A novel programming paradigm, dubbed program-as-weights, has been proposed to tackle complex tasks that defy traditional rule-based implementation, such as log line analysis and search result ranking. This approach, known as fuzzy-function programming, enables the compilation of functions from natural-language specifications into executable code. By leveraging large language models, developers can create customized functions that operate locally, ensuring improved reproducibility and reduced costs. The proposed methodology has significant implications for various applications, including alert systems and data processing pipelines. Researchers have introduced this concept as a potential solution to overcome the limitations of traditional programming methods, which often struggle to handle nuanced tasks1. This development matters to practitioners because it could lead to more efficient and autonomous systems, potentially transforming the way developers approach complex programming tasks and changing the dynamics of the workforce in the process.