Recently, there have been media reports that during rush hour or rainy days in some cities, a contradictory phenomenon of taking a taxi frequently occurs: on one hand, passengers are eagerly waiting for dozens of queue prompts on their mobile phones, and even if they increase the price, it is difficult to call a car; on the other hand, ride hailing drivers listen to the system's voice reminder of "surge in orders", but have no orders to take. On the surface, this seems to be the result of a shortage of transportation capacity during peak hours. However, upon further analysis, the algorithm based dispatching mechanism of the platform is the key driving force behind exacerbating the supply-demand imbalance. In order to pursue overall matching efficiency, the algorithm will prioritize orders to drivers with high ratings and fast responses. In theory, this may improve overall efficiency, but in reality it creates new waste: a large number of ordinary drivers are "idle" by the system, and the massive demand of passengers cannot be effectively met. The core of the problem lies in the "black box" attribute of the algorithm. The platform holds key data such as passenger bids, driver locations, and real-time road conditions, and can intervene in benefit distribution through opaque rules. However, passengers find it difficult to grasp the real queuing situation and can only fall into a passive price increase situation. This information asymmetry allows the platform to both exploit passengers' "ride hailing anxiety" to explore price limits, and regulate driver competition and profits through dispatch rights. To solve the dilemma of drivers and passengers during peak hours, it is necessary to break through the algorithmic "black box". The platform should publicly disclose the basic rules of dispatching, priority considerations, and dynamic pricing mechanisms to the public, and accept social supervision. Regulatory authorities should conduct regular reviews of the fairness and rationality of algorithms, and resolutely eliminate discriminatory dispatching behavior. The platform still needs to optimize the dispatch rules. The platform prioritizes the dispatch of drivers with good reputations, with the original intention of improving the quality of driver services and optimizing the overall travel experience through positive incentives. However, in special scenarios such as morning and evening rush hours and severe weather, this rule still needs to be refined and adjusted based on reality. Priority should be given to meeting the basic travel needs of passengers and ensuring the right of drivers to obtain reasonable orders and income. We can explore a more flexible mechanism for mutual selection between drivers and passengers, design differentiated service models and revenue sharing ratios for peak hours, and achieve "warm matching". Returning algorithms to a tool centric approach, highlighting fairness and justice in rules, and freeing passengers and drivers from being trapped in data silos, is not only an important step in solving the difficulty of hailing a taxi during peak hours, but also a necessary path for the digital society to move towards refined governance. (New Society)