Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Google OpenAI Anthropic Begin Forming SAFA – Standards Authority for Frontier AI (proactiveinvestors.com)
    —discuss
  2. I don't read code anymore (duyet.net)
    —discuss
  3. Kids turned the comment section of an NPR podcast into a group chat (thisamericanlife.org)
    —discuss
  4. The NSA Is Spending Billions to Test AI Models, Classified Estimates Show (gadgetreview.com)
    —discuss
  5. GPU Glossary (Readable) (modal.com)
    —discuss
  6. Flock (YC 2017) Wants Detailed Map of Its Surveillance Cameras Taken Offline (theintercept.com)
    2comments
  7. Deming in the machine running an AI assisted film (medium.com/zrkjsy)
    1comments
  8. Scoop: Top AI companies probing tens of thousands of security incidents (axios.com)
    —discuss
  9. Ask HN: How Jev in production for generative UI working out, anyone tried ?
    —discuss
  10. China and U.S. to open AI 'communication channel' after summit (japantimes.co.jp)
    —discuss
  11. US hiring appetite is healthy as economy powers ahead (businessmirror.com.ph)
    —discuss
  12. Advanced Autocorrect Powered by Jev (levmiseri.com)
    1comments
  13. Entropy as a Measure of Uncertainty (jonathanwarden.com)
    —discuss
  14. 102.8B tokens and 474K LOC with Gemini Flash – Telemetry and post-mortem (github.com/janson79jc)
    —discuss
  15. Stop Building for Steps. Build for States (clyxconsulting.substack.com)
    —discuss
  16. DICS: Staged DI for C# and Unity without runtime reflection (github.com/playq)
    1comments
  17. Dental Scope: Explore dental anatomy in 3D (dental-scope.vercel.app)
    —discuss
  18. NIH's new PubMed tool to strengthen research replication and reproducibility (nih.gov)
    —discuss
  19. Why Your Paycheck Buys More Fridges but Less House [video] (youtube.com)
    —discuss
  20. Booms, Bombs and Bonds: Interest Rates, Part I (paulkrugman.substack.com)
    —discuss
  21. Dogs Show a Surprising Speech-Processing Skill Once Thought Unique to Humans (discovermagazine.com)
    —discuss
  22. Intelligence Explosions Are Social (deepmind.com)
    —discuss
  23. RF Capture Guide for VHS (github.com/oyvindln)
    1comments
  24. Katz's Management Explains the $50 Lost Ticket Fee (eater.com)
    —discuss
  25. Meta-Stripper – Strip Image Metadata Locally (devkram.de)
    —discuss
  26. Anton Poulter (adamapples.blogspot.com)
    —discuss
  27. SNL Weekend Update: Anthropic CEO Dario Amodei on A.I.'S Threat to Humanity [video] (youtube.com)
    4comments
  28. Ember-1 (fireworks.ai)
    —discuss
  29. Gut Feelings (2007) [pdf] (hadinur.net)
    —discuss
  30. How Jev works: calibrated decision models (victordibia.com)
    —discuss

I was impressed by Jev, please explain why I shouldn't be

4 pointsby 1h ago
5 comments
I'm not a computer science guy, but I had a play with Jev and found it impressive. It allowed me to offer non-technical researchers the ability to classify arbitrary things their participants enter, without prior training or classification, at near instant speed, and negligible cost. My users may find that very helpful in the qualitative research world.

I showed them an innocuous working example of how they could classify thousands of participant entries as "mentioning Italian food" whether it mentioned "pasta", "penne", "rigatone", etc etc. Jev did this task with high confidence, and showed it could differentiate between the inverse, eg had very low confidence when I switch entry to mentioning "hamburgers". I find that a very useful tool to pass on to non-technical colleagues.

But on HN, there are lots of much more qualified people saying Jev is a con, a step backwards, or bonkers that people are impressed. Will one of them tell me plainly what software was doing this prior to Jev, and the like? I don't doubt it existed, but I never came across it. I'm interested to hear from people more expert than myself.

To be precise, software must be able to: - Classify thousands of entries with arbitrary themes or topics, even if entry doesn't mention that theme or topic directly, with high enough confidence to be useful. - Do it with thousands of entries in 0-3seconds. - Do it at negligible cost, eg sub 0.1cents

1h agoHN ↗

You're allowed to be excited about new technology.

56m agoHN ↗

What are you allowed to feel when it's old technology?

51m agoHN ↗

Anything you want. Do as you please, stop waiting for the internet to tell you how to feel about things.

56m agoHN ↗

what software was doing this prior to Jev

BeRT, BART and FLAN-T5, for text classification. All of those models are ~half a decade old now, are based on the transformer model, and small enough to run locally. They're not perfectly SOTA, but perform quite well in embedded applications. Fine for low-stakes stuff, I think.

Image classifiers are a hugely competitive space. You've got BeiT, DeiT, MobileOne, ConvNext, FastViT and several more models that all support image classification pipelines.

28m agoHN ↗

Thanks for quick reply. I downloaded FLAN for comparison and you are quite right, gives similar results, with the benefit of privacy and self-sovereignty in serving the model.

So other than untestable claims of _how_ Jev generates responses, it's main feature right now is its API design?