The Genie coefficient proposes a new metric to quantify the distance between a user's prompt and the AI's actual output. Current benchmarks track capability but ignore unspoken human assumptions. This gap often leads to literal but useless results. Practitioners need this measurement to improve how LLMs interpret implicit intent without requiring exhaustive prompting.