In the AI era, will jobs disappear or transform? The debate between Elon Musk and Jensen Huang

Recently, at an investment forum held in Saudi Arabia, two tech leaders gave completely opposite answers to the same question: How will AI change the future of work.

One side believes that within the next ten to twenty years, AI and robots will fundamentally change the nature of labor. People will have the choice to work or not because technological progress will eliminate poverty and enable everyone to achieve financial freedom. The other side points out that the reality is more complex—AI will not make jobs disappear but will change the form of work, with increased efficiency leading to more labor-intensive tasks.

Clash of Ideals and Reality

Elon Musk used farming as a metaphor: just as people can choose to buy groceries at the supermarket or grow vegetables in their backyard, future work will also become optional. This vision is based on a premise—that technological progress will inevitably lead to widespread wealth distribution.

However, Jensen Huang provided a set of thought-provoking data. He pointed out that radiologists have not experienced mass unemployment due to AI; instead, their hiring demand has increased. The reason is not that jobs have not been replaced, but that the nature of the work has changed. AI has taken over standardized image reading tasks, allowing doctors to see more patients and handle more cases, resulting in an overall increase in workload rather than a decrease.

This phenomenon applies to other professions as well. After lawyers use AI to handle paperwork, they end up managing more cases; programmers use AI to write code, prompting companies to demand more features; designers find that posters that once took a week to hand-draw can now be generated in ten minutes by AI, but clients require twenty versions delivered in a week. Lawyer acceptance rates may not decrease; rather, the composition of lawyers’ work has changed—the same logic applies to other knowledge-intensive professions.

The Paradox of Efficiency

There is an underlying paradox here: increased efficiency should bring more free time, but the actual result is more intensive labor. Data from the U.S. Department of Labor supports this observation—unemployment in the tech industry surged significantly compared to the same period last year; ResumeBuilder surveyed 750 bosses using AI, and 44% said they had laid off employees in 2024 due to AI efficiency gains.

These statistics reveal a harsh reality: The wealth created by technology is not distributed equally. During the Industrial Revolution, people predicted machines would liberate humans, but the result was workers working sixteen hours a day; the digital age promised paperless offices would shorten working hours, yet the global population still needs 24/7 online email responses.

Asymmetric Distribution of Computing Power

Another key insight from the forum: six years ago, 90% of the world’s Top500 supercomputers used CPUs; now, that proportion has dropped to 15%, while the share of accelerated computing (GPUs) has soared from 10% to 90%. These seemingly technological progress figures actually represent a reshuffling of hundreds of billions of dollars in computing resources.

Elon Musk and Jensen Huang announced plans to build a 500 MW AI data center in the desert, reflecting not inclusive technological democratization but massive capital concentration. Those controlling these computing resources, owning the models, and occupying platform positions will be the true beneficiaries of the AI era.

The Eternal Scarcity of Resources

Elon Musk once said that with AI, money would become unimportant. But this statement overlooks a fundamental fact: Even if material production costs approach zero, resource scarcity will not disappear. Land, computing power, attention, and power—these are always limited.

When everyone can use AI to write papers, will admission rates at top universities increase? When everyone can use AI to start a business, will market share competition become milder? The answer is clearly no. Human societal competition has never been just about satisfying basic needs; it is more about the struggle for relative status.

Tesla aims to produce one billion humanoid robots annually, priced at $20,000 each. This means each robot corresponds to $20,000 in purchasing power. Who will buy the first one? Who can buy ten thousand? The answer is obvious—that’s a game for the wealthy and large organizations.

Shift in the Power of Defining Work

Jensen Huang mentioned that the past fifteen years have been the “recommendation system era,” driven by algorithms from social media to advertising. Now, this is shifting from CPU architectures to generative AI on GPUs. This change will not happen overnight but will gradually permeate. Before people realize it, the definition of work has already been rewritten—and who holds that power?

Elon Musk himself also demonstrates this. He says he will become busier because of AI, as he has too many ideas in his mind. For someone who controls computing power, models, and platforms, the meaning of “optional work” is entirely different from that of ordinary users who only use tools.

Warnings from the IMF

The International Monetary Fund (IMF) predicts that AI will impact nearly 40% of jobs worldwide, with 70% of occupational skills changing, and the impact in developed countries could reach 60%. This is not alarmist talk but a warning about the amplification of existing power structures.

Technology has never automatically brought equality; it only reinforces existing asymmetries. For those who control the technology, work becomes a hobby because their income derives from capital appreciation and ownership of technology; for most people, AI will only make work more unstable, fragmented, and more like a forced obligation to survive.

The truth of the AI era is: work will not disappear, but the power to define work has shifted. And this new definition of work is increasingly in the hands of fewer and fewer people.