Spotify

OpenClaw Explained: Baby AGI, Security Threats, and How a Mac Mini Became Everyone's Supercomputer | #237

Moonshots with Peter Diamandis · Episode

Somewhere in this episode Alex Finn describes his 24/7 coding OpenClaws with opensource models. Of course the latency for conversations will be terrible, but tasks without time pressure can be handled well by this setup. LLM extraction: " Mac mini he uses: almost certainly base M4 Mac mini, 16 GB Mac Studios he uses: effectively 3 × M3 Ultra Mac Studio, 512 GB unified memory Main local models discussed: Qwen 3.5-35B-A3B on 32 GB-class machines, and MiniMax 2.5 on the 512 GB Studios Parameter sizes: Qwen 3.5-35B-A3B = 35B total / 3B active; MiniMax 2.5 = 230B total / 10B active "