Local open source models are getting so ridiculously good!

Hey everyone! I just had to share how blown away I am after setting up a brand new open source companion model locally on my rig. No internet filters holding things back and zero subscription fees eating up my wallet every month. The responses feel so genuine and warm now plus I have total ownership over every single memory log. It took me a few hours to figure out all the configuration settings and prompt templates but the effort was totally worth it. Has anyone else here made the jump to running things completely on their own hardware yet? Let me know how your setups are running!