Wang Xingxing Predicts Five Core Resources of the Future; Turing Award Winner Diffie Says AI Will Sweep the Globe
Wang Xingxing, founder of Unitree Robotics, proposed at the 9th Hongqiao International Economic Forum that data, computing power, chip production capacity, energy, and minerals will become the five most core resources of the future world, and that competition around computing power will be "very, very fierce." He believes that multimodal models are not yet fully aligned with the input and output of the physical world, and that precision operations still have errors of a few millimeters; this bottleneck is expected to be resolved within one to two years. Whitfield Diffie, winner of the 2015 Turing Award, said that the relationship between humans and robots will become the most important issue of the 21st century.
Body · AI translation
On November 5, at the "Humanoid Robots and Embodied Intelligence Industry Innovation and Development Cooperation" sub-forum of the 9th Hongqiao International Economic Forum, Unitree Technology founder Wang Xingxing shared his latest thoughts on the future development of the robotics industry.
In his view, data, computing power, chip production capacity, energy and minerals will become the five most core resources of the future world, and the competition centered on computing power will be "very, very fierce."
At the same forum, Whitfield Diffie, winner of the 2015 Turing Award and a member of the US National Academy of Engineering, also shared his judgment on the development of artificial intelligence. He believes that AI will certainly sweep the globe in the future, but at the same time, the issue of "the relationship between humans and robots" will become the most important topic of the 21st century.
Wang Xingxing: Computing power competition will be very fierce
In Wang Xingxing's view, competition in the AI industry is gradually moving beyond models and algorithms themselves and extending to the underlying resources that support AI development.
He mentioned that data and computing power are currently the resources that the industry pays more attention to, but in fact, with the development of AI, demand for chip production capacity, energy and minerals will also grow.
"These five resources will become the most core assets of the future world," he said.
Among them, computing power deserves particular attention. Citing SpaceX as an example, Wang Xingxing said: "Represented by Elon Musk, his company is transforming toward a large computing power company, participating in global computing power competition through large-scale computing power production and the construction of computing power data centers."
Wang Xingxing believes that future competition around computing power will be "very, very fierce."
Speaking about technological progress in humanoid robots and embodied intelligence, Wang Xingxing believes that the key problem the industry needs to break through is no longer entirely whether models can understand instructions, but whether models can translate understanding into precise physical actions.
He gave an example: if the world's best multimodal model today is used directly for robot operation, with the robot planning and grasping a relatively large object, the success rate is already close to 100%. But if it involves more precise operations, such as picking up mechanical parts, the success rate will decline.
In his view, the core issue behind this is that multimodal models and the physical world's inputs and outputs have not achieved sufficient alignment. Even if the model has understood the task requirements, when ultimately implemented in the robot's specific actions, there may still be errors of a few millimeters. And for precision scenarios such as industrial manufacturing, such errors may directly affect whether a task can be completed.
"The last few millimeters still have relatively large deviations," Wang Xingxing said. He believes that if the alignment error between model input and output can be solved, embodied intelligence models will not have major problems at the technical level, and what remains is mostly engineering and commercialization promotion.
He expects this technical bottleneck to be well addressed within the next one to two years.
He introduced that some large-model companies have already formed an AI self-evolution loop internally, that is, using AI itself to collect data, complete labeling, carry out model training, then use AI to evaluate training results, and further optimize the model accordingly. In this process, the degree of human participation gradually decreases.
By contrast, robotics companies have made relatively slower progress in this field. Wang Xingxing envisions that in the future, AI models can write robot control code themselves, then train through real robots, and AI can evaluate indicators such as action completion and task success rates, thereby driving continuous model iteration.
As the number of deployed robots increases, companies will be able to obtain more training and testing data from real scenarios, further accelerating model evolution.
Turing Award winner: "Human-robot relations" will be the most important topic of this century
Compared with discussions at the technical and industrial levels, Turing Award winner Diffie is more concerned about the profound impact of AI development on human society.
He believes that AI will certainly sweep the globe in the future, but at the same time, people's concerns about AI safety are also increasing.
Diffie mentioned the "Three Laws of Robotics" proposed by science fiction writer Isaac Asimov. This idea first appeared in complete form in Asimov's 1942 short story "Runaround." The three laws are: a robot may not injure a human being, must obey human orders, and must protect its own existence as long as this does not violate the first two laws.
Diffie believes that although people today no longer fully understand and design robots according to such principles, humanity still needs to establish corresponding "guardrail mechanisms."
In his view, the most important topic of the 21st century will be "the issue of the relationship between humans and robots." He mentioned that global warming, geopolitical conflicts and other issues are all important problems facing the world today, but with the development of AI, "the issue of the relationship between humans and robots" may ultimately become a more decisive topic, and its importance will surpass other major issues.
Beyond technological breakthroughs, the humanoid robot industry is accelerating its search for commercial application scenarios.
At the sub-forum, the "Humanoid Robot Application Scenario Map" was released, covering ten major fields including scientific research and teaching, automobile manufacturing, 3C manufacturing, commerce and logistics, healthcare, elderly care and disability assistance, domestic services, operation and maintenance inspection, emergency rescue, and modern agriculture, showing that humanoid robot applications are extending from industrial manufacturing to more economic and social fields.
Yicai reporters also learned from the forum that in the first half of this year, national humanoid robot complete-machine output exceeded 40,000 units, and products are developing from "able to run and jump" toward "able to think and work," while accelerating penetration into manufacturing, services, livelihood and other fields. With support from the Ministry of Industry and Information Technology, the National and Local Joint Humanoid Robot Innovation Center is also accelerating construction in Shanghai to provide support for intelligent technology innovation and industrial commercialization.