The 2025 Global Developer Pioneer Conference and International Embodied Intelligence Skills Competition opened on December 13. Yao Qizhi, an Academician of the Chinese Academy of Sciences, said that the development of embodied intelligence should move from imitation to reasoning. Current methods rely heavily on imitation and lack explainable world models and physical causal reasoning. In the future, reasoning, planning, and control should be placed in the same closed loop framework. The cost of physical data collection is high, and it is necessary to explore new data collection techniques, build a data factory that can be planned and scaled, and verify the large-scale path of embodied intelligence. In addition, embodied intelligent robots need to evolve from local skills to full-body collaboration. The movement ability and operation ability of existing robots are still two systems. In the future, full-body control and detailed hand operation planning will need to be unified to support multi-step general skills. He pointed out that the entire industry should go from individual combat to unified evaluation, establish open benchmarks and security standards, encourage open source reproduction, and promote excellent algorithms to be repeatable, verifiable, and industrialized through challenges.

The 2025 Global Developer Pioneer Conference and International Embodied Intelligence Skills Competition opened on December 13. Yao Qizhi, an Academician of the Chinese Academy of Sciences, said that the development of embodied intelligence should move from imitation to reasoning. Current methods rely heavily on imitation and lack explainable world models and physical causal reasoning. In the future, reasoning, planning, and control should be placed in the same closed loop framework. The cost of physical data collection is high, and it is necessary to explore new data collection techniques, build a data factory that can be planned and scaled, and verify the large-scale path of embodied intelligence. In addition, embodied intelligent robots need to evolve from local skills to full-body collaboration. The movement ability and operation ability of existing robots are still two systems. In the future, full-body control and detailed hand operation planning will need to be unified to support multi-step general skills. He pointed out that the entire industry should go from individual combat to unified evaluation, establish open benchmarks and security standards, encourage open source reproduction, and promote excellent algorithms to be repeatable, verifiable, and industrialized through challenges.