Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. The 'Colossal Headache' of Building a Made-in-America Monopoly Game (wsj.com)
    1comments
  2. Choices – Enterprise System-1 Router (mahabodi.github.io)
    —discuss
  3. Meta's New Muse AI Agent Read My Private Messages. I Never Asked It To (inc.com)
    —discuss
  4. Base31.org -best Site Award (base31.org)
    —discuss
  5. Classify 6700 pages for $1 with a Jev-compatible VLM API (twitter.com/spillai)
    —discuss
  6. Show HN: Talktome – Voice calls with your AI agent (github.com/rohanprichard)
    1comments
  7. Muse gives user's home address without permission, sending buyer to his house (theguardian.com)
    —discuss
  8. The Second Life of Autoconf for MS-DOS (dblock.org)
    —discuss
  9. Firebase reporting crashes due to Firebase (github.com/firebase)
    1comments
  10. Anthropic warns of 'existential risks to humanity' in IPO prospectus (ft.com)
    —discuss
  11. Claude Code's Next Era (latent.space)
    —discuss
  12. The machine layer under Nvidia OpenShell: why containment is not safety (endstop.systems)
    —discuss
  13. Training NanoGPT in 39.9 Seconds (github.com/kellerjordan)
    —discuss
  14. Show HN: Interactive Diesel Price Forecaster (sightspinner.ai)
    —discuss
  15. What I Learned About Asset Pricing (pedrosantaclara.substack.com)
    —discuss
  16. AI agent memory can be poisoned – and later treated as the user's own past (astraobscura.net)
    —discuss
  17. Managing access to catalogs in ScalarDB Analytics (medium.com/scalar-engineering)
    —discuss
  18. Supporting native Rust in Workers with the Emscripten target for WASM-bindgen (cloudflare.com)
    —discuss
  19. Paperclip/MtG inspired pharma strategy game (pharmasim.pages.dev)
    —discuss
  20. Calnode vs. calrs vs. Tymeslot: three self-hosted answers to Calendly (github.com/calnode)
    —discuss
  21. Bluegraph – Explore NOAA buoy data, rebuilt in 3D from measured spectra (bluegraph.io)
    —discuss
  22. Deterministic Testing for Multithreaded Python (lwn.net)
    —discuss
  23. Tell HN: Firebase service causing iOS crashes across many apps
    2comments
  24. Read anything aloud–with a voice that fits (ireadall.ai)
    —discuss
  25. Ask HN: Do you "Enable no talk mode?"
    —discuss
  26. Highlights from Git 2.56 (github.blog)
    —discuss
  27. Singapore's Bet on Chipmaking (ft.com)
    —discuss
  28. OpenAI shelves new AI model after internal safety tests: Report (channelnewsasia.com)
    1comments
  29. How well do you know AI? (reddit.com)
    —discuss
  30. Show HN: Diabetes Risk Analyzer (Flask-based web app) (diabetesriskanalyzer.org)
    1comments

Ask HN: Do you "Enable no talk mode?"

2 pointsby 50m ago
0 comments
After spending years reading mostly human generated content on the internet for years, and treasuring the efforts spent absorbing content, the latest fad of llm's spitting out text like there was no tomorrow has drastically shifted my tolerance for content.

Is anyone else slapping up a "ENABLE NO TALKING MODE" before their prompts to ensure that the response coming back is straight to the point and factual?

Sadly I can't convince claude on the desktop app to follow this rule continuously but it's an easy manual prompt starter..

Any other tricks that may be applied to keep chatter and tokens down?

A quiet thread, for now.Start the conversation on HN ↗