Tencent Tang Daosheng: Harnessing engineering capabilities is a key factor for AI implementation

On March 27th, at the Tencent Cloud Shanghai Summit, Tang Daosheng, Senior Executive Vice President of Tencent Group and CEO of the Cloud and Smart Industry Group, stated that the implementation of AI is not just an algorithmic problem, but also an engineering challenge. With the same model capabilities, different harnesses—such as tool invocation, layered context engineering, long-term memory management, and workflow design—will all impact the actual effectiveness of AI implementation.

Tencent Cloud enhances model implementation capabilities from multiple dimensions. First is the Tencent Cloud Intelligent Agent Development Platform (ADP), which connects agents to a professional “library” through RAG and knowledge base capabilities, allowing industry experts to be always online and continuously accumulate the company’s operational know-how. Then there is Claw, which runs in the Agent Runtime’s secure sandbox. Claw serves as the neural hub of this intelligent system, discovering and downloading skills from the skill library to continuously learn and accumulate the ability to connect to external systems, using large models to send and receive commands and trigger actions.

This sandbox solution for the Agent Runtime can also be used for the verification of program results in large model reinforcement learning, capable of launching over 100,000 container sandboxes within one minute, with a startup speed in the hundred-millisecond range, destroying them once they are used, significantly enhancing the training efficiency of reinforcement learning.