Blog

Technical writing on running open-weight models, agentic AI, and the engineering behind decentralised data systems. Posts are published on Substack — follow there for new ones.

16 August 2026 · Substack

Choosing Hardware for Local LLMs: From Model Size to Memory Budget

Part 2 of a practical guide to sizing hardware for open-weight LLM inference

Read on Substack

12 August 2026 · Substack

Choosing Hardware for Local LLMs: Start With Memory, Bandwidth, and Compute

Part 1 of a practical guide to understanding the hardware requirements of open-weight LLMs

Read on Substack