• podcast
  • Using Private Open Source AI Models on Hardware Asylum
  • Using Private Open Source AI Models on Hardware Asylum

    Published:

    Hosts: Dennis Garcia and Darren McCain
    Time: 32:46

    Subscribe Options
    RSS (MP3)
    iTunes (MP3)
    Spotify (Stream)
    Amazon (Stream)

    Originally recorded April 2025

    Show Notes

    Saying that AI is “hot right now” would be a massive understatement.  There has been a boom in popularity after the commercial launch of OpenAI’s ChatGPT and the rest of the world has been scrambling to catch up.  Thing is, AI has been around for years but, up until a few years ago it wasn’t all that useful and rather difficult to use.

    One of the goals that Dennis had when building the new Ninjalane server was to support AI activities for various projects including things like a “Site Search Chatbot”, “Retro PC Guru in the Ninjalane Labs BBS” and some other automations behind the scenes.  It will take some time to get those projects built so, in the meantime we are experimenting with an OpenWebUI install backed by Ollama to test different models and just have fun with a completely private AI installation.

    In this episode the duo discusses what it takes to setup a private AI and many of the ways you can customize how different AI models react to user input including a fun little project to help players setup their D&D characters and campaigns.

    Related Links
    Ollama
    Open WebUI
    Huggingface

    Episode 168 featured music:
    Little People - Start Shootin' (http://www.littlepeoplemusic.com/)