Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Xiaomi MiMo v2.6(xiaomi.com ↗)
    65comments
  2. The NASA/ESA Mars Sample Return mission has been canceled(science.org ↗)
    119comments
  3. What Sun got wrong(dtrace.org ↗)
    237comments
  4. Attention is all you have(alicegg.tech ↗)
    133comments
  5. Transformers Explained Visually(poloclub.github.io ↗)
    4comments
  6. Why does mathmain need an encrypted loader?(safedep.io ↗)
    20comments
  7. The Advisory Group on Mathematics and Artificial Intelligence(terrytao.wordpress.com ↗)
    15comments
  8. Divide by Depth for Instant 3D(gabrieloc.com ↗)
    3comments
  9. Turn off and restrict access to Apple Intelligence features on Mac(support.apple.com ↗)
    107comments
  10. In Search of a Compositional Theory of Self-Stabilization(muratbuffalo.blogspot.com ↗)
    3comments
  11. Grok 4.7(x.ai ↗)
    338comments
  12. Apple Copland D11E4 Booting in the Browser(pagetable.com ↗)
    8comments
  13. AI coding has made CI a bottleneck, so we reworked ours to keep up(linear.app ↗)
    41comments
  14. Frontier AI on Your Own Hardware(timdettmers.com ↗)
    6comments
  15. US halts flights at busy East Coast airports, says fiber line cut(reuters.com ↗)
    80comments
  16. RoboHarm: Do Frontier Robot Policies Refuse Unsafe Instructions?(robocurve.org ↗)
    3comments
  17. Show HN: A website that tracks US food prices every day(kadoa.com ↗)
    3comments
  18. Kev: Tiny Jev-like family of decision models built on top of Qwen3.5(github.com/jaredpalmer ↗)
    163comments
  19. Python Workers are now generally available(cloudflare.com ↗)
    21comments
  20. This Digital Radio Gets Messages to the World’s Remotest Locations(ieee.org ↗)
    30comments
  21. How do Traffic Signals Work (2019)(practical.engineering ↗)
    31comments
  22. Avoiding the babbling-idiot failure in a time-triggered communication system(ieee.org ↗)
    6comments
  23. A restored PDP-11/83 serving this page on 211BSD Unix(pdp1173.com ↗)
    28comments
  24. Fable 5 – Median thinking declined in August(twitter.com/lon ↗)
    194comments
  25. Grim Fandango Puzzle Document (1996) [pdf](jmac.org ↗)
    87comments
  26. Show HN: Foremerge – Catch intent conflicts between parallel coding agents(github.com/naw103 ↗)
    discuss
  27. Noodle Gallery- Open-source, self-hosted alternative to Google Photos and Immich(digitalescapetools.com ↗)
    34comments
  28. M5 Ultra Mac Studio Review(macstories.net ↗)
    192comments
  29. Heretic removes restrictions from language models(heretic-project.org ↗)
    88comments
  30. macOS 27: Workaround to avoid downloading AI models and save storage(reddit.com ↗)
    82comments

Claude says its DeepSeek when asked in Chinese

18 pointsby 6mo agotwitter.com
4 comments
6mo agoHN ↗

Tell me why anyone thinks these chatbots are intelligent.

6mo agoHN ↗

I got the same answer in Claude workbench which does not have a system prompt. On Web UI I was always getting Claude as answer.

Anthropic recently complained about DeepSeek using Claude to train their models. This response however shows that Claude is also doing the similar.

6mo agoHN ↗

Hmm, I tried asking in Chinese and it still identified itself as Claude by Anthropic.

6mo agoHN ↗

You need to ask via API to make sure that there is no system prompt.