Some of you may know that I dabble in lots of self-hosted fun. I’ve been self-hosting for several years, both internally on my local network as well as externally on a VPS. On my local setup, I use a whole bunch of different apps to help with tasks or store important data, such as recipes we like or even Linux ISOs.I’ve been running Home Assistant for well over 6 years (if not longer by now). I’ve also been a big fan of Music Assistant. I recently learned of an AI Radio plugin that allows you to use a local LLM (or a paid LLM such as OpenAI) along with TTS software such as Piper to have a DJ sometimes chat between your songs. Out of the box you have options such as “Morning DJ, and Music Nerd”. I don’t have the most powerful “AI” setup at home at the moment since I’m just using an N100, but it still is fun to check out and see how these work. One of the cool features is that there is a way to load custom voices into Piper; you can make your own model (with your own recordings), or other people have provided voice models they have created. I have Jean-Luc Picard being a morning DJ, and Jarvis being a Music Nerd…
I’ve been using Claude a lot lately to help me harden my network (with me reviewing its suggestions before I make them, of course). Also, LLMs are very helpful for hunting down and tracking small issues you may be having internally when you can’t find the specific answer to your question when you do a normal search. Are AI LLMs the solution for everything? No, but they are a great tool to help you narrow down the solution to a problem you are having.
Now I’m considering getting a cheap GPU for my server instead of using the N100… but these prices are so out of my league even for most “cheap” options at the moment.
