Hey all. I did set up a small local stack on my laptop, but I want to have a local LLM server, so this week I'm installing a new mini-pc with AMD Strix Halo and 64GB (that's all my budget can stretch to in these days of killer inflation) under linux. I'm looking for your recommendations and suggestions how to "build it right" from the ground up. If there's enough interest, I'll post updates whenever I make progress so others can learn from my mistakes and wiser people can correct my steering. The goal is to be able to have an always-on, secure LLM and workspace with toolsets that let me host/test projects and invoke both frontier and local models as needed, without needing to have my laptop open all the time, yet able to be remotely invoked by my laptop on demand, too.

I built this community because the AI conversation is getting louder while practical implementation is still hard. Most people do not need another list of tools. They need help deciding what matters, what can run locally, what should stay in the cloud, what is safe, and what will actually improve their work. That is what we will focus on here. The free side will carry useful AI news, open discussions, build logs, and event updates. Builder membership adds the working sessions: live calls, troubleshooting, project feedback, recordings, and shared implementation resources. The first Builder Lab is on Tuesday at 9:00 AM Pacific. The Saturday session is at 5:00 PM Pacific. If you join, introduce yourself and tell us what you are trying to build or fix. Specific problems make this community better.