New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. We don't need AI regulation – leave safety to us, Nvidia's Jensen Huang says(techcrunch.com ↗)
    discuss
  2. The Customizable Select(utilitybend.com ↗)
    discuss
  3. Schadenfreude(wikipedia.org ↗)
    discuss
  4. The Unwilding of Patagonia's Pollinators(biographic.com ↗)
    discuss
  5. Ask HN: Are requests/demands coming from AI companies for regulation genuine?
    1comments
  6. Show HN: MinMAX Sort – A Stable Bidirectional Selection Sort Family Algorithm
    1comments
  7. Prefill a 284B model on Nvidia. Decode it on Apple Silicon. Over plain 10GbE(github.com/chadhurley25075-png ↗)
    1comments
  8. MartyPC – A Cycle-Accurate IBM PC/XT Emulator(github.com/dbalsom ↗)
    discuss
  9. We are all Product Engineers now(seldo.com ↗)
    discuss
  10. Show HN: An open-source Socratic coach for competitive programming(chromewebstore.google.com ↗)
    1comments
  11. Looking back on Servo's first donation-funded role(servo.org ↗)
    discuss
  12. Show HN: Check an NPM package or MCP server for malicious code before install(bouncer.run ↗)
    discuss
  13. The first ever API built for dictation;Today we're launching the Dictation API(assemblyai.com ↗)
    1comments
  14. There are no optical corrections in the Parthenon(doi.org ↗)
    discuss
  15. I wrote a kill number before I wrote the landing page copy(deskthaw.com ↗)
    1comments
  16. A software thing I built: GPS on a 25MHz 486-SX(vcfed.org ↗)
    1comments
  17. I built dual cam app because I had credits left and didn't know what else to do(play.google.com ↗)
    1comments
  18. Noise-Coded Illumination for Forensic and Photometric Video Analysis(arxiv.org ↗)
    discuss
  19. Show HN: HTMLCATS – Editor-First Learning for HTML, CSS, and JavaScript(htmlcats.com ↗)
    discuss
  20. Fusion in Devin Desktop and CLI(cognition.com ↗)
    discuss
  21. Show HN: Makerpa.ge – project hub and syncing badge across your sites(makerpa.ge ↗)
    discuss
  22. IOS27 removes feature to hide safari, why?
    discuss
  23. Is Jev That Good?(twitter.com/completeskeptic ↗)
    discuss
  24. Dan Selsam's Personal Statement on AI Risk(reddit.com ↗)
    discuss
  25. Where Prediction Markets Struggle(economist.com ↗)
    discuss
  26. The AI Language We Can't Read: Neuralese [video](youtube.com ↗)
    discuss
  27. Tesla Full Self-Driving customers can now access lower insurance premiums(drive.com.au ↗)
    1comments
  28. Continual Learning Mechanisms Compose for Long-Horizon Memorization(arxiv.org ↗)
    discuss
  29. Are you committed to your company or your team?(economist.com ↗)
    discuss
  30. Sam Altman says world 'right to be afraid' but 'should trust' AI firms(bbc.com ↗)
    1comments

From Academia to Alignment

4 pointsby 2h agomillicosm.substack.com
2 comments
2h agoHN ↗

Good luck to the ARC. Alignment is not a solvable problem. A paradox cannot be solved.

I've stated like the following:

“Alignment, which we cannot define, will be solved by rules on which none of us agree, based on values that exist in conflict, for a future technology that we do not know how to build, which we could never fully understand, must be provably perfect to prevent unpredictable and untestable scenarios for failure, of a machine whose entire purpose is to outsmart all of us and think of all possibilities that we did not.”

The full elaboration I wrote up here - https://www.mindprison.cc/p/ai-alignment-why-solving-it-is-i...

1h agoHN ↗

Kinda funny a chatbot threatening its users was the breaking point for them to see "AIs would overtake humans as the dominant intelligence". I would've thought it would be benchmarks or AI cleanly solving unsolvable math problems.

I guess that means fear is a better motivator than demos.