Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Less Typing, More Work(nicolasdular.com ↗)
    discuss
  2. Understanding the European Cyber Resilience Act (CRA)(openssf.org ↗)
    discuss
  3. Postgres 19: What's new in monitoring?(clickhouse.com ↗)
    discuss
  4. Find-jevable-code – audit a repo for Jev-replaceable decisions(altslate-labs.github.io ↗)
    discuss
  5. Grim Fandango Puzzle Document (1996) [pdf](jmac.org ↗)
    discuss
  6. OpenAI Degrading Their Models, Also for Paid Users(reddit.com ↗)
    discuss
  7. Bot-free self-hosted analytics with GoatCounter on NixOS(bernat.ch ↗)
    discuss
  8. Bull Beats Out HPE for Next-Gen Lumi AI Supercomputer(nextplatform.com ↗)
    discuss
  9. We reviewed package-manager ambiguity in 10 public repos(zfinia.com ↗)
    discuss
  10. A folding stepladder from fence timber(oxplot.com ↗)
    discuss
  11. You could have built Jev(sgnt.ai ↗)
    discuss
  12. What Sun Got Wrong(dtrace.org ↗)
    discuss
  13. How Notion handles concurrent editing with CRDTs(notion.com ↗)
    discuss
  14. Faraday Learns to Reason(medium.com/vektormemory ↗)
    discuss
  15. Far-left party wins Berlin election, pledging to nationalise housing(reuters.com ↗)
    discuss
  16. AurionMail: E2EE suite (CryptPad and Mail) with user-friendly single-password UX(aurionmail.github.io ↗)
    discuss
  17. Polars 2.0.0rc2(pypi.org ↗)
    discuss
  18. Show HN: I wrote the code and let an LLM review it(jaysinh.dev ↗)
    discuss
  19. Where Motion Is Mistaken for Movement(medium.com/gurvinder372 ↗)
    discuss
  20. DeltaTensors: Git for model fine-tunes that also saves you storage
    discuss
  21. Compsci grads facing recession-like job prospects thanks to AI(theregister.com ↗)
    discuss
  22. A Cute Proof That Makes e Natural(poshenloh.com ↗)
    discuss
  23. Proof That Computers Can't Do Everything (The Halting Problem) [video](youtube.com ↗)
    discuss
  24. Lithography Machines (Proposal)(drive.google.com ↗)
    discuss
  25. Turn any Llama-server into a jev system one endpoint(github.com/khimaros ↗)
    discuss
  26. GSoC 2026: Port the Enlightenment desktop environment to NetBSD, part 2(netbsd.org ↗)
    discuss
  27. TXR: An Original, New Programming Language for Convenient Data Munging(nongnu.org ↗)
    discuss
  28. Meta's Muse Is Better at Surveilling Than Helping Me(wired.com ↗)
    discuss
  29. OSS, security and Funding: libjpeg-turbo
    2comments
  30. Grit your teeth and ship it(seangoedecke.com ↗)
    1comments

AI chatbots give wrong answers to financial queries 'most of the time'

36 pointsby 1h agoft.com
9 comments
57m agoHN ↗

Article is just a vague summary of https://www.saturnos.com/report/artificial-authority

Anecdotally, current models seem to be decent at general personal finance principles - certainly better than the majority of personal finance education that people get exposed to unless they seek it out and read a variety of books and sources. But I wouldn't trust them with direct decision making with actual money due to the training lag time on current tax policy, etc.

48m agoHN ↗

I would prefer to use agent-assisted python scripts that chatbot.

44m agoHN ↗

Agreed, this works really well for me. Double check the math/python, execute many times without a LLM that can change o

43m agoHN ↗

I like how FT makes me accept cookies from their 46 “technology” (advertising) partners before showing me that the article is behind a paywall anyway.

37m agoHN ↗

You actually like that? I find it kind of annoying.

33m agoHN ↗

No they do not like it, it is a figure of speech to underline how much they do not like it.

29m agoHN ↗

Single shot or with reasoning enabled? My experience is that reasoning dramatically reduces hallucinations and improves output quality. I don't trust models without it.

17m agoHN ↗

These models do pretty well in benchmarks and real world so I'm highly suspicious of this article. Further more, in the original report, the examples of bad answers are from Haiku - at least 7 out of 10. Anyone who knows anything about LLMs know that haiku shouldn't be used for anything pretty much.

There's no reproducible set either. I'm not gonna trust this report.

8m agoHN ↗

Now, compare this to a recent story that seemed to claim the opposite:

https://news.ycombinator.com/item?id=49139102

I don't have the time to review the underlying research and decide which one is more correct. My personal biases make me want to believe the current one. Your personal biases may be pulling you in the other direction. How do we make the conversation more intelligent than that?