
  10. DOE releases national quantum computing roadmap (fnal.gov)
    —discuss
  11. The Normalization of Inexplicable Failures (ihatethefuture.com)
    —discuss
  12. I Gave My Life over to Meta's A.I. Agent and Was Blown Away (nytimes.com)
    —discuss
  13. We Outsourced the Verb (yesigiveafig.com)
    —discuss
  14. Show HN: High-performance, zero-dependency charting engine for Svelte (bonguynvan.github.io)
    —discuss
  15. Thoughts on HTML's proposed previewsrc attribute (shkspr.mobi)
    —discuss
  16. Numeric just declared war on the journal entry (karthikkolapalli.substack.com)
    —discuss
  17. What's Up with Monomorphism? (2015) (mrale.ph)
    —discuss
  18. A20 Pro and on-device AI: It's showtime (rickytakkar.com)
    —discuss
  19. Show HN: Updated Ceptile AI Search Improved Answers and Research (ceptile.com)
    —discuss
  20. Is AGENTS.md useful? What the research found (gethrbr.com)
    —discuss
  21. Snowmoon (vitalik.eth.limo)
    —discuss
  22. What Improves Developer Productivity at Google? Code Quality (acm.org)
    —discuss
  23. Show HN: PedalTalk, a hands-free group bike intercom that keeps your music on (play.google.com)
    —discuss
  24. Show HN: Digital Audio Workstation built with stock WebAudio nodes (seven.systems)
    —discuss
  25. Ju! Ju! Tsu! (codeberg.page)
    —discuss
  26. Luarocks.org Security Incident, September 2026 (luarocks.org)
    —discuss
  27. Good News for American Renters (economist.com)
    —discuss
  28. Show HN: The Standup, releases, advisories and outages for developers (thecompound.tech)
    —discuss
  29. Teaching GPU programming in p5.js: now with compute shaders (davepagurek.com)
    —discuss
  30. Text Gag (textgag.com)
    —discuss

Ask HN: How are you getting inference for personal projects?

1 pointsby 1h ago
3 comments
It seems like the major providers are starting to cut back the usage included with their monthly subscriptions. How are people handling inference for side projects these days?

Are you still using Codex/Claude subscriptions, paying for hosted open weight models, running local models, etc

1h agoHN ↗

Claude subscription (the cheapest $20/month one) -- I very rarely hit the limits, particularly the weekly ones. But then again, the largest codebase I ever work on is 50k lines.

I'd prefer local models massively for both work and "pleasure" (I love LLMs as random cognitive sparring partners), but then again, my bank account massively disagrees :/

1h agoHN ↗

I've been on the Claude Max plan for quite a while now but I also run my own local models and every week there are new local models that are slowly but very surely closing the gap to frontier-level quality and so, pretty soon, I won't use APIs at all. Can't wait!

27m agoHN ↗

I have a Proxmox machine with dual P40 gpus that pass thru to a vm with ollama running for local inference. Been working great for over 2 years.