When a large model meets a mass-produced robot: the world may really only need 5% of people Once large-scale models and professional models from various industries are truly popularized, if Musk's Optimus and Yushu Technology's robots are really mass-produced and implemented on a large scale, a cruel but increasingly clear conclusion is emerging: 95% of humans in this world may really become "redundant" in the traditional sense. Speak clearly first. In the past, all technological revolutions - steam engines, electricity, computers, the Internet - replaced a certain kind of labor. But this time is different. The big model is systematically taking over cognitive labor: writing code, writing reports, designing, diagnosing, strategizing, making decisions, and creating ideas. Almost all white-collar jobs that require "brainstorming" have been compressed into a three-stage process of "questioning review execution". And professional models go further by directly encapsulating expert knowledge from vertical fields into callable tools. Doctors, lawyers, analysts, programmers, designers, teachers... these high cognitive positions that once required more than ten years of training are rapidly being downsized. At the same time, once humanoid robots achieve cost controlled mass production, the last line of defense against physical labor will also be breached. Musk repeatedly emphasized that Optimus' goal is to keep labor costs extremely low; Yushu Technology has already entered a commercial phase in quadruped and humanoid robots. When robots can walk, grab, collaborate, and work 24/7 without rest, positions that rely on human labor such as factories, warehouses, logistics, catering, cleaning, nursing, and agriculture will no longer require as many people. Cognitive labor is consumed by large models, while physical labor is consumed by robots. What's the rest? The truly scarce ones may only have three types of people left: People who can define problems and propose truly valuable directions (top decision-makers and strategists); People who can implement, maintain, and iterate AI and robot systems (system integration and advanced engineering talents); And those who create new experiences, emotional connections, art, and meaning (a very small number of truly irreplaceable creators). In total, it may only be around 5%. The remaining 95% will experience a sharp decline in marginal value contribution within the existing production system. It's not that these people are 'worthless', but rather that they have become dispensable in creating material wealth and routine services. This is not an exaggeration, but the result of technical logic deduction. Every productivity leap in history has resulted in some people being 'surplus' - it's just that this time the speed and breadth far exceed the past. Previously, it was a replacement that took decades to complete, but now it may be compressed to 5 to 10 years. Of course, reality will not turn into wasteland overnight. The mass production of robots still has hard thresholds in terms of engineering, cost, safety, and regulations; The large model is still far from being perfectly reliable. Society will have buffers, new professions, and policy offsets. But the direction is clear: the traditional survival model of "selling time and skills for wages" is being systematically dismantled. The real question is no longer whether AI will take over jobs, but how to redefine the distribution, meaning, dignity, and order of human society when 95% of people are no longer needed economically? This is where the big model and robot really exploded. It is not just about improving efficiency, but shaking the underlying assumption that 'people must prove their value through labor'.