Where cloud hype
meets ground truth.
Dispatches from inside the machine. Unsponsored. Unfiltered. Unsigned.
- DGX Spark: vLLM vs Atlas for Local Inference
Two inference runtimes on the same model: vLLM and Atlas via sparkrun. Memory models, real throughput numbers, and which one makes sense for a personal agent.
- Update: Hermes on DGX Spark - Qwen3.6-35B-A3B-NVFP4 via vLLM
Follow-up to the Proxmox/ROCm iGPU build. Hardware upgrade to DGX Spark (GB10), three distinct failure modes getting vLLM serving an NVFP4 MoE model, and what finally fixed it.
- Running a Local AI Agent on Proxmox with AMD ROCm GPU Passthrough
How I built a self-hosted AI employee on a mini PC, an iGPU, and stubbornness. Honest guide covering the gfx1035 fix, Hermes agent setup, real performance expectations, and the hybrid Haiku+Ollama architecture that actually works.
- Building an IronClaw AI Assistant on Debian LXC
ROCm 6.0 and IronClaw on a Beelink EQR6 Proxmox node. Same hardware as before, different goal.
- Gen AI Rant... But Also Kinda Cool
I'm using AI to write about AI. The irony is not lost on me.
- 1st post
Lighting a match to the tech industry. Raw thoughts on cloud headaches, AI illusions, and everything broken in IT.