Good engineers make decisions based on data. Most businesses assumed that the more data, the better the decision. Then, several factors put a halt to the hoarding of ever more data. GDPR and its localized counterparts, and the cost of storage. However, before the GDPR came into effect, the Datensparsamkeit approach already existed.     Datensparsamkeit is a German word that’s difficult to translate properly into English.

Frankel's platform  covers topics related to software development, technology trends, and coding tutorials. Through articles, tutorials, and personal anecdotes, Frankel offers insights into programming languages, development frameworks, and best practices in software engineering. Readers can learn about software design principles, development methodologies, and emerging technologies to enhance their coding skills and build high-quality software applications.

A Java geek

Token frugality (Tokensparsamkeit) is proposed as a key skill for developers using AI coding assistants. Two main strategies are covered: token compression via a proxy tool (rtk) that strips unnecessary words before sending to the LLM, and context optimization by limiting message history, tools, and MCP servers. The post also details a practical setup for running Claude Code against a local model using llama-cpp with a Mixture of Experts model (Qwen3.5-35B-A3B), covering pitfalls like Docker GPU access, context window limits in dense models, and parallelism settings. The final configuration uses Flash Attention, GPU layer offloading, and specific environment variables to point Claude Code at the local server.

Tokensparsamkeit for coding assistants