Show HN: Sunk Cost – How long until a local LLM rig pays for itself?
Summary
The article analyzes the break-even point of running a local LLM rig, outlining the factors that influence ROI such as memory, bandwidth, and electricity versus continuing API costs. It explains that local speed is a function of memory bandwidth and tokens per read, with API speed affecting only time-to-value, and presents a framework to estimate when a local setup pays off for SMB users.