Hi guys! I was wondering:

What self hosted software are you missing? What would you which existed?

Background: I am a quite Senior Dev and have 6 months between gigs. Would be fun to start a side project to keep my skills sharp.

  • Cazzah@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    That’s because LLMs don’t do that.

    The companies that offer those services basically do some tricks behind the curtain.

    Like let’s say you want an LLM to learn your corporate docs. LLMs can’t do that because they need millions of text from across the internet just to learn to speak English… You can’t feed your 1000 docs and 10,000 emails in and point to it and say “Forget the billion documents you injested and pay attention to this… but also retain the ability to speak English”

    What they actually implement is a standard text search engine, that returns matching paragraphs from the relevant documents, prompts to LLM with something like "This paragraph may contain an answer to user question X. If it does, please paraphrase it.

    • kataflokc@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      Yes, that’s exactly what I want it to do

      Most of my 60-70 email replies per day are answered in almost the exact same way

      I want it to read an email and then, using paragraphs or sentences from my previous emails, automatically generate a response

      There are already companies out there who are generating what they term small language models - basically hybrid models of say gpt 3.5 plus a large volume of corporate data - but they are all cloud based

      Others offer plugins to help answer your emails

      I’d like a combination of the same to run locally

      • Cazzah@alien.topB
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 year ago

        There are already companies out there who are generating what they term small language models - basically hybrid models of say gpt 3.5 plus a large volume of corporate data - but they are all cloud based

        I think you will find most of these are not small language models, but are instead the thing I said above - a llm like gpt + a search engine. Even small language models require millions of texts and only perform very specialised tasks.