DeepSeek Harness First Impressions: Architecture, the Plugin Market, and the 20 Plugins That Stayed
A week with DeepSeek Harness: 30 in, 20 kept.
LLM, AI Native Infra and Agentic AI.
A week with DeepSeek Harness: 30 in, 20 kept.
Bigger models make GPU partitioning matter more.
Why GPU scheduling matters: how the default scheduler misses node, card, and topology placement, how HAMi v2.10 fills the gaps, and what mutex really means on kind.
GPU sharing is moving from soft allocation to governance—scheduling decisions and runtime isolation can finally be reconciled.
HAMi moves from cluster to desktop with Olares.
A GPU explainer for Kubernetes veterans new to AI. Maps token, model, training, inference, Transformer, Tensor Core, HBM, and KV cache to concepts you already know.
From GPU utilization to productive GPU-hours.
A practical AI Infra review of Agentic AI reliability, covering a five-dimension framework, fault tolerance, recovery, observability, and hybrid architecture design.
From GPU hardware, Kubernetes scheduling, inference engines to token cost — understanding the 8-layer observability architecture for modern AI infrastructure.
How I built a personal AI infrastructure using ChatGPT, OpenClaw, Obsidian, GitHub, Lark, GLM-5.1, and a Mac mini M4.