Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Recursive self-improvement of AI research agents(arxiv.org)
    discuss
  2. Show HN: Analysis of Games from Chess Olympiad with Opus 5.5 and Stockfish(unremarkable.info)
    discuss
  3. Microsoft disrupts AI-assisted platform that compromised 12,000 accounts(arstechnica.com)
    discuss
  4. Why Tool AIs Want to Be Agent AIs(gwern.net)
    discuss
  5. Fearless SIMD v1.0 is here(linebender.org)
    discuss
  6. M6 Mac Mini uses soldered NAND storage again(macrumors.com)
    discuss
  7. Show HN: Yass – Yet Another Spec Syntax Enters Beta(github.com/shakefu)
    1comments
  8. Building a docs site that screenshots itself(semaloop.com)
    1comments
  9. There is no epidemic of loneliness, but there is an epidemic of scurvy(experimental-history.com)
    discuss
  10. Claude Code Hooks Explained: The Deterministic Layer Around Your Agent(blakecrosley.com)
    discuss
  11. Jevify skill – Gets your existing agents running on Jev
    discuss
  12. Show HN: use-citations – Claude Code skill for citation backed answers(github.com/sshah03)
    discuss
  13. Show HN: Daily-updated open dataset of flight fare history by route and month(github.com/deusacc)
    discuss
  14. CobaltC now has a compiler (for Linux)(github.com/strawberry9)
    discuss
  15. First Petabit-Class Transoceanic Subsea Cable(fb.com)
    discuss
  16. Ask HN: Engineers on highly polished products, how are you using agentic coding?
    1comments
  17. Which AI model should your team use for Ansible(steampunk.si)
    discuss
  18. Unreal Agent An async-first agent harness(github.com/unreallabsai)
    discuss
  19. Transformers now runs llama.cpp quants(huggingface.co)
    discuss
  20. Australia's giant data-centre boom(economist.com)
    1comments
  21. Avhash: BlurHash replacement, using a AV1 frame with container headers stripped(github.com/keithamus)
    discuss
  22. Show HN: Ceptile AI Search platform which combines web results with AI chat(ceptile.com)
    discuss
  23. Claude Opus 5.5 – Pelican Game(riba2534.cn)
    2comments
  24. I Gave My Life over to Meta's A.I. Agent and Was Blown Away(nytimes.com)
    1comments
  25. El Niño projected to cause 450k deaths in next six months(theguardian.com)
    1comments
  26. Mastering LLM Inference Optimization(machinelearningmastery.com)
    discuss
  27. The Rio Grande's pecan problem (2025)(hcn.org)
    discuss
  28. Show HN: Visually orchestrate Claude Code AI agents(github.com/rondoflow)
    discuss
  29. Benzi – Code Intelligence MCP/Harness beating Claude Code and CodeGraph(varianttech.net)
    1comments
  30. A Summer of AI Optimization(lemire.me)
    discuss

Google Found the Pages. ChatGPT Found the Entity

2 pointsby 2h agorichresults.ai
1 comments
1h agoHN ↗

Hi, This is Stefan from richresults.ai. I am the author of this post.

We made the AEO and the Entity Architecture for JUNO PORTALS. Seven days after JUNO PORTALS launched, I did a test with 20 Chats. Not logged in, via VPN and private window. 10 in German and 10 in English. ChatGPT cited Crunchbase, even though juno-portals.com was not indexed in Bing. On Google the pages of juno-portals.com already appeared in the blue links, but Google AI Overview cited completely different "Juno Portal" entities.

ChatGPT recognized JUNO PORTALS correctly every time, in 20 out of 20 chats.

What was really surprising is the issue about JUNO PORTALS MAG in the Chat. I started the chat addressing JUNO PORTALS MAG. ChatGPT could not deal with JUNO PORTALS MAG and cited other hits like a music label. After that, ChatGPT did not find JUNO PORTALS. Even when I directly addressed JUNO PORTALS.

This shall be a clear and true test, handmade! So my question is: Did anybody observe similar differences between indexing and entity resolution?