Summary of Key Points
This article uses the metaphor of a "sheep eating people" land-grabbing movement to discuss whether the combination of AI and embodied intelligence (robots) will repeat history at the Shanghai World Artificial Intelligence Conference (WAIC). It suggests that in the short term, this could lead to job losses and the displacement of resources, but in the long run, it may drive industrial upgrading. The article points out that robots are currently at a stage where they are more for show than practical use; they are often used as interns to collect data rather than replacing workers. Additionally, the computational power required behind these technologies is quietly consuming public resources such as electricity and water, with much of the public investment flowing to a few tech companies. While ordinary people may bear the short-term costs, the technological revolution still has the potential to create new opportunities in the long run.
1. The Current State of Robots: Exciting, but Not Very Practical
The robots at the exhibition were more like performers—playing drums, playing the piano, dancing, and even folding clothes in servant attire, using sound, light, and electricity to attract attention. However, there were hidden issues: the dancing robots required remote control by staff, and the robotic arms used for folding clothes often failed to grasp the clothes properly. The robotic traffic police needed human assistance during peak hours. These examples show that robots are not yet capable of handling complex real-world situations independently, and the perfect performances in promotional videos are often scripted.
For instance, the robotic traffic police at intersections could provide warnings and simple gestures during off-peak times, but they would be overwhelmed in busy conditions, requiring human intervention. The dancing robots in the exhibition halls always had staff holding control handles, making it impossible for them to perform spontaneous dances. Even engineers admit that real-world scenarios are much more complex and require gradual improvement.
2. Robots in Factories: Not Working, but Collecting Data
It is often said that robots will replace workers in factories, such as screwing screws. However, using humans is actually more cost-effective. A skilled worker earns six to seven thousand yuan per month and can handle unexpected situations, while a robot costs hundreds of thousands of yuan and requires maintenance and repair, making the overall cost higher. So why do factories still use robots?
The answer lies in data collection. The success of large models relies on free online images and text data, but robots need specific physical interaction data (such as the force needed to screw screws or the feel when grasping parts), which is not readily available. Therefore, robots are used in production lines to collect this data while working. This is similar to students gaining experience through internships; factories provide the space, and manufacturers provide the robots, turning the production line into a "data mine," with increased capacity as a byproduct. Goldman Sachs' research also indicates that the industry is still in the stage of "concept validation to small-scale testing," far from replacing workers.
3. A New Form of "Sheep Eating People": The Quiet Displacement of Resources
The land-grabbing movement turned arable land into pastures, displacing farmers; today, it's about robots competing for resources, with ordinary people bearing the cost. For robots to become more intelligent, they need to be connected to the internet and use computational power, which in turn consumes electricity and water. Data centers consume a significant amount of water daily, and chip operation requires a lot of energy.
In the United States, some states have enacted laws to limit the construction of new data centers due to increased electricity prices and shortages of industrial water. In China, efforts like "East Data West Computing" (moving computing needs from energy-rich regions in the east to the west) are being made, but resource allocation issues persist, potentially leading to higher electricity costs and tighter water supplies for ordinary people.
There are also invisible resources being utilized: government subsidies, industrial funds, and venture capital are flowing into AI, but this industry primarily uses money rather than labor. Building data centers and developing robots requires hundreds of millions, yet only a few highly educated engineers are needed, unlike factories that employ thousands of workers. In the short term, the investment does not create many jobs for ordinary people, with benefits concentrated in the hands of a few tech companies.
4. The Two Sides of the Technological Revolution: Short-Term Pain and Long-Term Hope
The industrial revolution of the spinning jenny replaced hand weavers, leading to new industries and workers. The same is true with robots: there are short-term costs (such as job losses and resource displacement), but in the long run, it may create new opportunities.
Short-term costs include the replacement of some jobs (e.g., repetitive factory tasks) and the displacement of resources (electricity, water, public investment). However, technology will also create new demands, such as for robot maintenance, data annotation, and AI trainers, potentially changing lifestyles (just as the internet led to e-commerce and delivery services).
As the article concludes, many changes happen quietly, and by the time people realize them, society has already changed. The costs of a technological revolution are inevitable, but they are also essential for social progress.
5. What Should Ordinary People Do?
There's no need to panic or be overly optimistic. Robots are not yet powerful enough to completely replace human jobs (like in 20 years), but it's important to recognize that technology will change the job market. In the future, people will need to collaborate with robots, such as maintaining and repairing them, rather than just performing repetitive tasks.
It's also crucial to address the issue of resource distribution: can the benefits of technological progress be more fairly distributed to ordinary people? This could be achieved through policies that require tech companies to take on more social responsibility or by training people for new roles. After all, the ultimate goal of a technological revolution is to improve the lives of most people, not just a few.
This article does not doom robots nor overstate their potential; instead, it uses real observations and analogies to show us the dual nature of technological revolutions—both opportunities and challenges. Understanding these aspects helps us approach the future of AI and robots more rationally.