Through the last decades, we are witnessing rapid changes in the world of computing. Computing products and services are becoming more consumer-oriented and personalized. In this emerging paradigm, the better we understand individual users, their history, preference, and context, the better we can improve their experiences and maintain their loyalty. Mobile phones, due to their rich sensing modalities, constant cloud connectivity, and personal association to the users, provide an interesting platform for sensing and understanding human as subjective entities more than objects in motion. In this paper, we articulate the vision of subjective sensing and its role in advancing mobile services. We present the Munich project and summarize some initial results on continuous low power mobile sensing, accuracy and latency tradeoff, and feature extractions for mobile search and ads.