I honestly don’t see how the average working class adult find that time. I feel like gaming is a luxury for rich/middle-class teens.

  • Jakeroxs@sh.itjust.works
    0·
    3 days ago

    Yes, as I chose not to have kids and my wife also enjoys video games.

    Edit: However, I do tend to hyperfixate on other interests for periods of time.

        • PeriodicallyPedantic@lemmy.ca
          0·
          2 days ago

          Still working on replacing Gemini. Gotta make a custom integration, and I’ve never use python and the HA dev docs are a bit confusing to a beginner

          • Jakeroxs@sh.itjust.works
            0·
            18 hours ago

            Lol I have that on my to do list too, already have text Gen webui setup and even tested some tts/stt to make sure I worked, actually connecting it all together and have it overtake gemini on my Google home for instance, not sure how possible that actually is.

            • PeriodicallyPedantic@lemmy.ca
              0·
              13 hours ago

              So here is what I’ve found so far:

              Getting home assistant set up with either Google gen ai, OpenAI chat GPT, Anthropic, and OLlama (if you already have OLlama running somewhere), is pretty simple and it is pretty configurable.
              Even though I’m trying to get away from Google, it’s free so I’m using Gemini 2 for testing purposes.

              You can cast your dashboard to a Google/nest Hub, but idk if it can use the mic or listen for wakewords. It looks like you can’t use Google home mini as a voice satellite without physically altering them, afaict (articles are a bit old).

              But if you’re ok with getting new hardware, you can buy HA voice satellite devices, or DIY it, which can serve the purpose of Google home mini.
              Likewise you can get a cheap tablet and load the HA app on it, and use that in place of a nest hub.
              Both of these should work with wakewords and invoke the voice assist functionality.

              Right now I’ve got it set up so it can do voice assistant stuff using local TTS and STT, and with Gemini as the LLM, running on my Pi and interacting with it via the home assistant app (haven’t set up wake word yet, but lots of tutorials).

              What I’m dealing with is that I don’t wanna use Google/OpenAI/Anthropic for my LLM, and I’m not ready to build an AI home server. I noticed that CloudFlare offers AI hosting for free (below a certain usage), and since I’m already using CloudFlare for my ingress I wanted to make a new CloudFlare AI conversation agent integration, since none seems to exist.
              I’ve built a toy agent that just echos back what you say in all caps, and it works (although there is an error in the dev console that I can’t figure out, even though everything seems to be working). And I’ve figured out how to invoke CloudFlare AI workers (CloudFlare docs are great). But the actual hard part of mapping the CloudFlare inputs/outputs to the HA assist inputs/outputs is still to be done.
              I still don’t really fully grasp what the HACS development cycle is supposed to look like. How do I test my HACS integration locally, and safely?