Brian Jackson
@jaxz.in
Day job: Platform Engineering Supervisor at Industrial Light & Magic
24/7 job: Father of two amazing teens, husband to an amazing wife, son and brother
Helping to make Star Wars, like Vader Immortal, Tales from the Galaxy's Edge, The Mandalorian, Ahsoka
24/7 job: Father of two amazing teens, husband to an amazing wife, son and brother
Helping to make Star Wars, like Vader Immortal, Tales from the Galaxy's Edge, The Mandalorian, Ahsoka
Update, I’ve got my Davis weather station data flowing into Home Assistant every few seconds instead of every 5 minutes. I did an end-run around their cloud service and I’m intercepting the equipment’s wireless signal using a USB radio and a Raspberry Pi. Pushed to HA over MQTT. #homeautomation
March 18, 2025 at 6:43 AM
Update, I’ve got my Davis weather station data flowing into Home Assistant every few seconds instead of every 5 minutes. I did an end-run around their cloud service and I’m intercepting the equipment’s wireless signal using a USB radio and a Raspberry Pi. Pushed to HA over MQTT. #homeautomation
March 1, 2025 at 9:56 PM
Bye bye Alexa, I’ve got a fully #selfhosted voice assistant! Whisper for speech-to-text, Llama 3.2 for the conversation agent, Kokoro-84M for text-to-speech, and #homeassistant connecting them all together behind their new Voice Preview device! All running on my home network on a mix of OS and GPUs
February 2, 2025 at 7:10 PM
Bye bye Alexa, I’ve got a fully #selfhosted voice assistant! Whisper for speech-to-text, Llama 3.2 for the conversation agent, Kokoro-84M for text-to-speech, and #homeassistant connecting them all together behind their new Voice Preview device! All running on my home network on a mix of OS and GPUs
I converted my old Windows gaming PC into a Linux machine learning sandbox. Now I'm hosting my own LLM with Ollama and Open WebUI and Llama 3.2 performs well even with my 5 year old Radeon GPU. And I can access it from anywhere thanks to Tailscale so I might end my ChatGPT sub. TBD #selfhosted
January 12, 2025 at 10:54 PM
I converted my old Windows gaming PC into a Linux machine learning sandbox. Now I'm hosting my own LLM with Ollama and Open WebUI and Llama 3.2 performs well even with my 5 year old Radeon GPU. And I can access it from anywhere thanks to Tailscale so I might end my ChatGPT sub. TBD #selfhosted