Author: Zhang Feng
I. The New Yorker's Basic Viewpoint: "Artificial Intelligence Will Relegate the Most People to the Permanent Bottom"
A widely circulated article in The New Yorker magazine depicts a disturbing future scenario: with the rapid development of artificial intelligence, society will split into a very small "elite class" that masters AI technology and a huge "useless class," while the majority of people will be permanently relegated to the bottom of society. The core logic of this viewpoint can be summarized as follows: First, AI will replace a large number of white-collar and knowledge-based jobs. Unlike previous industrial revolutions that primarily replaced manual labor, artificial intelligence directly impacts cognitive labor, analytical judgment, and even a certain degree of creative work. Traditional middle-class professions such as lawyers, accountants, programmers, doctors, and teachers may be largely replaced by AI. Secondly, the speed of technological iteration far exceeds the speed of labor force transformation. Historically, the widespread adoption of the steam engine and electricity took decades or even centuries, while AI's capabilities experience qualitative leaps every few months. People don't have time to learn new skills before their existing skills become obsolete. Thirdly, capital's monopoly on technology will exacerbate inequality. Large companies that control AI technology and computing resources will become the new "feudal lords," while ordinary people will find no bargaining power in this system because AI is cheaper, more efficient, and more stable than anyone else. Fourth, the logic of "creating new jobs" fails. While past technological revolutions eliminated old jobs, they created many new ones. However, AI not only replaces physical labor but also mental labor. The newly created jobs will either be extremely high-end (only a very few people can fill them) or quickly swallowed up by AI. Ultimately, most people will lose their value in participating in the economic system, relying only on basic income to survive, becoming "pets feeding algorithms." This viewpoint is not alarmist; it has sparked profound anxiety among academics, the tech community, and policymakers. However, a closer examination of the essential characteristics of artificial intelligence reveals that *The New Yorker*'s conclusion is built on a fundamental misjudgment—it views AI as an external force replacing human brainpower, failing to recognize that AI is essentially the infrastructure for mental labor. II. The Rationality and Irrationality of *The New Yorker*'s Logic First, we must acknowledge the rationale behind *The New Yorker*'s viewpoint. There is ample evidence to support the claim that artificial intelligence will indeed have a dramatic impact on the job market. Large language models like GPT-4 have already demonstrated performance close to or exceeding that of ordinary professionals in tasks such as code generation, text writing, data analysis, and even legal consultation. A 2023 Goldman Sachs research report estimated that approximately two-thirds of jobs in Europe and the United States are exposed to the risks of AI automation, with one-quarter to one-half of these jobs potentially being performed directly by AI. Secondly, the speed of technological substitution is indeed unprecedented. During the Industrial Revolution, it took textile workers two generations to transition; while AI went from failing the Turing test to passing the bar exam in less than a decade. This exponential rate of change makes the traditional "retraining-job transition" model ineffective. Thirdly, the trend of wealth and power concentration is indeed worrying. A few companies, such as OpenAI, Google, and Microsoft, have established significant advantages in basic models, computing power, and data. Once this monopoly solidifies, ordinary people may indeed lose their voice in the economic system. However, *The New Yorker*'s logic contains a fundamental fallacy: it equates "artificial intelligence replacing a certain type of labor" with "the executors of that labor becoming useless." This assumption ignores the fact that the relationship between laborers and production technologies in the economic system is not a simple substitution relationship, but a complex restructuring one. The first flaw lies in the trap of "zero-sum thinking." Viewing AI as a competitor that will "steal" jobs from humans is itself an industrial-age mindset. In fact, every technological revolution has eliminated old professions, but it has also unleashed new demands and possibilities. Agricultural mechanization in the 19th century reduced agricultural employment in the US from 80% to less than 2%, but it didn't result in an 80% unemployment rate—instead, people shifted to manufacturing, services, and the now unimaginable field of "knowledge workers." AI will similarly create new professional fields that are difficult to imagine today. The second flaw lies in ignoring the diversity of human labor's value. The New Yorker's view implicitly assumes that economic value exists only in productive labor that can be measured by efficiency. However, human creativity, emotional connection, ethical judgment, aesthetic experience, community building, education, and many other activities cannot yet be efficiently and completely replaced by AI; they are precisely indispensable parts of human life and even the economic system. The higher the efficiency of AI, the more valuable these "inefficient but unique" human abilities will become. The third, and most important, flaw lies in the misinterpretation of the nature of AI. The New Yorker views AI as a kind of "superhuman intelligence," as if it were an independent entity capable of taking over all human mental work. However, true artificial intelligence is not "another kind of intelligence" at all, but rather "extracted and industrialized mental labor capabilities." To understand this, we need to delve into the essential characteristics of artificial intelligence. III. The Essence of Artificial Intelligence: The Infrastructure of Mental Labor An analogy: The Industrial Revolution was the infrastructureization of physical labor. To understand artificial intelligence, we need to go back to the Industrial Revolution. The Industrial Revolution wasn't some mysterious "machine age," but rather the industrialization of repetitive, mechanical, and manual labor. Before the Industrial Revolution, forging a shovel required a blacksmith's physical skills—the force, rhythm, and angle of the hammer swings—all accumulated over generations as "physical knowledge." The Industrial Revolution, through steam engines, stamping presses, and assembly lines, removed these repetitive, predictable physical movements from humans, standardizing, mechanizing, and scaling them up. From then on, forging skills that previously required ten years of apprenticeship could be mastered by a farmer operating a machine after two months of training. This isn't about "machines replacing humans," but rather "the ability to perform manual labor becoming an accessible infrastructure for everyone." You don't need to be a blacksmith; you just need to connect to the industrial system to achieve outputs far exceeding those of a blacksmith. The Industrial Revolution transformed "physical labor," once an extremely scarce individual skill, into a cheap and universally beneficial public resource. The consequence wasn't that workers became poorer; quite the opposite—the Industrial Revolution led to the first sustained and significant improvement in the living standards of ordinary people in human history. Because when the bottleneck of manual labor was broken, humans began to focus on organization, design, management, and innovation—work that truly requires unique human abilities. Artificial Intelligence: The Industrialization of General Repetitive and Mechanistic Mental Labor. Artificial intelligence is precisely the continuation of this logic in the field of mental labor. The essence of artificial intelligence is the industrialization of general, repetitive, and mechanistic mental labor. What is "general, repetitive, and mechanistic mental labor"? Let's break it down: It's not the genius-level creative mental labor of Einstein's discovery of relativity, but rather the standard problems encountered by ordinary professionals in their daily lives—writing a business email, organizing meeting minutes, translating a text, writing a standard sorting code, analyzing the basic trends of a set of financial data, and identifying common lesions in medical images. **Repetitiveness:** These tasks have clear patterns, and the processing methods are highly similar across a large number of cases. A doctor's judgment logic when reviewing 1000 CT scans is similar each time; a programmer's logical structure for writing 100 sorting functions is similar each time. **Mechanism-Based:** These tasks have clear rules, methods, and processes, which can be described using an "if-then" logical framework, or in other words, can be written as algorithms. The steps are definite, and there is a clear mapping relationship between inputs and outputs. This type of mental labor constitutes the majority of white-collar jobs in modern society. It requires professional knowledge, training, and thinking—but it is not the kind of cutting-edge creative breakthrough, nor is it the kind of work that requires deep emotional connection and complex situational judgment. Artificial intelligence, through large-scale pre-training, deep neural networks, reinforcement learning, and other technologies, "extracts" these mechanistic mental labors from the human brain, turning them into standardized, readily available, and near-zero marginal cost services. You don't need to study accounting, you don't need to memorize all the articles of tax law; you only need to describe your question to AI, and it can complete the tax calculations that previously required a professional accountant half an hour to complete. This is not "AI replacing humans" at all, but rather "the ability of mechanistic mental labor has become an infrastructure that everyone can access." Just as the Industrial Revolution gave everyone the "metallurgical strength" previously only possessed by blacksmiths, artificial intelligence is giving everyone the "computational brainpower" and "analytical brainpower" previously only available to professionals. Why will this give the majority more opportunities? Understanding the essence of AI helps us understand why it benefits the majority rather than oppresses them. First, AI significantly lowers the entry barriers for knowledge and professional skills. In the past, becoming a data analyst required learning statistics, programming languages, and database knowledge, investing hundreds of hours in training. Now, a marketer can directly ask AI in natural language: "Analyze our sales data from the past year and find out which product combinations are most frequently purchased together." AI can not only provide the answer but also explain its analytical methods. This means that professional knowledge itself is no longer a scarce resource; what is truly scarce is the "ability to ask the right questions" and the "ability to judge the quality of answers"—and these are precisely what ordinary people can gradually cultivate. Secondly, AI liberates humanity from repetitive mental labor. A doctor spends a significant amount of time each day writing medical records, reviewing routine imaging, and searching literature—this accounts for 70% of their working time, and these are precisely the mechanical, intellectually demanding tasks. When AI takes over these tasks, doctors can focus their energy on where human intervention is most needed: in-depth communication with patients, developing personalized treatment plans, and conducting innovative medical research. Doctors won't become fewer; instead, they will become more valuable—because they can now focus on aspects that AI cannot replace. Thirdly, the near-zero marginal cost of AI will democratize "high-end intellectual services." Previously, only large companies could afford top law firms, McKinsey consultants, and Goldman Sachs investment banks. Now, a small business owner can use AI to generate draft legal contracts, write business plans, and analyze financial statements. This isn't eliminating the market for these professionals, but rather expanding the overall market – as costs decrease, demand explodes, and professionals will actually gain more high-quality work in collaboration with AI. Fourth, AI enables a leap in personal productivity. What one person could do in the past was limited; now, one person working with AI can accomplish what used to require a small team. This won't cause unemployment; instead, it will spur countless micro-entrepreneurship and individual economic activities. One person can simultaneously be a product manager, designer, programmer, and marketer, because AI provides powerful support for routine tasks in these fields. Creativity, judgment, and a sense of responsibility—these core qualities have become more important than ever before, while the barriers to achieving them have been lowered like never before. IV. New Social Forms and Division of Labor in the Future When artificial intelligence becomes widespread as a fundamental infrastructure for cognitive processes, human society will enter a completely new organizational form. This is not a utopian fantasy, but a reasonable deduction based on current technological trends. The distribution of basic material needs on demand will become possible. With AI-driven productivity levels, the distribution of basic material needs on demand is no longer a pipe dream. We need to understand why this is happening: **Intelligent Production:** AI scheduling systems can achieve optimal configuration of raw material procurement, production planning, and logistics distribution, significantly reducing waste and inventory costs. In manufacturing, intelligent manufacturing systems can automatically adjust production lines the instant demand arises. **Energy Efficiency Revolution:** The application of AI in grid dispatching, energy consumption forecasting, and renewable energy integration will lead to a continuous decline in energy consumption per unit of GDP. As energy and computing power become cheaper, the "marginal cost" of producing materials begins to approach the cost of raw materials themselves. **Maturity of Automated Production Systems:** Combining AI control and robotics technology, a high degree of automation can be achieved throughout the entire process from raw materials to final products. This is similar to the "tap water" we're all accustomed to today—you don't need to know how the water plant works; you just turn on the tap and there's water, used as needed, and paid a very low fee based on usage. When the marginal cost of producing most basic necessities (food, basic clothing, standard housing modules, basic transportation, and common appliances) drops low enough, society is fully capable of achieving the distribution of basic necessities according to need. This is similar to the basic education and healthcare provided by Nordic countries today—not unlimited luxury, but a basic guarantee of a decent life. It's important to emphasize that "distribution according to need" is not the same as "distribution according to demand." It should be understood as a baseline of protection, upon which people can still obtain more resources, experiences, and recognition through their creative activities. Spiritual needs and creativity become core values. Once material constraints are overcome, what becomes scarce? Meaning, experience, creation, relationships, and aesthetics. These are precisely the areas where AI falls short—not that AI is completely incapable, but that no matter how well AI does it, it cannot replace the meaning of "human participation." Why do people attend live concerts instead of AI-generated perfect performances? Because "the performance of that specific person at that moment" itself has meaning. Why watch the Olympics? Because the process of real, flesh-and-blood people pushing their limits is moving. Why chat face-to-face with friends instead of chatting with AI? Because the other person is "another free, self-aware subject." These activities—artistic creation, scientific research (genuine cutting-edge exploration, not just literature reviews), education (especially the cultivation of values and aesthetics), community building, psychological therapy, sports, handicrafts, and philosophical discussions—will become major activities and sources of value in future society. The social division of labor will undergo the following transformation: From "finding a job" to "finding a mission": When material security is no longer an issue, people will choose to engage in activities not primarily for survival, but because they are meaningful, challenging, and bring flow experiences and self-actualization. From "Executor" to "Definitive Analyst, Evaluator, Integrator": AI can write code, but humans need to define "what software we want to write and what problem it wants to solve." AI can generate design solutions, but humans need to judge "whether this solution fits the project's character." AI can collect massive amounts of information, but humans need to integrate it into a compelling narrative. From "Efficiency Competition" to "Uniqueness Competition": You'll always lose to AI in terms of efficiency, but "my unique perspective, experience, emotions, and judgments" are something AI cannot replicate. In the future, an individual's core competitiveness will no longer be "how fast and accurate I can do this," but rather "why this must be done by me." This means that the class distinctions in future society will no longer be between "those with AI and those without AI," but rather between "those who can fully collaborate with AI to unleash their creativity" and "those who haven't yet learned this." The latter are not the lower class, but a potential group waiting to be liberated. This is precisely the mission of education. However, the aforementioned promising prospects will not automatically materialize. It depends on the development and governance of AI following the right path. If AI is monopolized by a few companies and becomes a new tool of privilege, then The New Yorker's prediction may indeed come true. Therefore, we need the coordinated development of a series of supporting technologies. Synergy with Web3: Preventing Value Monopolies. The core value of Web3 lies in its decentralized ownership and governance mechanisms. Combining AI with Web3 can avoid the monopoly of computing power, data, and models. Decentralized Computing Power Markets: Through blockchain technology, ordinary people can contribute their idle GPU computing power and receive token rewards, while large model training does not have to rely entirely on the data centers of a few companies. Although decentralized training currently faces technical challenges, the decentralization of computing power in the inference stage is already a feasible direction. Data Ownership and Proof of Contribution: The vast amounts of data generated by users when interacting with AI should have their value returned to the users themselves. Blockchain enables transparent tracking and value distribution of data contributions. Currently, people use data (such as publicly available text on the internet) to train large models for free. If in the future everyone can choose to contribute their optimized data from interacting with AI and receive rewards, then the evolution of AI becomes a process of universal participation and benefit. The protection and development of open-source models: Meta's Llama, Alibaba's Tongyi Qianwen open-source version, etc., demonstrate that high-performance AI models do not necessarily need to be closed. Web3's incentive mechanisms can provide continuous funding for open-source developers, avoiding a "winner-takes-all" scenario. Collaboration with quantum technology: Breaking the monopoly of computing power. The emergence of quantum computing may completely disrupt the current computing power landscape of AI. Quantum computing's parallel computing capabilities and exponential speedup for specific problems could free AI training tasks from relying on stacking more traditional chips. This has the potential to break down the computing power barriers currently built by a few companies like Nvidia and TSMC, enabling more research institutions, SMEs, and even individuals to train large models. More importantly, quantum key distribution and quantum random numbers can build truly secure and unpredictable AI systems, preventing the nightmare scenario of "super AI monitoring society." Digital governance: preventing the abuse of power in AI. Technology itself is neutral; governance determines its direction. The future AI society requires the following governance mechanisms: **Algorithm Transparency and Auditability:** Everyone has the right to know the basis of AI decisions, especially in scenarios involving their own interests (credit, employment, healthcare). This requires regulations mandating that AI systems provide explainable output. **Anti-Monopoly and Interoperability:** Large AI platforms should be required to open their interfaces to third-party developers, allowing users to switch between different AI service providers at low cost, preventing lock-in effects. Similar to number portability in the telecommunications sector and one-click data transfer services in the banking industry. **Digital Identity and Data Sovereignty:** Everyone possesses their own digital identity and data sovereignty; AI should only access relevant data with explicit authorization. This is not only for privacy protection but also a necessary means to prevent AI from developing spying and manipulation capabilities. Universal Basic Computing Resources (UBC). Similar to the concept of Universal Basic Income (UBI), future societies could allocate a certain amount of "free AI computing quota" to each individual daily—for example, allowing 100 questions to be asked of a large model per day, and free access to basic services such as speech synthesis and image generation. This ensures that even the poorest individuals are not excluded from AI infrastructure. Collaboration between Technology and Civilization. The synergy between AI and Web3, quantum technology, and digital governance essentially ensures that productivity tools do not deviate from a "human-centered" direction. What we need is not a super AI controlled by a few companies that disciplines everyone, but an open, auditable, low-barrier, and universally accessible intellectual infrastructure. Just like today's electricity system—anyone can plug it in and use it, and no one can enslave others by monopolizing electricity. The future of AI should be the same. The New Yorker's concerns are profound and alarming, reminding us that technology does not automatically bring justice. However, to conclude that AI will permanently relegate the majority to the bottom exposes a fundamental misunderstanding of AI's nature. Artificial intelligence is not another "super-intelligent species," but rather the industrialization of human beings' own mechanistic mental labor. It is infrastructure, a tool, an amplifier of capabilities. Its true historical significance is not replacing humans, but liberating humanity from repetitive mental labor, enabling everyone to achieve higher levels of creativity, judgment, and emotional connection at a lower cost. In the future, basic material needs will be met on demand by highly automated AI systems, and humanity will focus on spiritual creation and the pursuit of meaning like never before. What we should be wary of is not AI itself, but the possibility of AI being monopolized. Through the distributed governance of Web3, the breaking down of computing power barriers with quantum technology, and the ensuring of transparency and fairness through digital governance, we can forge a path of "human-machine collaboration, benefiting all." At every turning point in history, some have predicted that new technologies will destroy the opportunities of the majority. History has repeatedly shown that when technology is a true infrastructure rather than a shackle, the opportunities it creates far outweigh the jobs it eliminates. Artificial intelligence will not relegate most people to the bottom rungs of society permanently—on the contrary, it will give most people, for the first time, the opportunity to escape the burdens of survival and truly become the masters of their own lives and the creators of their own meaning.
Weatherly
Anais