What I'm exploring:

• Running the latest LLMs locally — Ollama, LM Studio, llama.cpp, and whatever dropped last week
• AI orchestration harnesses: LangChain, LlamaIndex, Open WebUI, etc
• Coding harnesses: Claude, Copilot-CLI, OpenCode and freinds
• Local infrastructure: Proxmox, Docker, and DGX Spark
• macOS as a surprisingly capable local inference machine (Apple Silicon is no joke)
• Staying current — because in this field, a month-old mental model is ancient history

Philosophy: If it needs an API key to phone home, I'd rather run it myself.