At the forum, Wang Xingxing, Chairman of Unitree Robotics, highlighted that while AI technology has been extensively utilized in programming and development for numerous years, its integration into the robotics sector has been comparatively limited. To bridge this gap, Unitree Robotics has recently dedicated itself to fostering the self-evolution of physical AI robot models. Specifically, harnessing the power of state-of-the-art AI large models, the company establishes rules, accumulates experiences, and utilizes constraint tools. This enables robots to autonomously seek out the latest research papers, findings, and open-source solutions, and subsequently generate machine control codes through self-programming. These generated control codes can be executed in simulated environments or deployed and tested by employing large models to control physical robots, with their performance evaluated and scored accordingly.