Wang Xingxing: The core challenge of physical AI lies in aligning models with the physical world

At the 9th Hongqiao International Economic Forum, Wang Xingxing, founder and chairman of the humanoid robot company Yushu Technology, stated at the sub-forum "Innovation and Development Cooperation of Humanoid Robots and Embodied Intelligence Industry" that the current core challenge of physical AI mainly lies in the alignment between models and the physical world.

Wang Xingxing believes that the understanding capability of multimodal models is already quite good, and the biggest issue is the error in input-output alignment, including the deviation between sensor measurements at the input level and the actual state, as well as the deviation between action commands at the output level and actual movements. He predicts that this part of the problem may be solved in the next 1 to 2 years.