August 9 News.Yushu Technology CEO Wang Xingxingrevealed at today's World Robotics Congress that he onceoppose most stronglyHumanoid robots, which the company has not launched at all by 2022.

But it was in 2022, ChatGPT represented the significant progress of AI technology, as well as Musk and other high-profile for cutting-edge technology, when "some customers directly find Yu Shu" and are willing to give money to the order.lead directly toYuuki started making humanoid robots.
Speaking about the industry, he said that the robotics industry is currently focusing too much on data and too little on modeling, and that the biggest problem right now is actually theModeling issues, current robot model architectures are not good enough or uniform enough to be used even though there is a lot of better data to train on.
Wang Xingxing has stated thatThe ultimate goal of Yushu Technology is to have robots free up human productivity to do hard work, "Yushu robots go dancing and fighting is the company's hope to make robots, especially humanoid robots, do all kinds of full-body movements through AI technology. Before the ultimate goal is not realized, we participate in some performances and fighting competitions to show people the real situation and generate part of the commercial value".
According to Wang Xingxing, "Our company has always been a great believer in theUniversal Robots with Universal AI will be the mainstream route for future tech.. In recent years, we have noticed that humanoid robots are very hot at home and abroad, and we have invested a lot of energy in their development. Undoubtedly, the industry is currently in a relatively early stage, we can give some more patience."
According to 1AI previously reported, Wang Xingxing also highly appreciated the ability of AI at the 2025 World Conference on Artificial Intelligence in July this year: let's say I want to generate a lottery software, it can help you automatically generate (code) in all aspects, and the current success rate is very high, basically it can reach ninety percent or so. "Now I'm lazy too, and I'm too lazy to write a lot of code text, so why don't I just tell a big model that I need a feature and it can automatically write it for me."He added that last year there had been a poor success rate in trying, with a lot of errors inside the written code, but by the first half of this year the success rate was really, really high.