Descriptions:
Try Tank for yourself! Join the community 👉 https://mrc.fm/cmc
I’ve been itching to see how a regular Mac mini really holds up as a local AI machine, so in this stream I’m putting one through its paces live. I’ll run a few local LLMs, push some inference, and see whether that unified memory earns its keep for the sort of self-hosted, local-first setup I love. We’ll get into RAM tiers, real-world token speeds, and whether this compact little box is the one worth buying or whether you’re better off holding out for what’s coming. Bring your questions and your prompts and we’ll test them together.
0:00 Unboxing Mac Mini M4
5:12 Setting up macOS screen sharing
9:25 Configuring headless Mac Mini settings
12:50 Selecting local AI models
18:06 Optimizing macOS for headless servers
23:31 Controlling Chrome with MCP servers
28:56 Extracting Google Maps reviews with AI
37:39 Designing a multi Mac AI network
41:17 Filtering search results with AI
46:37 Managing Claude API usage limits
51:58 Troubleshooting Claude code token limits
55:52 Repurposing old MacBooks as servers
