The Deep Feed

For much of the last twenty years, the most lucrative frontiers for capital were digital. We built layers of abstraction, moving from hardware to operating systems, then to the web, and finally to the cloud. Success was measured in lines of code and user engagement. But a shift is happening. The most significant breakthroughs are moving back into the physical world, where the rules are governed by the laws of thermodynamics and molecular biology rather than the logic of silicon. We are seeing a return to the 'hard' sciences, where the stakes are not just a better user interface, but the ability to generate electricity through fission or to rewrite the biological code of a living organism.

The Neutron Milestone

The announcement from Antares marks a turning point in American energy policy and engineering capability. Their Mark-0 low power reactor has achieved criticality at the Idaho National Lab, a self-sustaining fission reaction that has not been seen in a novel reactor design for over fifty years. This is not merely a technical achievement; it is a political and industrial signal. For decades, the United States had largely abandoned the development of new nuclear technologies, leaving the field to aging infrastructure and foreign competitors. Antares has broken that stagnation in less than three years. The transition from making neutrons to making electrons is the next logical step, and it represents a massive leap toward a domestic energy supply that is both stable and scalable.

We’ve made neutrons. Next up: electrons.

This acceleration is driven by a new breed of founders who view the regulatory and engineering hurdles of nuclear power not as dead ends, but as problems to be solved through rapid iteration. The goal is to meet the ambitious targets set by recent executive orders, aiming for a fleet of modern reactors before the nation's 250th anniversary. This is a race against time and old-world thinking. If successful, the deployment of these small, modular reactors could decentralise the power grid and provide the massive amounts of energy required to fuel the very AI models that are currently driving the digital economy.

Rewriting the Biological Script

While Antares is mastering the atom, NewLimit is attempting to master the cell. The company is moving into human trials with a technology that aims to do something previously thought impossible: epigenetic reprogramming. By using RNA to deliver specific transcription factors via lipid nanoparticles, they are attempting to restore the intrinsic ability of cells to regenerate and withstand stress. This is not about metabolic speed; it is about cellular age. In animal studies, this therapy allowed older subjects to behave with the resilience of much younger organisms, effectively reversing the physiological decay associated with age.

The convergence of AI and biology is the engine behind this progress. NewLimit's Ambrosia model can run in reverse: you define the desired cellular state, and the AI proposes the molecular combination required to achieve it. This removes much of the guesswork that has historically slowed drug discovery. The timeline for these therapies has collapsed from decades to years. What was once a theoretical pursuit of longevity is becoming a clinical reality, with the first human trials scheduled for 2027. The implication is clear: the most valuable companies of the next decade will be those that can bridge the gap between digital intelligence and physical reality.

In the high-stakes theatre of big tech, the most dangerous impulse is envy. It is the desire to chase a competitor's success simply because they achieved a hit, regardless of whether that hit aligns with your company's core identity. Satya Nadella, the CEO of Microsoft, has built a career on resisting this urge. As the industry undergoes a massive shift driven by generative AI, Nadella is steering Microsoft not by trying to out-feature every startup, but by doubling down on the company's fundamental role: the provider of the platform. He understands that in a period of radical change, the winner is not necessarily the one who builds the best individual tool, but the one who builds the ecosystem that others must use to create value.

The Envy Trap

Nadella's philosophy is built on a distinction between what the world expects of a company and what a company wants to do out of imitation. He notes that Microsoft is at its best when it fulfills its traditional role as a trusted purveyor of platforms. When a company attempts to mimic a competitor's specific product success, it often loses sight of its unique capabilities. This is the lesson of the Zune and other failed diversions. In the age of AI, the temptation to chase every new model or every niche application is immense. However, Nadella argues that Microsoft's competitive advantage lies in its brand permission—the trust that enterprise customers place in them to provide the underlying infrastructure for their own innovation.

We are at our best when we do what the world expects us to do, we are at our worst when we do things out of envy.

This approach requires a shift in how competitive position is measured. Most executives view competition as a zero-sum game, where every gain for a rival is a loss for them. Nadella views it through the lens of opportunity and capability. The question is not 'How do we beat OpenAI?', but 'What is Microsoft uniquely capable of doing in a world where frontier models are ubiquitous?' The answer lies in creating a frontier ecosystem where many different stakeholders can operate their own intelligence, all while sitting atop Microsoft's infrastructure. This is a move from being a product company to being the essential layer of the new intelligence economy.

The Intelligence Ecosystem

The rise of agentic platforms represents the next frontier for Microsoft. As AI moves from simple chat interfaces to autonomous agents that can perform complex tasks, the need for a unified, trusted platform becomes even more acute. Microsoft is positioning itself to be the glue that connects these agents, the data, and the enterprise workflows. This is a much larger play than simply integrating a chatbot into Word or Excel. It is about defining the architecture of how work itself is performed in the coming decade. By focusing on the platform, Microsoft can capture value from the entire ecosystem of AI development, regardless of which specific model becomes the dominant intelligence.

Ultimately, Nadella's strategy is one of disciplined expansion. It is about identifying the conceptual model of where the opportunity lies and having the courage to ignore the distractions. In a world of infinite technological possibilities, the most successful companies will be those that can define their boundaries and master their core competencies. For Microsoft, that means being the foundation upon which the next generation of builders creates the future. It is a strategy of strength, not reaction.

The central question of the AGI era is not how much wealth will be created, but where that wealth will accrue. If intelligence—the ability to process information, solve problems, and generate ideas—becomes as cheap and abundant as electricity, the traditional economic models of labor and wages will collapse. In a world where machines can perform almost any cognitive or physical task, the concept of 'work' undergoes a radical transformation. To understand the future, we must look past the hype of automation and focus on the fundamental economic principle of scarcity. Value flows to what is scarce. If intelligence is no longer scarce, we must identify the new scarcities that will define the next economy.

The Relational Economy

One of the most significant candidates for future scarcity is the 'relational sector'. This refers to goods and services where the value is derived specifically from the fact that a human is involved. We see this today in the preference for a human barista over a vending machine, or a live musical performance over a digital recording. Even if a machine can produce a perfect cup of coffee or a perfect symphony, the human element provides a unique form of social and psychological value. As automation handles the logic and the production, the economy will likely bifurcate into a massive machine-driven sector and a highly valued human-to-human sector.

Humans are naturally scarce. If we have automation where other things stop being scarce, we will still have scarcity in the things that humans are involved in.

However, this human economy faces a structural challenge. While humans will continue to trade services with one another, they will also be consuming goods produced by the machine economy. This means wealth will inevitably flow from the human sector to the machine sector. The question is whether the human economy can remain a significant part of the total economic output, or if it will become a shrinking niche of luxury services for the wealthy. The value of human connection may increase, but its share of the total economic pie could decrease unless new forms of scarcity are identified.

The Redistribution Problem

The shift in labor share—the portion of national income that goes to workers rather than capital owners—poses a massive political risk. If AGI allows capital owners to capture almost all the gains from productivity, inequality will not just increase; it will explode. Current taxation models, which rely heavily on labor income, will become obsolete. We will need to rethink how we tax the production of intelligence itself. Whether through robot taxes, land value taxes, or new forms of capital levies, the goal will be to ensure that the gains from automation are redistributed in a way that maintains social stability and prevents a permanent underclass.

Ultimately, the economics of AGI will be a struggle to define what it means to be human in a world of machines. If we define ourselves by our utility—our ability to solve problems and produce output—we will find ourselves obsolete. If we define ourselves by our relationality, our creativity, and our unique presence, we may find a new kind of economic purpose. The transition will be messy, and the winners will be those who can navigate the shift from a labor-based economy to one based on the management of scarcity and the curation of human experience.

Modern engineering teams are currently fighting a two-front war. On one side, there is the pressure to adopt AI tools at breakneck speed to keep pace with a rapidly evolving market. On the other, there is the fundamental requirement for reliability, maintainability, and human understanding. This tension has created a divide between the 'enthusiasts', who see AI as an existential necessity, and the 'skeptics', who see it as a source of systemic instability. This is not just a technical debate; it is an organizational crisis that threatens the very foundation of software engineering.

The Race Against Time

For the enthusiasts, the argument is simple: the rate of progress is discontinuous. This is not a normal technology cycle where you can wait for the dust to settle. Companies that sit on the sidelines while their competitors use AI to write code, automate testing, and accelerate product cycles risk being rendered obsolete before they even begin to compete. The fear is not just of falling behind, but of being completely wiped out. In this view, the risk of inaction is far greater than the risk of implementing imperfect tools. Speed is the only way to survive the transition.

Teams that sit this out while competitors are hustling could be out of business before the dust settles.

The Race Against Entropy

The skeptics, however, are looking at the hidden costs. When a team ships code faster than its engineers can actually read and understand it, they are making a massive withdrawal from a technical trust account. This 'technical debt' is not just a metaphor; it is a real, compounding liability. Reliability degrades, institutional knowledge evaporates, and systems become so complex and opaque that no single human understands how they work. The result is a product that may be fast to build but is impossible to maintain, leading to on-call rotations that burn out engineers and products that fail in unpredictable ways.

This creates a feedback loop of instability. As the system becomes more complex, the need for more automation increases, which in turn increases the complexity. The 'entropy' that the skeptics fear is the natural tendency of unmanaged systems to move toward disorder. If the speed of AI-driven development exceeds the speed of human comprehension, the entire engineering organization becomes a house of cards. The goal is not to stop the speed, but to ensure that the human capacity to govern that speed grows at the same rate.

The solution is not to choose one side over the other, but to treat this as a fundamental design problem. Organizations must build structures that allow for rapid experimentation while maintaining rigorous guardrails. This means developing new ways to verify AI-generated output and creating a shared reality where the speed of the enthusiast is tempered by the wisdom of the skeptic. The winners will be the teams that can harness the velocity of AI without succumbing to the entropy of unmanaged complexity.

When humans feel helpless, our most common reflex is anger. It is a corrosive and often misplaced response to a world that feels increasingly out of our control. In periods of intense uncertainty—whether personal, political, or social—we experience a profound sense of vertigo. This vertigo often drives a primal longing for authority. We look for a figure who promises certainty, a hand that claims to be helping but is actually a fist. This impulse is deeply human, reminiscent of a child seeking a parent to resolve the overwhelm of a world they cannot understand, but it is also the mechanism through which many of history's most dangerous regimes have risen to power.

The Allure of Order

The late Leonard Cohen, a thinker who spent much of his life examining the cracks in democracy and the human spirit, understood this danger perfectly. He observed that during periods of bewilderment, the public yearning for order invites uncompromising individuals to impose it. This imposition of order often comes at the cost of liberty and human dignity. The 'sadness of the zoo'—a metaphor for a society that has traded its agency for the comfort of a controlled environment—is the inevitable result of surrendering to the voice of authority. We are often tempted to trade our autonomy for the illusion of safety.

The public yearning for Order will invite many stubborn uncompromising persons to impose it.

Cohen's work suggests that the antidote to this impulse is not found in more anger or more riots, but in a more disciplined form of resistance. He warns against wasting anger on things that do not change the underlying structure of our suffering. Instead, he advocates for a form of resistance that is inward-facing and deeply personal. True resistance, in Cohen's view, is the act of maintaining one's own integrity and capacity for love in the face of a system that seeks to commodify or crush it. It is a refusal to be 'bred for pain' by those who profit from our misfortune.

The Act of Love as Resistance

The most radical act of resistance is often the most quiet: the cultivation of love and empathy. In a world that rewards polarization and the dehumanisation of the 'other', choosing to see the humanity in those who harm or help us is a profound act of courage. It is an attempt to 'let them off the hook'—not to excuse their actions, but to refuse to let their actions dictate our own internal state. This is what Cohen calls 'Radio Resistance'—a continuous, subtle broadcast of human dignity that refuses to be silenced by the noise of authority or the paralysis of despair.

To live this way is to accept the vertigo of the summit and the despair of the valley without letting them break our spirit. It is to recognise that while we may be subject to the forces of history and politics, our internal response remains our own. In an age of increasing technological and political control, the ability to maintain this internal independence is perhaps the most important skill a human being can possess. Resistance is not always a shout; sometimes, it is a quiet, steady refusal to be anything less than human.