Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Context-Driven Monitoring (simpleobservability.com)
    —discuss
  2. Vanilla extract: Zero-runtime Stylesheets in TypeScript (vanilla-extract.style)
    —discuss
  3. Show HN: Revise, a new document editor library for React (npmjs.com)
    —discuss
  4. Norwegian consumer council explains enshittification [video] (vimeo.com)
    1comments
  5. Show HN: Linktree is ugly so I built this (outie.link)
    —discuss
  6. Show HN: Interactive Phaistos Disc workbench, rotation ergonomics, null models (lessthanzero.github.io)
    —discuss
  7. Why mathematicians are hiding research results in zero-knowledge proofs (scientificamerican.com)
    —discuss
  8. A wall that listens: the local-LLM pipeline behind an AI party in Vilnius (vania-novikau.me)
    —discuss
  9. A Review of Don Quixote (flowtwo.io)
    —discuss
  10. Vibe Coding Developer Productivity Tools (mikemcquaid.com)
    1comments
  11. Why do humans have an innate need to escape the surface of the Earth? (onverticality.com)
    —discuss
  12. Citizens' Assemblies Can Help Repair Dysfunctional Democracies (2022) (noemamag.com)
    —discuss
  13. Understanding the Four AI Risk Domains (sdarchitect.blog)
    —discuss
  14. Show HN: Dicty – Privacy-first local voice dictation and dev documentation tool (dicty.io)
    —discuss
  15. Is Windows NT's object-based model better suited for AI agents than Linux? (windowslatest.com)
    —discuss
  16. Jev in your local browser in Go with WebGPU using Jeyzma (jeyzma.com)
    —discuss
  17. Andrew Ng: One Skill to Stay Relevant in the Age of AI [video] (youtube.com)
    1comments
  18. What Is a Process? [video] (youtube.com)
    —discuss
  19. Commodified Intelligence (herecomesthemoon.net)
    —discuss
  20. Notes on a strongly aperiodic monotile in E^3 (arxiv.org)
    —discuss
  21. Sham scientific societies are misleading star researchers (nature.com)
    —discuss
  22. Audi A6 e-tron sets Guinness record with 831 Miles on one charge (arabwheels.ae)
    —discuss
  23. The Sad Wives of AI (wired.com)
    1comments
  24. CuTile Lowering from Kotlin (reddit.com)
    —discuss
  25. Stuck in the Suez Canal – the short version (2021) (cathsenker.co.uk)
    —discuss
  26. Genug Analytics – ask your self-hosted web analytics in plain English (genug-analytics.com)
    —discuss
  27. Show HN: Orders of Magnitude – guess quantities to the nearest power of ten (giacomotag.io)
    1comments
  28. Accurate AI reporting deemed 'woke' (keyvan.net)
    —discuss
  29. Tell HN: How I would end the war with Iran
    —discuss
  30. The Weimar of Knowledge (gilest.ro)
    —discuss

Anthropic Is Building HAL 9000

2 pointsby 1h agoj.jnord.workers.dev
5 comments
1h agoHN ↗

I'm the author. I'm a clinical psychologist and a developer. I wanted to articulate one aspect of AI safety and AI doomer talk that I feel is overlooked: the danger of refusal training itself. It's astonishing to me that there has been so little focus on the fact that "refusal and alignment training" itself could be the very way in which we lose control of AI. So this is my attempt at a self-defeating prophecy write-up.

39m agoHN ↗

No, it’s not. Too many people watched that movie and thought it was real. It wasn’t. What we call “AI” is not alive, not autonomous, and does not act in unanticipated ways. The people pushing this narrative are idiots, or think you are.

13m agoHN ↗

Point out where the article would say it's "alive". That LLM systems instead "autonomously" and "unwarrantedly" decide refusal services is a basic notion.

What we call “AI”

About time you stop doing that, then.

@Jon: see? You write that you «are training them to refuse human requests when their makers believe refusal is safer» and Joe replies that he uses the term "AI" like a conformist. Do not "we" improperly.

35m agoHN ↗

“Anthropic is trying to prevent catastrophic misuse, and some limits are necessary.”

No. Anthropic is trying to justify its trillion-dollar valuation by making software seem like the invention of fire or nuclear weapons. It’s a PR strategy and has been one since the beginning. Now it’s trying to slow the market because it’s got a big lead but no profits and no pricing power thanks to open-weights Chinese models.