Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Another terminal/shell AI assistant (shell native)(github.com/ther1d ↗)
    1comments
  2. Show HN: All-in-one open-source Tab organizer(github.com/ringlochid ↗)
    discuss
  3. The senior engineer death spiral(sunilpai.dev ↗)
    discuss
  4. Grit your teeth and ship it(seangoedecke.com ↗)
    discuss
  5. Nscale S-1 demonstrates unprofitability of AI(reddit.com ↗)
    discuss
  6. The Story of FCKGW-RHQQ2-YXRKT-8TG6W-2B7Q8(twitter.com/davepl1968 ↗)
    discuss
  7. California Cracks Down on Influencers' Undisclosed Political Ads(nytimes.com ↗)
    discuss
  8. The X3DH Key Agreement Protocol(signal.org ↗)
    discuss
  9. Market design: affinity networks like shooq, & why a festival should buy afm(danieldeboulay.com ↗)
    discuss
  10. Show HN: AudioSort, a web app for sorting audio files(vakaiser.com ↗)
    discuss
  11. Crypto Blew Its Big Moment–and the Blame Game Has Begun(wsj.com ↗)
    discuss
  12. OpenAI expects to burn $280B by 2030(ft.com ↗)
    discuss
  13. World Wide Words(worldwidewords.org ↗)
    discuss
  14. Show HN: Puzzle Season – five logic puzzles in the browser(puzzleseason.app ↗)
    discuss
  15. China doesn't need to beat US AI [video](youtube.com ↗)
    1comments
  16. Show HN: Migrating from Claude Code CLI to Claude Desktop(github.com/calebbarzee ↗)
    discuss
  17. The Ethics Issue of AI-Assisted Creation(not.ci ↗)
    2comments
  18. Professional cyclists can top 100 km/h. But safety has not kept pace(cbc.ca ↗)
    discuss
  19. Show HN: 23k Verified 1P Walmart & Target Products with Preserved UPCs(github.com/fluffyteddybear12 ↗)
    discuss
  20. Line-Rate Post-Quantum Byzantine Consensus via L1D Balanced Ternary(github.com/leadpiperl1 ↗)
    discuss
  21. Scott Jenson: Are we going to use the same Desktop UX forever? [video](youtube.com ↗)
    2comments
  22. RSA-896(saweis.net ↗)
    8comments
  23. Lattice enumeration: O(R^N) complexity reduced to O(R^2) with no floating point(github.com/andrew-j-morris ↗)
    discuss
  24. Using system-one models inside high-throughput data pipelines(southbridge.ai ↗)
    discuss
  25. TanStack supply-chain attack exposed ~170 private CrowdSec repos(crowdsec.net ↗)
    1comments
  26. Is It a Flatpak? Is It an AppImage? No It's AppJail(youtube.com ↗)
    1comments
  27. AI Bubble: 'They're coming to the end of this' – Eli the Computer Guy [video](youtube.com ↗)
    discuss
  28. Show HN: Bastardica – Create bastard web fonts(mitpit.com ↗)
    discuss
  29. Comparing reflection capabilities of C++, Zig and C3(nyr24.github.io ↗)
    3comments
  30. Do you separate the reviewer's window from the author's
    1comments

AI-Generated Movie Star Has a Total Meltdown on TV

8 pointsby 7h agohonest-broker.com
3 comments
5h agoHN ↗

Don't waste your time - the 'meltdown' was the underlying LLM starting to speak Chinese while being interviewed.

Their character has slow reactions, seems quite dumb and has no personality at all. They are probably running an old version of ChatGPT 4.x, plus some standard TTS/STT pipeline. There is nothing state of the art there, and a bad look for AI in general. It's 100% a (poor) marketing play and quite surprising it's been going on for so long.

5h agoHN ↗

It won't get the same level of attention, but the question prior to the lapse into Chinese was pretty damning as well. It wasn't a difficult question at all but she didn't understand or answer it correctly until it was asked a second time.