Jensen Huang Resigns as NVIDIA CEO: Technology Can Create the Future, So Why Does Humanity Keep Repeating the Past?
On November 12, 2026, NVIDIA announced that Jensen Huang had resigned from the positions of CEO, president, and director, and the company he founded and led for thirty-three years began searching for its next leader. Drawing on this event and on Oppenheimer's experience of losing his security clearance in 1954, which was only formally revoked in 2022, the article points out that technology creators and the institutions that decide on technology are not the same kind of power, and that correcting errors can be delayed for decades. The author argues that the real bottleneck for AI is authorization and oversight, that the growth of intelligence does not automatically bring justice, and that more powerful intelligence may also strengthen surveillance and the concentration of power.
Topics
Body · AI translation
**Key points**
**1. Technological leaps have not eliminated the difficulties of human coordination**
**2. Institutional power and technological influence are not equal**
Oppenheimer lost his security clearance, and it was only formally revoked many years later, showing that the creators of technology and the institutions that decide on technology are not the same kind of power, and that correcting errors can be delayed for decades.
**3. The more technology becomes infrastructure, the harder it is to avoid governance**
Once technology carries the interests and risks of others, society will demand that it be held accountable; both the freedom to innovate and social constraints have legitimacy, and the conflict has no permanent correct answer.
**4. The real bottleneck for AI is authorization and oversight**
**5. The growth of intelligence does not automatically bring justice**
AI can help with transparent analysis and error correction, but it can also strengthen surveillance, manipulation and the concentration of power; the progress of civilization depends not only on what is created, but also on whether power can be constrained, errors faced, and people can live together. 2026-11-14 11:32 We are already able to create the future, so why are we still repeating the past? AIGC from 0 to 1 quick overview
This article comes from the WeChat public account: AIGC from 0 to 1, author: Wang Lingyi
On November 12, 2026, Nvidia released a not-very-long announcement.
Jensen Huang resigned from the positions of CEO, president and director. The company he founded and led for thirty-three years began looking for its next helmsman.
The wording of the announcement was calm. It thanked him for past contributions, announced new arrangements, and expressed confidence in the future. For a listed company, everything needs to continue operating according to established procedures.
But behind these few lines of text is an extremely unusual contrast of the times.
Not long ago, Jensen Huang was still one of the most symbolic figures in the global artificial intelligence wave. He and his company had almost become a way for this generation to understand the future.
From graphics computing to CUDA, from data centers to large-model training, behind humanity's continuously expanding computing power there has always been Nvidia. When people talk about the next generation of artificial intelligence, they often cannot avoid this company; when they talk about this company, it is hard to avoid Jensen Huang.
Yet even such a person, who almost participated in defining the future world of computing, still could not alone decide his own fate in this company.
Of course, regarding his departure, public information is not yet sufficient to reconstruct the entire decision-making process. Personal conduct, board judgment, company risk, and the broader social environment cannot be lightly reduced to a political persecution. Technological achievements also cannot become a reason for personal conduct to be exempt from scrutiny.
But even if we temporarily set aside these still-unclear causes and effects, this matter still leaves people with a certain indescribable sense of loss.
A person can spend more than thirty years changing the world, yet cannot thereby obtain unlimited control over his own situation.
Technology can create such enormous power, but it cannot guarantee that the people who create these powers always have the power to control them.
The inspiration comes from Qu Yuan's "Heavenly Questions" and the Southern Song Suzhou stone-carved "Astronomical Chart." The ancients looked up at the stars and asked about heaven, earth and the human world; people today create intelligence and are still asking about the direction of civilization. Over a thousand years, we have changed our ability to understand the world, yet we are still learning how to live together.
01
Over the past more than two hundred years, humanity has been accustomed to understanding progress on an ever-rising scale.
The steam engine extended human physical strength, electricity reshaped production and life, and computers handed increasingly complex mental labor to machines. In the age of artificial intelligence, we have even begun trying to give machines a certain degree of autonomous judgment, planning and action capability.
Many boundaries once considered insurmountable are being rewritten one by one.
The history of technological development is almost a history of humanity continuously escaping limits.
Distance once limited our movement, so we invented railways, airplanes and the internet.
Our bodies limit our strength, so we built machines.
Knowledge and calculation limit our judgment, so we created computers and began creating artificial intelligence.
Every breakthrough enables humanity to do what was previously impossible.
But when we turn back to examine the world between people, we see another picture.
Trust can still break down. Power still has to be contested. Interests are still difficult to coordinate. Within an organization, the most capable person does not necessarily have the final decision-making power; a seemingly rational choice does not necessarily bring good results.
Technology can help us predict the weather more accurately, but it cannot make everyone agree on who should bear the losses from a storm.
It can help us understand disease, but it cannot automatically decide how medical resources should be allocated.
It may even help us explore the origins of the universe, but it cannot answer by calculation alone: when different people's wishes conflict, who should give way?
None of this has disappeared with technological progress.
Thus, a strange temporal dislocation appears.
Technologically, we are getting closer and closer to the future, but in human relationships, we still keep encountering the past.
This is probably what is truly sad about this event.
Not because some great figure suffered a setback, but because we suddenly realize: it turns out that humanity's brilliant technological progress has not freed us from those older conditions of existence.
We can make tools that change the world, yet we still need to live amid mutual suspicion, dependence, competition, cooperation and checks and balances, an indescribable sorrow.
For the first time, humanity sent an object of its own making into Earth orbit. That species, which had long been able only to look up at the starry sky, finally let its creation enter the world of celestial motion.
The following year, Hannah Arendt published The Human Condition.
She did not merely praise this technological achievement; instead, she used it to ask: when humanity gains the ability to change its own conditions of existence, are we ready to decide how these abilities should actually be used?
This is a more difficult question than technological success itself.
Because technology can tell us how to achieve a goal, but it cannot by itself decide which goals are worth pursuing.
More importantly, human action never takes place in a vacuum.
We live in a network of relationships made up of countless other people. One person makes a decision, others respond; these responses in turn give rise to new actions. The final outcome often exceeds the expectations of the original actor.
This means that even a person with extremely powerful abilities cannot fully grasp all the consequences their actions set off in society.
A person's ability can approach some kind of technological extreme indefinitely, but he is always not the only actor in this world.
More than half a century ago, Oppenheimer also faced a huge gap between technological power and institutional power.
He participated in leading the development of humanity's first atomic bomb, yet in 1954 he lost his security clearance from the U.S. government.
Decades later, in 2022, the U.S. Department of Energy formally revoked that year's decision, acknowledging that the original procedures had serious flaws and injustices.
History did not become simpler because of this.
Oppenheimer's experience at least reminds us: the creators of technology and the institutions that decide how technology is used are not the same kind of power. Institutions can constrain individuals, but they may also make mistakes when exercising such constraint.
More unsettling is that some mistakes were regarded as necessary when they occurred, and by the time they received formal correction, decades had passed.
Technological progress can be measured in ten years, five years, or even a few months.
Yet for an institution to learn to acknowledge its own mistakes, it sometimes takes a generation.
The distance between these two speeds is precisely one of the difficulties that modern civilization has still not resolved.
03
However, if we merely conclude that "technology changes, human nature does not," we are probably still being too hasty.
Humans do retain many ancient psychological tendencies.
We fear loss, crave security, care about fairness and status, and become defensive when our interests are threatened. Technology will not automatically erase these behavioral tendencies just because computing power increases.
But this does not mean that humanity is doomed to repeat the same outcomes.
The same fear can produce different behaviors in different social environments.
When power is unchecked, information cannot be verified, and no one bears responsibility for mistakes, fear easily turns into suspicion and suppression.
When rules are clear, procedures are fair, and responsibility can be traced, people may also choose cooperation in the face of uncertainty.
So the repetitions in human history are not necessarily only because human nature has not changed.
It may also be because, in different eras, we repeatedly built certain similar power structures, faced similar conflicts of interest, and made similar choices.
If that is so, the problem is no longer just how to change human nature.
Rather, it is whether we can change those social conditions that continually provoke similar behaviors.
This is what makes governance such an important yet often underestimated topic.
Engineers bring new possibilities into the world, while laws, boards of directors and regulatory agencies are responsible for judging how far these possibilities can go.
But the real world does not have such clear dividing lines.
When a technology is small in scale, it may simply be the creator's own product. As it gradually becomes infrastructure on which the whole society depends, it begins to enter the lives of countless people, and also enter the risks of countless people.
Developers depend on it, enterprises depend on it, consumers depend on it, and capital and public institutions also begin to depend on it.
The creator gains unprecedented influence, and at the same time loses the possibility of deciding all consequences alone.
The more successful the technology, the harder it is for it to belong only to the person who created it.
Because once a technological system carries other people's livelihoods, investments, public safety and future, society will naturally demand that it be accountable to more people.
But this also cannot prove that every governance action is correct.
Governors may also fear, may also misjudge, may also be driven by interests, and may also, in order to avoid taking responsibility themselves, choose the measures that are easiest to implement but not necessarily the most reasonable.
Thus, we encounter an even more difficult situation.
Innovators need freedom, because discoveries that truly matter often come from breaking existing rules.
Society needs constraints, because no individual can, by virtue of their own excellence, demand that all others bear risks unconditionally.
There may even be situations like this: each party believes it is doing its duty, each party has reasons worthy of serious consideration, and yet after these reasons interact, they produce an outcome almost no one wants to see.
The deepest difficulty of civilization may not be that humanity does not know what is right, but that many right things cannot always be realized at the same time.
If this is the situation humanity must face, what exactly does the arrival of artificial intelligence mean?
It can write code, analyze data, find patterns, and call tools. More autonomous Agent systems may also continuously execute tasks, coordinate multiple systems, and even participate in the research and development of new technologies.
Some skills that humans previously needed years to accumulate may in the future be called upon in an extremely short time.
This change is enough to reshape a large number of industries.
But the truly profound change may not have begun yet.
Because as machines become more and more capable of acting on our behalf, a problem that could previously be temporarily obscured by technological progress will become unavoidable.
Who has the right to make it act?
It can analyze a patient's condition, but that does not mean it can decide the treatment plan on its own.
It can coordinate a company's vast resources, but that does not mean it can decide by itself whose interests should be sacrificed.
It can predict the consequences of a social policy, but that does not mean it has the right to decide what goals society should pursue.
As machines' capacity for action expands, these questions will only become more important.
Rather, humanity has not yet built sufficiently mature mechanisms to decide how to authorize these capabilities, how to supervise them, and how to handle the consequences they cause.
In the past, we used technology to continuously expand what we could do.
In the future, we may have to expend equally enormous effort to reinvent the institutions that decide what should be done.
This brings about a possibility that is worth earnestly anticipating and, at the same time, must be treated with vigilance.
Can those misjudgments that arose in the past from insufficient information be discovered earlier?
Can those problems that cannot be coordinated because there are too many participants and the interests are too complex be better analyzed with the help of new intelligent systems?
Can those decision-making processes hidden within organizational hierarchies and difficult to trace be made more transparent?
Even more, is it possible for those major decisions that in the past had to rely on individual authority and experience to be placed into procedures that are more open, more verifiable, and easier to correct?
These changes are not impossible.
More powerful intelligence can also bring more efficient surveillance, more refined manipulation, more concealed concentration of power, and automated decisions that are even harder to question.
Without independent oversight and error-correction mechanisms, the growth of technological capability may even make already existing asymmetric relationships more entrenched.
Therefore, the growth of intelligence does not automatically equal progress in governance.
A smarter system does not necessarily produce a more just society.
Technology can lower the cost of understanding the world, but it cannot automatically establish for humanity the legitimacy of living together.
One is something we are already very familiar with: making machines more and more powerful.
The other is far more difficult: teaching humanity how to use these powerful forces together.
The former has chips, algorithms, laboratories, and technical indicators that can be continuously measured.
The latter, however, involves trust, rights, responsibility, interests, dignity, and the lives of different people in society.
It has no clear performance curve, nor an endpoint that can be declared complete after a single successful training run.
And what we have often neglected in the past is precisely the latter.
05
We once believed so deeply in the power of technology.
We believed that as long as we were smart enough, worked hard enough, and judged far enough ahead, we could continuously break through the limits of existing conditions.
This is not an entirely wrong belief.
Without such a belief, many technologies that changed the world might never have been born.
A person can change the world, yet can never make the world run solely according to their own will.
Moreover, this limitation does not belong only to great figures.
No capability can allow its possessor to completely withdraw from this relationship.
The sorrow we feel today may come precisely from reconfirming this fact.
Humanity has already come so far.
We have left the ground and entered space; we have broken through the limits of the body and continuously extended the boundaries of the mind.
Those things once endowed with supernatural power by myth are bit by bit becoming ordinary engineering problems.
Yet, returning to the relationships between people, we still need to learn how to trust, how to get along, how to constrain power, how to face mistakes, and how to continue living together when none of us can be fully satisfied.
These matters are so ancient that they have accompanied almost the birth of human civilization.
Yet they are also so difficult that, even today, they remain unfinished.
Rather, it lies in whether, when we acquire these capabilities that were unimaginable in the past, we finally have the opportunity to reexamine those social structures that have long determined humanity's fate.
If we merely acquire more powerful tools but continue to distribute power, handle conflicts, and bear responsibility in the same crude ways, then even if the future holds ever more miracles, familiar history may still repeat itself again and again.
But if new intelligence can help us better understand each other's circumstances, make power answerable, make decisions testable, and ensure that mistakes no longer can be corrected only through greater harm, then perhaps this technological revolution will truly touch the depths of civilization.
We do not have to wait for the human brain to biologically re-evolve before it becomes possible to create a different society.
What truly needs to change may be the way we enable countless minds to live together.
This path is obviously far more difficult than building faster chips and larger models.
And no one can guarantee that it will succeed.
We have learned to build ever more powerful intelligence. As for how to make these forces compatible with human freedom, dignity, and living together, humanity is still learning.
We also hope to see that, amid those ancient conflicts and repetitions, humanity finally finds some new ways of getting along.
Because the progress of civilization cannot be measured only by what we create.
It must also be measured by whether, after we have gained unprecedented power, we have learned not to harm each other in the same ways.AI industry signal channel: Frontier technology. This content is published with the author's authorization. The views represent only the author and do not represent the position of Huxiu. If you have objections or complaints about this manuscript, please contact tougao@huxiu.com. Those who are changing and want to change the world are all in the Huxiu APP [ ]