‹ BackHN Continuity

Thread

I built non-autoregressive decision models with RL a year ago

1363 points · 319 comments · nandakishor_ml

  1. pknerd · · focus · HN ↗
    Correct me if I am wrong, can I use Jev and this tool for ticket classification? I mean, for instance, a level 1 ticket contains a screenshot of the login page that displays an error, LLM can do it perfectly, can Jev do it?
    1. dgritsko · · focus · HN ↗
      At least for now, Jev is not multimodal. So a screenshot alone wouldn't cut it.
      1. adverbly · · focus · HN ↗
        I find that a bit interesting because the most system one part of the brain is probably the part used for visual processing.

        It's trying to use a human analogy but the analogy breaks down if you try to apply it directly

        1. verdverm · · focus · HN ↗
          I think I saw that the System 1/2 framing is based on the Thinking Fast/Slow book's framing
      2. Reubend · · focus · HN ↗
        Oh, I missed that! So the Doom demo was potentially just "harnessmaxxing"?
        1. nacs · · focus · HN ↗
          It was yes, lots of words in the input to describe state of the game-world at each checkpoint.
    2. rgbrgb · · focus · HN ↗
      It can’t do images but it can do a pretty good job of triaging urgency or choosing when to escalate. So I’d guess yes but it depends how detailed your classification is.
    3. someguy101010 · · focus · HN ↗
      you can use this for screenshots

      <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;thaitea&#x2F;laya-vision-smolvlm-256m" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;thaitea&#x2F;laya-vision-smolvlm-256m

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.