Ding DaoshiID: dingdaoshi789
Senior Internet observer, dedicated to research on the Chinese internet industry and enterprises.
On April 21, 2025, we published an article titled “From the ’72-Hour Network Survival Experiment’ to the ‘Yizhuang Robot Marathon’.”
I believe that despite the numerous issues faced during this robot marathon, its historical significance cannot be overstated. The participating robots, whether they fell or needed battery changes, reflect the determination of the Chinese technology community to break through challenges, even when they know it may not succeed.
After the article was published, there were many controversial comments and opinions, with a significant number of netizens believing that the robots’ performance in this competition did not meet expectations.
I also noticed that in recent days, some video bloggers compiled humorous videos of robots falling or making mistakes, which received a lot of likes and shares from netizens.
So, will the development of humanoid robots be overshadowed? Will these exposed issues put the brakes on the industry and significantly slow down its development?
In light of this, I was invited to the “Beijing News Radio” to continue discussing these hot topics and share some insights and opinions.
Here is a summary of the key points from the program, which I am now sharing with everyone (with additions).
—– This is the dividing line —–
1. Host: Teacher Ding, you were present at the humanoid robot marathon. Which scenes left a deep impression on you?
Ding Daoshi: To watch this competition, I canceled all my activities in advance, even my business trip. I arrived in Yizhuang early at 7 AM on the day of the event. Rather than focusing on the robots falling or their highlights discussed online, I want to share a few easily overlooked scenes.
The first scene was many middle and primary school students coming in groups with banners to visit. They continuously cheered for the robots on the field, gasping in amazement at the exciting moments. This reminded me of the many groups of students I saw when watching rocket and satellite launches in Jiuquan. I believe the best education should be like this—through on-site observation, allowing children to develop a basic understanding of robots. Perhaps among these children, there will be future scientists researching robots?
Another scene that impressed me was the many elderly people with gray hair standing by the sidelines, recording videos with their phones, waiting for the robots to pass by and watching them leave. Despite their unsteady steps, their eyes could not hide their excitement. For the older generation, it may be hard to imagine that our country has made such breakthroughs in technology, especially in the field of humanoid robots, which have fully reached an internationally advanced level. This excitement is the most straightforward affirmation of China’s technological progress.
These moments of observation made me realize that the significance of a competition has long surpassed the competition itself—it is the soil for sowing the seeds of science, a bridge connecting the past and the future, and a testament to the era of national technological rise.
2. Host: Regarding this competition, some netizens exclaimed: “My robot is amazing!” while others said the robots stumbled and fell, seeming not very high-tech. What do you think about the different interpretations and opinions from netizens?
Ding Daoshi: I have indeed seen these opinions, and I have read many voices from both sides.
I want to say that the historical or symbolic significance of this competition far exceeds its practical significance.
Think about it. The ENIAC computer was born in 1946, capable of only 5,000 addition operations per second, weighing 30 tons, requiring multiple people to operate, and consuming a lot of energy for cooling. Yet, it was this “clumsy” machine that opened a new era of human computation; the Wright brothers’ “Flyer One” flew for only 12 seconds, covering 36 meters, with poor maneuverability, yet it opened humanity’s longing for the sky and propelled the aviation industry from nothing to something.
Similarly, the Yizhuang Robot Marathon is the same; the robots on the field may have stumbled and had limited functions, still unable to meet the practical needs of large-scale industrial production or daily life. However, it has built an innovative competitive platform, gathering cutting-edge technology and creative ideas, showcasing humanity’s courage to explore the unknown in the field of robotics. This pioneering spirit, like Watt’s steam engine and the Wright brothers’ airplane, illuminates the path of robot technology development, and its historical and symbolic value will far exceed the practical results of the event and should be recorded in history.
3. Host: During the competition, there were indeed instances of robots falling, needing to change robots, and changing batteries. Do you think these are within an acceptable range? What data do these failures accumulate for future research and what experiences can the industry summarize from them?
Ding Daoshi: I believe it is completely within an acceptable range. Although there were many falls and malfunctions, some robots did reach the finish line. For example, the first-place Tiangong robot, I specifically observed it on-site, and after it ran past, there were still hundreds or even thousands of human competitors behind it. Isn’t that a form of progress?
Of course, the occurrence of failures is also a good thing; problems exposed in advance can be quickly improved and iterated. Tencent Technology recently published an article with a viewpoint I particularly agree with. They said, why let robots run marathons? Because the perfect conditions of the laboratory cannot simulate the complexity and uncontrollability of real road conditions. Only in a real environment over 20 kilometers can the joint structure, energy efficiency, heat dissipation capability, and motion algorithms be fully exposed, which is the first threshold for the humanoid robot industry to land.
4. Host: I also saw some people suggesting that for tasks like carrying things, robot dogs can do it, like the very popular Taishan robot dog, which makes carrying things very convenient; for cleaning rooms, we can use vacuum robots; we just need to create machines that can replace certain labor forces, so why do we need to create “humanoid robots”?
Ding Daoshi: Why “humanoid” and not another form?
Imagine this: the Earth has existed for billions of years, and humans have evolved for millions of years, with countless factors leading to the conclusion that the “human” form is more suitable for doing what needs to be done. Especially in household scenarios, actions like washing clothes, cooking, mopping, and organizing are more suited for a “humanoid” to perform.
Dogs, chickens, horses, and cows cannot do what humans can do.
The “human” form is naturally more suitable for doing these household chores or work. This is because, through the survival of the fittest in nature, humans with ten fingers can operate various tools more precisely, which is a crucial foundation for completing household and work tasks; those who can walk upright can easily carry items, operate machines, and even perform complex physical labor; the physiological curvature of the spine helps cushion the impact generated during upright walking and various activities, protecting the spinal cord and nerves from damage; the human brain has a high level of cognition, decision-making, and planning ability, which allows us to efficiently organize and manage household and work tasks and respond to various complex situations.
Understanding this logic clarifies why fixed-wing aircraft have developed for over 100 years but still generally resemble birds in appearance, as the fundamental principles remain unchanged.
5. Host: This year, we have frequently seen videos of humanoid robots doing flips, practicing martial arts, serving tea, and screwing in factory parts… various humanoid robots showcasing various human skills in different scenarios, which has raised high expectations. So, how far has the domestic humanoid robot industry developed? How far are we from having them “on the job” in our lives? [What are the key breakthroughs needed for humanoid robots to truly become our assistants?] [Is the cost of research and development a challenge behind application development?]
Ding Daoshi: Humanoid robots are the culmination of two major industrial categories: precision industrial manufacturing and AI.
Starting from the industrial level, future humanoid robots will have a body, limbs, and facial features similar to humans. The joints will be precisely manufactured to simulate the flexibility and load-bearing limits of human joints, allowing robots to move more human-like and carry items steadily; the internal wiring and power system layout should be compact and reasonable like human blood vessels and muscles, avoiding tangled wires and power imbalances, ensuring stable operation for long periods. For example, in hand manufacturing, it is necessary to replicate the skeletal, muscular, and ligament coordination of the human hand, allowing fingertips to sense the material and shape of objects, with grip strength control precise to the gram, such as being able to gently pick up an egg without breaking it or firmly grasp a tool without slipping.
Now, regarding AI technology, it is the key to injecting “soul” into robots, making them more than just a pile of silicon-based materials. First, perception ability is its “five senses” for interacting with the world. Visually, similar to ideal cars, it uses a combination of cameras and various radars, paired with advanced image recognition algorithms, to instantly recognize family faces and distinguish various household items, from kitchen utensils to bedroom clothes; hearing relies on sensitive microphones and voice recognition and semantic understanding modules, allowing it to hear calls from the bedroom even from the living room, accurately interpreting needs, whether it’s “help me get a book” or “set the alarm for tomorrow morning”; tactile sensors are spread across the body, providing real-time feedback on force and temperature when touching obstacles or grasping items, adjusting actions to avoid knocking over vases or hurting the owner.
Of course, we all know that dreaming is easy, but practice is difficult.
The widespread adoption of household humanoid robots still needs to overcome many challenges. Technical difficulties are like a thicket of thorns; anxiety over battery life needs to be resolved urgently, with breakthroughs in miniaturizing large-capacity batteries and fast wireless charging technology to prevent robots from running out of power midway through their “work”; cost control is like a tightening noose, with high-end components and research and development labor being quite expensive, requiring economies of scale and supply chain optimization to transform from luxury goods into consumer products; data privacy protection is like a solid shield, with deep collection of household life data requiring strict prevention of leaks, considering encrypted transmission, storage, and strict access authorization mechanisms.
Additionally, on the social acceptance front, ethical issues pose a significant challenge. For example, similar to the debate over who is responsible when autonomous vehicles have accidents, the accountability for robot errors is also difficult to define (it can’t all be blamed on the research and development company), requiring expert discussions and public participation to establish norms. At the same time, legal regulations need to be improved, from detailing product safety standards to accident compensation guidelines, to lay a solid legal foundation for industry development.
6. Host: The humanoid robot marathon has also sparked discussions about the robot industry, with two viewpoints emerging. One believes that the end of the humanoid robot half marathon signifies the beginning of the industry’s marathon, seeing the commercial landing of the industry; however, others believe that exploring the business model of humanoid robots remains fraught with difficulties. What do you think about the market prospects for humanoid robots?
Ding Daoshi: Although the robots have been ridiculed by netizens in recent days, we cannot throw the baby out with the bathwater.
As I mentioned in my previous article, a few years ago, humanoid robots only existed in science fiction. However, in the past two years, with the development of AI large models and advanced industrial manufacturing, the industry singularity for humanoid robots has arrived, significantly shortening the time for landing from once unattainable concepts. As an on-site observer, I felt that since this was the first competition, the performance of the robots had a significant gap compared to human competitors. However, I believe that in five or ten years, when similar robot competitions are held again, robots may surpass the vast majority of human competitors.
The true value of robots lies not in how fast they are, but in whether they can be widely applied in work, life, and production across various fields. For example, in the case of household robots, in the future, they may be able to take on tasks such as washing clothes, mopping, cooking, and organizing household chores. Based on today’s competition performance, this possibility is quite large. It is expected that within 3 to 5 years, household robots will likely enter homes on a small scale; in 10 to 20 years, they could be widely popularized at lower prices (such as under 100,000 yuan), providing services to the public.
As our country’s industrial manufacturing capabilities and artificial intelligence algorithm capabilities continue to improve, the application scenarios for robots will become even more diverse. Therefore, the industry needs more similar robot competitions to stimulate innovative vitality and promote industry development.
————————————————-My work “Exploring Rural Revitalization: The Digital Economy in Progress” is now available on e-commerce platforms and is selling well. Friends are welcome to click the bottom “Read the original text” to place an order and support.Ding Daoshi
-
In 2005, I first proposed the concept of “free media person,” which was later abbreviated to self-media, influencing to this day.
-
In 2011, Ding Daoshi joined Sutu Network, serving successively as the chief editor of Sutu column, executive editor of Sutu, and director of Sutu Research Institute.
-
In 2014, Ding Daoshi, as a representative of self-media, led the drafting of the “Self-Regulatory Norms of Chinese Self-Media,” sharing it at a symposium hosted by the Cyberspace Administration, which attracted widespread attention.
-
Currently, he is a columnist for mainstream technology media and news clients such as Penguin Account, Yidian Zixun, Baidu Baijia, Toutiao, iResearch Column, and Xueqiu Finance.
For cooperation, please contact WeChat:
dingdaoshi
Long press the QR code to recognize and follow ↑↑↑