I have spent most of my working life in IT, and over the past few years I have developed a number of applications built around AI. At the same time, I have continued thinking about philosophy and psychology, while paying close attention to the broader effects of technology on society. Because I occupy both positions — working with the technology while also trying to understand what it means — my response to the current development of AI is very direct. I find it more profound and more unsettling than almost any previous technological change I have witnessed.
I do not think the present discussion around AI is simply another technology boom. Few technologies in human history have challenged so many areas of life in such a short time. AI is forcing us to reconsider knowledge, work, education, economics, power, human agency and even consciousness itself. Earlier technologies mainly extended physical strength, perception or calculation. AI is moving into language, reasoning, creativity and judgement — capacities that we once regarded as distinctively human. It is not merely changing the tools we use. It is forcing us to ask again what intelligence is, what knowledge is, what work means and what place human beings will occupy in a future society.
This is no longer an issue that can be left for some distant future. The technology is advancing every day, while our philosophical understanding, social institutions and psychological preparation are clearly falling behind. Before we have properly understood what this transformation means, we have already begun handing more work, judgement and social functions to AI. That growing distance between technological capability and human understanding gives the AI question an urgency that few earlier technological revolutions possessed.
For that reason, I would not casually dismiss Elon Musk’s bold predictions about AI. He has been closely involved in electric vehicles, commercial spaceflight, artificial intelligence and humanoid robotics. His early role in online payments was commercially significant, and he has helped transform electric cars, reusable rockets and satellite communications. It would be unfair to treat him as someone who merely makes extravagant claims.
But what may become technically possible and what human society will ultimately become are two different questions. The more closely I work with AI, the less willing I am to equate technological progress with civilisational progress. AI can increase intelligence and productivity, but it cannot by itself resolve the distribution of wealth, the concentration of power, conflicts within human nature or the organisation of society. This is why Musk’s recent interview with The Economist caught my attention. His judgement about the direction of technology may be exceptional. Yet when he moves from technological capability to predictions about the future of society, the limits of his established way of thinking become much clearer.
The 2036 Musk Envisions
The interview was conducted by The Economist’s editor-in-chief, Zanny Minton Beddoes. It was recorded at Tesla’s Gigafactory in Texas on 20 July 2026 and published on 23 July. The discussion did not remain at the level of technological possibility. Beddoes repeatedly pressed Musk on wealth distribution, government responsibility, war, politics and the extraordinary amount of personal power he now holds.
Musk’s basic projection is straightforward. In roughly five years, he believes AI may exceed the combined intelligence of humanity. By 2036, artificial intelligence will be vastly greater than human intelligence and will be able to perform almost everything better than we can. Digital AI will provide cognition and decision-making, while humanoid robots will become its means of acting upon the physical world. Once vast numbers of intelligent robots can manufacture goods and provide services, economic output will approach something close to infinity. People will no longer have to work in order to survive, and money, taxation and saving will gradually lose their importance.
He compares work in this future to growing vegetables in a backyard today. We do not need to grow our own food, but some people still choose to do it because they enjoy it. Work, he suggests, will eventually become the same kind of voluntary activity. If mass unemployment arrives before complete abundance, governments can simply send people money. Provided that the supply of goods and services grows faster than the money supply, the problem will not be inflation but deflation.
Even more strikingly, Musk acknowledges that there may be a 10 to 20 per cent chance of AI escaping control or even destroying humanity. He also considers it unlikely that human beings will remain in control ten years from now. Yet his attitude has shifted from warning to a form of technological fatalism. The process cannot be stopped, so the best we can do is reduce the risk and accept what is coming. His practical safety proposal is mainly that the leading AI companies should meet regularly, test one another’s frontier models before release, and inform governments when they discover a serious danger. That response seems remarkably small compared with the scale of the risk he himself describes.
Nor did Musk’s image of the future emerge from nowhere. During the interview, he explicitly referred to Iain M. Banks’s Culture novels as the best depiction of the future he expects. Those novels portray a post-scarcity civilisation managed by superintelligent machines, where material abundance has made most human work unnecessary. Musk also spoke about the powerful effect Star Wars had on him as a child. His vision of the future is therefore shaped not only by engineering judgement, but also by a science-fiction narrative that has been forming in his mind for most of his life.
From an Engineering System to Human Society
In the interview, Musk reduces the economy to a combination of digital and physical intelligence. AI solves cognitive problems. Robots manipulate the material world. If intelligence, energy, chips and physical execution can all be expanded, productive capacity can grow by orders of magnitude.
This is a powerful engineering model, and it reflects the kind of thinking at which Musk excels. In an engineered system, we can define a goal, identify the main constraint, concentrate resources on removing the bottleneck and move the system into a new operating state. Many of Musk’s achievements are connected with precisely this ability.
Society, however, is not an engineering system that automatically reaches a desired state once a technical bottleneck has been removed. Every social institution is entangled with power, interests, legitimacy, cultural psychology and historical inertia. Change one condition and the other participants change their behaviour in response. Technology can advance suddenly, but its absorption into society remains constrained by existing institutions and established interests. A technological leap does not automatically produce a smooth or rational social leap.
An increase in productive capacity does not automatically distribute wealth. If robots, energy, land, data and AI systems remain concentrated in a small number of companies and individuals, the first result may not be abundance for everyone. It may instead be falling labour income combined with a further concentration of capital and power. How much machines can produce is one question. What right ordinary people have to receive that production is another. Musk moves from the first question to the second with almost no institutional analysis between them.
His suggestion that the Treasury could send money directly to people who lose their jobs is a temporary technical answer, not a path through social transformation. Who decides how much is paid? Does the ownership of the robots change? How are the gains created by AI distributed? Can different nations enter this condition at the same time? Will the first states and companies to obtain superintelligence use that advantage to dominate others? These are the questions that will determine the future order.
Musk’s experience with the Department of Government Efficiency, or DOGE, illustrates the limitations of applying an engineering approach directly to society. In this interview, he admitted that he had become too involved in politics. DOGE began with extremely ambitious savings targets, but later independent investigations found that many of its claimed savings were difficult to verify and that some cancelled contracts would produce no actual savings. The point is not that government contains no waste. The point is that a public institution cannot be operated simply as a company conducting layoffs and cutting costs. A programme that looks inefficient on a spreadsheet may also support public health, risk control, regional stability and institutional continuity.
The clearest example in the interview was Musk’s categorical insistence that the abrupt cuts to USAID had caused no deaths. He treated the disruption that followed as a failure by private foundations to step in quickly enough. It is certainly difficult to prove a direct causal chain between every individual death and a particular policy decision. Nevertheless, a Lancet modelling study projected that, if the deep cuts continued, they could produce more than fourteen million additional deaths by 2030. Musk is entitled to question the model. But to declare that the number of deaths was zero does not demonstrate rigour. It shows his tendency to compress complex social consequences into a simple and absolute conclusion.
This does not mean that every judgement Musk makes about politics or international affairs should be dismissed as naive. His position on Ukraine is highly controversial, but it is primarily a strategic and moral judgement. Disagreement alone does not prove that he misunderstands reality. His analysis of China’s position in the AI race was, in fact, relatively concrete. He pointed to China’s advantages in electricity supply, manufacturing, robotics, efficient models and the effort to develop an independent chip industry. These are genuine constraints. The strongest criticism is not that every social judgement he makes must be wrong. It is that he repeatedly reduces a highly complex problem to a handful of variables and then reaches a confident conclusion too quickly.
Human Beings Will Continue to Create Differences
Musk has not fully grasped the nature of the human economic system. Throughout history, material scarcity has certainly been an important source of suffering, conflict and social tension, but it has never been the whole explanation. Human nature does not automatically improve when material supply becomes sufficient.
Relative surplus and relative deprivation have existed side by side for a very long time. One part of society may possess enormous excess while another part still lacks basic necessities. This tells us that access to resources depends not only on how much society produces, but also on who owns the resources, who writes the rules of distribution and who is allowed to participate in the system.
More importantly, many human desires do not arise independently. They emerge through comparison. Basic needs can be satisfied, but the need for status has no fixed point of satisfaction. A person may not merely want a house that provides adequate shelter, but a house that is better than somebody else’s. They may not simply want an adequate service, but access to a service from which others are excluded. They may not want wealth only for consumption, but as confirmation of their position within a social hierarchy.
Greater material abundance therefore does not necessarily end competition. It may simply move competition somewhere else. If food, clothing and ordinary transport become abundant, land, environmental quality, privacy, attention, prestige, influence, unique experiences, social relationships and decision-making power can remain scarce. Even where scarcity did not previously exist, people can create it through brands, rankings, exclusive memberships, premium access, limited editions and systems of status recognition.
The accumulation of wealth cannot be explained by consumption alone. Once a person possesses more wealth than they could possibly spend in a lifetime, wealth no longer functions mainly as a means of buying goods. It becomes a way of controlling resources, managing risk, influencing others and shaping the rules. At that point, wealth has become power. Musk asks why money would still matter if goods were more plentiful than anyone could consume. But money has never been only a tool of consumption. It is also a claim upon future resources and a capacity to influence how society operates.
Even if AI and robotics produce extraordinary levels of productivity, the existing pyramid may therefore survive and even become more rigid. Those at the top will not surrender control simply because total output has increased. They may instead build a new hierarchy around ownership of AI, robots, energy, compute, data and distribution platforms. Ordinary people in such a world might no longer lack basic food, yet they could possess less autonomy, less social purpose and less power to participate in collective decisions.
At the same time, it would be another mistake to explain everything through a single, fixed essence of human nature. Human beings display comparison, rivalry, envy, domination and the pursuit of status, but also cooperation, empathy, fairness and reciprocity. A pyramid is not the only social structure compatible with human nature. A more precise conclusion is that human psychology contains persistent tendencies that reproduce difference and hierarchy, while institutions determine whether those tendencies are amplified, constrained or redirected into less destructive forms of competition. Material abundance can reduce absolute poverty and some conflicts produced directly by survival pressure. It cannot eliminate social comparison or transform structures of power by itself.
From Musk to Marx
If we take the argument further, there is a real resemblance between the communist society envisioned by Karl Marx and the advanced civilisation described by Musk.
Both begin with the idea that, once productive forces reach a sufficiently high level, material scarcity will recede, compulsory labour will cease to be necessary, and money and traditional systems of distribution will lose their meaning. Humanity will then enter a freer and more equal stage of civilisation. Marx placed his hope in common ownership of the means of production and the disappearance of class. Musk places his hope in superintelligent AI and humanoid robots. The mechanisms are different, but both assume that a fundamental breakthrough in production will fundamentally transform social relations.
Both also search for a foundational cause beneath social conflict. Marx places the emphasis on ownership of the means of production and class relations. Musk places it on limited intelligence, limited labour and insufficient material supply. Each then tends to assume that, once the foundational condition is changed, many of the conflicts built upon it will also disappear.
Human society does not work that way. Private property can be abolished without abolishing the desire for power. Goods can become abundant without ending competition for status. Work can cease to be necessary for survival while people continue to compete for identity, prestige, control and the ability to dominate others. An old class structure can be dismantled, only for those who control organisation, information and distribution to form a new ruling group. Historical attempts to establish communist societies did not lead to the classless endpoint that had been imagined. They produced new pyramids based on political, organisational and distributive power.
It would still be inaccurate to say that Marx simply ignored social complexity or distribution. On the contrary, he understood more clearly than Musk that greater production does not automatically create fairness. Ownership of the means of production was central to his analysis. Our criticism of Musk — that productive abundance does not automatically give ordinary people access to what robots produce, and that ownership of AI and robotics is decisive — is itself partly Marxian.
What Marx underestimated were psychological mechanisms that may be more persistent than class relations. He saw human nature as being shaped largely by social relations and therefore expected human beings to change when the relations of production changed. This overestimated the plasticity of human motivation and underestimated the durability of status competition, group conflict, the pursuit of power, envy and self-expansion. He also failed to provide a sufficiently reliable institutional mechanism to prevent those who distribute resources in the name of the collective from converting public power into their own power.
Musk’s omission is more immediate. He barely enters the question of ownership or political power. He moves directly from robots being able to produce almost without limit to everyone being able to obtain almost anything, as though no institutions stand between production and distribution and no organised interests stand behind those institutions.
Marx tried to eliminate class by eliminating private ownership, but the result can be the conversion of economic power into even more concentrated political power. Musk tries to eliminate social conflict by eliminating material scarcity, but the result could be unprecedented power for those who control AI and robotics. One assumes that changing ownership can remake human beings. The other assumes that unlimited production can bypass human nature. Both underestimate the human capacity to construct hierarchy, accumulate power and redefine scarcity.
The Hardest Problem Has Never Been Production Alone
Musk’s mistake is not that he has imagined an absolutely impossible future. It is that he treats a possible technological capacity as though it will naturally become a particular social order. AI may exceed human intelligence. Robotics may sharply reduce the cost of many goods and services. These technological directions should not be dismissed.
But productivity is not distribution. Intelligence is not legitimacy. Material abundance is not equality. Technology can reduce poverty, but it cannot automatically eliminate comparison, rivalry, envy, alliance, exclusion and the pursuit of power. When an old scarcity disappears, human beings can create a new one. When an old class structure is broken, power can find another structure in which to accumulate.
Musk’s exceptional strength is his ability to see, often earlier than others, what may become technically possible. His limitation is that he often assumes human society will behave like an engineered system, moving naturally towards his intended destination once the principal obstacle has been removed.
The hardest question has never been how to produce more. It is how to build institutions that can limit power, moderate competition, protect access and correct themselves, while accepting that technology will not fundamentally remake human nature. Whether AI can create a world of extraordinary material abundance is a technological question. Whether human beings can preserve dignity, freedom and a reasonable degree of fairness within that world is another question entirely — and one that remains unresolved.
Sources
The Economist full interview
Interview transcript
https://singjupost.com/full-transcript-zanny-minton-beddoes-interviews-elon-musk/
Reuters investigation into DOGE’s first 100 days
https://www.reuters.com/world/us/100-days-doge-lots-chaos-not-so-much-efficiency-2025-04-24/
Associated Press review of DOGE contract cancellations
https://apnews.com/article/a65976a725412934ad686389889db0df
The Lancet study on the consequences of USAID cuts
https://www.thelancet.com/article/S0140-6736%2825%2901186-9/fulltext
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.