Jensen Huang's latest interview before visiting China: Whether you are eliminated by AI depends on one thing, NVIDIA requires all employees to use AI

On July 13, 2025, an exclusive interview unexpectedly became the prelude to the "eve of the visit to China." In an exclusive interview with CNN's Fareed Zakaria GPS, NVIDIA CEO Jensen Huang made a calm yet piercing judgment in front of a global audience: If the world lacks new ideas, then the productivity gains brought by AI will translate into unemployment. It’s not about models, chips, or blockades—what he fears most is whether we humans still "have ideas" in the AI era. After this statement, he was about to embark on his trip to China—according to Reuters, Jensen Huang will attend a media event in Beijing on July 16. This is his second visit to China this year and his first public appearance since NVIDIA's historic market value surpassed $4 trillion. This conversation is more penetrating than any financial report. He discussed AI-related unemployment, questioning abilities, the relationship between the American tech stack and Chinese developers, and even stated: "The programming language of future computers is English, or any other natural language of your choice." This is a period of accelerated industrialization of AI and the opening moment of the "cognitive elimination race." For developers, entrepreneurs, and workers, a real and harsh question emerges: Who will be the first to be eliminated by AI? It’s not about lacking ability, but having too few ideas. This article will outline the core content of this interview through "documentary × disassembly × insight," understanding a super CEO's real judgment about the future. After reading, you will gain a reference framework to examine your own 'creative ability' and see the deeper logic of NVIDIA's interaction with the Chinese AI ecosystem. Section One | Who Will Be Eliminated? Jensen Huang Warns of "Creative Drought" In this CNN dialogue, Jensen Huang did not repeat the overused AI buzzwords. He succinctly highlighted the focus of humanity's future with one sentence: "If the world lacks new ideas, productivity gains will directly turn into unemployment." Host Fareed Zakaria was asking whether AI would lead to large-scale white-collar unemployment, especially in roles like accounting, law, and consulting that follow processes. But Jensen Huang did not directly answer "who will lose their jobs"; instead, he countered with, "Are there still people in this world with new ideas?" The weight of this statement is heavier than "how many jobs AI will take away." Because he is not talking about job disappearance, but about the disappearance of creativity. The core issue is not unemployment, but the exhaustion of ideas. Jensen Huang's logic is straightforward: If society still has a large number of new ideas—such as faster logistics systems, smarter health diagnostics, and more efficient drug discovery— Then the arrival of AI will help us turn these ideas into reality, leading to more and richer jobs; But if everyone is just repeating old tasks, copying others' paths, and waiting for the system to feed answers, then AI can complete these processes in just a few seconds It's not about whether AI will replace you, but whether you have something worth it to help you with. Jensen Huang did not list a "dangerous profession checklist," but he provided a "portrait of the eliminated": Those who can only execute tasks but cannot design them; Those who do not think about the purpose and only wait for the system to output results; Those who rely on "asking others" rather than "asking themselves" for all their work. Conversely, he said: "AI is a technology equalizer that elevates ordinary people who do not understand technology—provided you have something for it to help you with." AI is not an "upgrade of tools," but rather a test of whether you have an independent sense of direction. Today, AI can quickly write copy, code, create charts, and list plans... what it lacks the least is the ability to get things done. The real dividing line is no longer "can you use it," but whether you can tell it what you want to do, why you want to do it, and what constitutes a job well done. This is not a matter of ability, but rather a question of ideas. In other words: "AI will not cut you out; AI merely accelerates the speed at which 'those without ideas' are replaced." This is the true meaning of "creative drought." It does not mean that everyone must become highly creative, but rather—do you have a way to make AI work for you, rather than passively waiting to be replaced? Section Two | Jensen Huang's Visit to China: Exporting Technological Views, Not Just Bringing Chips On July 16, 2025, Jensen Huang will hold a media meeting in Beijing. This is his second visit to China this year and the first time he appears as a "visitor" after NVIDIA topped the global market value. The outside world is generally concerned about whether he will announce "AI chips customized for the Chinese market," but from his statements in the CNN interview, what is more important than chips is the technological concept he wants to convey: We want the American tech stack to be the global standard. (We want the American tech stack to be the global standard). Redefining the AI Game from a "Developer's Perspective" In the CNN interview, the host directly posed a sensitive question: The U.S. has set more and more export restrictions on China's AI chips, You said this strategy is 'counterproductive,' what do you mean? Jensen Huang did not evade the question; he said: Depriving someone of technology is not the goal, just a means. Moreover, this approach will not help us achieve the real goal. So what is the real goal? His answer: The U.S. should become the leader in AI technology; Just as the U.S. dollar is a global currency, the American technology system should also become the "universal grammar" of global consensus; And this relies not on restricting others, but on attracting the world's best developers. He mentioned a fact that very few Western CEOs openly acknowledge: 50% of the world's AI developers are in China. This is not a politically correct statement, but rather an industrial reality. NVIDIA is not simply looking for "shipment" or "moat," but a deeper strategy: → To have global developers build content and create value on the AI technology platform in the United States. → To prioritize the "standard of technical language" to operate on this American framework. Platform Diplomacy Beyond Technology From CUDA, GPU to AI frameworks, NVIDIA aims to establish a globally unified technical foundation, and China is an indispensable part of it. His logic is— If Chinese developers are using American development tools and relying on NVIDIA's computing ecosystem, then regardless of where the model is trained or deployed, the underlying dependency ultimately returns to American standards; This is more long-term and effective than a sales ban. As he said: "We hope the best AI systems in the world can run best on American technology platforms." His visit to China is largely to maintain the influence of the most critical part of the technology stack. Because today's AI is no longer a hardware battle, but a competition of 'who defines the next generation of infrastructure.' And China is the most important gathering place for developers in this competition, a testing ground for real-world scenarios, and the market with the fastest industrial feedback. Summary Jensen Huang's visit to China, rather than being a simple market action or symbolic gesture, is more of a typical 'output of technological perspective.' He said: I want AI to run on the best technology platform. Not to force others to accept it, but to make others willing to rely on it. In the current complex technological relationship between China and the United States, he is one of the few CEOs still building bridges rather than tearing them down. And the bridge he is building does not lead to any specific vendor, but to a group of developers who are still tuning models at night and modifying code during the day. Section Three | The Hard Currency of the AI Era: Can You Ask Questions? In the middle of this exclusive interview, Jensen Huang was asked a very practical question: "Many studies now point out that people who use AI too much tend to become 'lazy'—they think less and rely more on the system. What do you think?" Jensen Huang did not directly deny the research conclusion, but he answered from a different angle: "I use AI every day. I don't think my thinking ability has diminished; in fact, it has strengthened. Because I am not letting it think for me, but rather letting it teach me things I don't know." He said the real key is not whether you use AI, but how you interact with it. He mentioned a severely underestimated ability: the ability to ask questions. When you interact with AI, you don't just throw out a random statement. You ask it questions and then continue to probe, like a doctor issuing a second or third round of test orders. You might even take an AI's answer and give it to another AI to point out errors, and then go back to the first AI and ask: 'Are you sure this is the best answer you can provide?' It sounds like "putting AI on the spot," but what he is talking about is a shift in future work habits: not inputting a question and waiting for an answer, but continuously asking questions to deepen, clarify, and reason out the problem He also mentioned a detail: the most important job of a CEO is actually to ask questions. — "90% of my work is actually about asking questions. Many of my instructions and judgments are actually hidden in one question after another." This statement seems understated, but it is actually the most powerful one in this conversation. The real gap opened up in the era of AI is not in technology or education, but in whether a person has the ability to ask high-quality questions. Why is "being able to ask questions" more important than "being able to operate"? Because today's AI tools are not lacking in execution capability—if you say "write a meeting minutes," "draw a flowchart," or "list a job posting," it can complete these tasks quickly. But what it is not good at is discerning which approach is more suitable for you, which goal is better, and which expression is more moving. In other words, AI is strong, but it does not know which task to perform. So, if you do not even know what answer you are looking for or what problem you are trying to solve, then AI cannot help you. At this point, the ability to ask questions becomes the bridge connecting you and AI. In the past, a person's intelligence was often judged by whether they could answer questions. Now, a person's capability is instead judged by whether they can ask the right and thorough questions. For example— Instead of asking: Can this project be done? ask: If we do it, can we reduce labor input by 30%? Instead of asking: How is this content written? ask: Is there another angle that could better resonate with the other person's emotions? AI can give you answers, but truly advanced questions can only be asked by yourself. Here, Jensen Huang makes it very clear: interacting with AI is essentially a 'question-and-answer process.' The deeper you ask, the more accurately AI will respond. This means that "asking questions" has become a core competency for future work. Section Four | Programming Without Code: The Future Language is English After learning to ask questions in Section Three, the next question is—how do you get AI to execute it? Many people think that AI is a programmer's job. But Jensen Huang made a surprising statement in the interview: The programming language of the future is English. Or any natural language you choose. In other words— You do not need to understand code to be qualified to use AI; You just need to articulate things clearly to mobilize AI. A real scenario: NVIDIA has already implemented "everyone using AI" internally. Under the lens of CNN, Jensen Huang casually stated a sentence, but it carries a lot of information: I encourage all employees to use AI, to the extent that it is even mandatory. It is not about encouraging engineers to try it out or having internal small-scale initiatives, but rather that the entire NVIDIA, from software engineers to chip designers, must integrate AI into their workflows. He stated: "100% of our software engineers and 100% of our chip designers are already using AI." This is not a conceptual "empowerment," but a "basic action" at the institutional level of the company. Want to write code? Ask AI first; Need to check design specifications? Ask AI; Finished the first version of the plan? Throw it to AI for optimization; All employees must first let AI provide references, structures, validations, and alternatives before moving forward. This is not a question of whether to use AI, but if you are not proficient in using AI, you cannot advance your work in this position. Why is "knowing how to use AI" not just a privilege for programmers? Because today's AI can already understand human language. You don't need to write complex functions; as long as you can accurately describe the goals, constraints, and expected results, AI can automatically generate code, modify drawings, check for vulnerabilities, and write copy. The change behind this is not technology, but the way people connect with technology: From "code translator" to "collaborative partner"; From "operating system" to "dialogue interface". This also means: the clearer your thinking, the more useful AI becomes; the better you can describe, the more it can help you implement. NVIDIA's real practices illustrate a trend: the boundaries between 'technical positions' and 'non-technical positions' in the future workplace will become increasingly blurred. Those who understand products can quickly generate demos through AI as long as they can articulate the requirements clearly; Salespeople can automatically write customized proposals with AI as long as they can list customer concerns; Content creators do not need to write prompt tutorials; as long as they express themselves adequately, they can generate structured content. AI is becoming everyone's "second brain," and what you need to do is master how to engage it. Because the future programming language is the language you speak. Section 5 | AI Begins to "Act": From Video to Reality of Picking Up a Cup AI gives the impression that it can only talk, write, and draw, being a "tool in the screen." But Jensen Huang mentioned something that seems small but is fundamentally significant in an interview with CNN: If you can generate a video of a person picking up a cup, then you can also generate the action of a robot picking up a cup. This is not just casual talk, but a technical judgment: → AI not only "generates content," → it is beginning to "execute actions." A brand new model capability: a closed loop from language to action. Today you can tell AI: "Help me write an apology letter," "Generate a 3D image of an office," "Give me a video: a person picking up a cup." What's next? If you can generate a video, it indicates that AI can understand spatial, action, and temporal relationships. Therefore, it can also "delegate this action to a robot" for actual execution. This is what Jensen Huang means by: "AI is entering the physical world." In other words: Previously, AI could "speak," but could not "act"; now, it is beginning to "do what it speaks." He said: What we see now is an art of possibilities. If you can describe an action, you can generate that action. This is not a hypothesis, but a reality in progress. The Veo 3 released by Google can already generate very realistic "human action videos" based on textual descriptions. You say, "A person picks up a transparent water cup in the kitchen," and Veo will reproduce the angle of the hand, the ripples of the water, and the reflections of the light. These models understand not just the images, but the underlying rules of the actions. And this capability, once combined with robots, will move from "generating videos" to "controlling robotic arms." AI's "next leap" — from the world of information into the real world. He said: If AI can understand images and comprehend language, then it can understand proteins, chemistry, and operations. We can teach AI not just nouns, but also the meanings of these words and how they interact. This means that AI is not just an assistant, but is becoming a junior researcher, technician, and even an engineering collaborator. Applications are much closer than we imagine. In medical laboratories, AI can control instruments to complete cell scans; In factories, AI-driven robots are starting to perform precision screw tightening; In logistics warehouses, AI is not just scheduling routes, but is personally "walking into the shelves" to execute grabbing tasks. This is no longer a matter of understanding problems, but of executing them. From inputting a sentence → generating action steps → directing machines to execute, the entire chain is basically usable, with just a few industrial interfaces and a lot of scenario training left to complete. Summary Jensen Huang said: "Today, this technology already exists and is already in operation. Within three to five years, you will see a large number of such robots entering reality." AI is no longer just a "generating tool," but is becoming a "collaborator in reality." It is jumping off the screen, picking up cups, opening doors, and taking over repetitive actions in the real world. And this is precisely what Jensen Huang refers to — the next stage of AI is "physical intelligence": it is no longer just giving you suggestions, but is starting to complete tasks for you. Conclusion | Who will be eliminated? Not AI In this interview, Jensen Huang neither presented technical data nor showcased product blueprints. Every word he said actually drew a clear line: Those with ideas use AI to get things done; Those without ideas will be eliminated by AI. He avoided specific career advice and focused on three questions: Be clear about what you want to do: AI does not judge direction; it only amplifies your existing goals; Learn to ask the right questions: What truly makes a difference is the quality of the questions; Start using it now: Whether you are an engineer, designer, salesperson, or student, AI is already waiting for you to give it tasks. "AI is the equalizer of technology." "It will not widen the gap, but it will expose who never thinks." This is Jensen Huang's answer, and it is also a reflection for each of us: Don't wait for the industry to change, for positions to disappear, for elimination to happen, while still asking how to learn AI It's not about what you have mastered, but whether you are using it—participating in the construction of the future. Now is the starting point. Source: AI Deep Research Institute, original title: "4 Trillion CEO Jensen Huang CNN Warns of 'Creativity Crisis': Who Will Be the First to Be Eliminated by AI?" Risk Warning and Disclaimer The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this is at your own risk