The Deep Feed

01 — Not Boring

The Scarcity Premium: Why the Future is Built on Old Assets

As AI scales toward infinity, the world is retreating into the unreplicable.

By Packy McCormick · 12 min read

Editor's note: A look at how the most valuable assets in a digital age are the ones that cannot be coded.

We are entering an era where the ability to generate intelligence, code, and content is approaching a marginal cost of zero. When anyone can spin up a billion-parameter model to write a marketing plan or a legal brief, the value of the 'thinking' itself begins to erode. This creates a strange economic gravity. As digital abundance grows, the premium on physical and cultural scarcity intensifies. We see this in the massive capital flows toward 'iconic franchises'—the sports teams, the historical landmarks, and the physical institutions that exist outside the reach of a software update.

The Billion-Dollar Courtside Seat

Consider the recent movement of venture capital into the realm of cultural relics. When firms like Thrive Capital move to acquire stakes in entities like the San Francisco Giants or the Lakers, they aren't buying a business model; they are buying an impossibility. You can build a new AI lab every week. You can spin out a thousand new 'big labs' with billion-dollar valuations. But you cannot build a new Los Angeles Lakers. The scarcity of these assets is absolute. In a world of infinite digital reproduction, the one thing you cannot replicate is a shared, physical history.

People keep getting richer, but they stay the same amount of scarce assets.

This isn't just about sports. It is about the fundamental tension in modern capitalism. We are seeing a bifurcation of value. On one side, there is the hyper-scalable, software-driven growth that promises to automate the world. On the other, there is the 'Scarce Asset' class—the land, the legendary brands, and the unique human experiences that serve as the ultimate stores of wealth. For the agency owner or the investor, the lesson is clear: scale is easy; scarcity is the real prize.

The Scarcity Hierarchy

Digital Intelligence (High abundance, declining marginal value)
Scalable Software (High growth, high competition)
Cultural Institutions (High scarcity, extreme value)
Physical Land and Legacy Brands (Absolute scarcity, the ultimate hedge)

Even in the realm of hard tech, scarcity dictates the winners. The recent DARPA Heavy Lift Challenge showed that while drones are becoming common, the ability to move massive weight over long distances remains a rare, difficult capability. The companies that solve the physical constraints of the real world—the ones that can actually move things, seed clouds, or build satellites—will find themselves holding the keys to a much more stable kingdom than those merely playing with bits and bytes.

Key Takeaway

In an age of infinite digital replication, value migrates to the things that cannot be copied.

02 — Stratechery

The CapEx Trap: Financing the AI Revolution

The massive capital requirements of AI may create a bubble that threatens the very margins it seeks to protect.

By Ben Thompson · 10 min read

Editor's note: A critical look at the financial engineering required to keep the AI buildout from stalling.

The current AI boom is not just a race of algorithms; it is a race of capital. We have moved past the stage of mere experimentation into a period of massive capital expenditure. The industry is facing a triple constraint: compute, power, and now, capital itself. While the promise of AI is immense, the revenue required to pay for the staggering costs of chips and data centres has not yet arrived at scale. This creates a dangerous gap between what is being spent and what is being earned.

Financial Engineering as a Bridge

To bridge this gap, we are seeing a new kind of financial maneuvering. Nvidia is not just selling chips; it is helping its customers find ways to fund the purchase of those chips. This involves tapping into long-duration capital and complex funding mechanisms. Google and other tech giants are leading the way in using equity to fund these massive builds. It is a high-stakes game of musical chairs. If the revenue from AI applications fails to materialize quickly enough, the entire structure of this capital buildout could collapse under its own weight.

The blast radius of an AI bubble expands every time a new funding mechanism is invented to protect margins.

This isn't just a technical problem; it's a systemic risk. When the primary way to sustain growth is through increasingly complex financial structures, the industry becomes more fragile. We are building a massive infrastructure on the hope of future utility. If that utility is slower to arrive than the debt or the equity requirements, the correction will be violent. The 'CapEx Train' is moving fast, but we have yet to see if it has enough fuel to reach the destination.

The Three Pillars of the AI Buildout

Compute: The raw processing power required for training and inference.
Power: The massive electrical loads required to run the data centres.
Capital: The liquid funds needed to finance the astronomical upfront costs.

Furthermore, the regulatory response to AI is adding another layer of complexity. The EU's mandate for watermarking AI-generated content is a prime example of how policy attempts to manage a technological shift. While intended to provide transparency, such mandates can create friction for developers and complicate the very tools they are trying to build. The intersection of massive capital requirements and increasing regulatory scrutiny creates a narrow path for sustainable growth.

Key Takeaway

The AI revolution is currently being funded by hope and complex financial engineering, creating a massive systemic risk.

03 — The Marginalian

The Biological Outlier: A History of the Ginkgo

How a single species survived the death of worlds to become a symbol of human connection.

By Maria Popova · 15 min read

Editor's note: A meditation on evolutionary resilience and the way nature survives even when its ecosystem vanishes.

The ginkgo is a biological anomaly. It is a living relic from an era before the dinosaurs, a species that has survived multiple mass extinctions and the shifting of continents. Most species that survive such upheavals do so by adapting to new niches. The ginkgo, however, survived by being an evolutionary generalist that eventually became a specialist in survival itself. It outlived its original pollinators and seed dispersers, becoming a mountain refugee in China before being discovered by humans.

The Error of the Written Word

The history of the ginkgo is also a history of human error. When the German naturalist Engelbert Kaempfer first described the tree in the 17th century, he mispronounced its Japanese name. That error was codified by Linnaeus and spread through the botanical canon. It is a reminder that our scientific understanding is often a collection of well-documented mistakes. Yet, even through these errors, the tree's unique geometry captured the human imagination, moving from the temples of Japan to the royal gardens of Europe.

Rarity precedes extinction, but the ginkgo is a temporal outlier.

There is a poetic symmetry in how the ginkgo moved from the wild to the cultivated. It was saved from the brink of extinction not by natural selection, but by human landscaping. Buddhist monks and later European aristocrats saw in its strange, fan-shaped leaves a beauty that defied the standard templates of nature. We did not just observe the ginkgo; we became its primary mechanism for survival, transplanting it across the globe and ensuring its continuation in an era where its original biological partners were long gone.

Why the Ginkgo Survived

Evolutionary versatility: The ability to switch growth patterns based on conditions.
Energy storage: Using lignotubers to resprout after disturbances.
Human intervention: Cultural and medicinal value that led to global cultivation.

To look at a ginkgo leaf is to see a shape that should not exist in the modern world. It is a reminder that life does not always follow a linear path of 'improvement.' Sometimes, survival is about being different enough to find a niche where the rules of the game no longer apply. The ginkgo did not just endure the end of the Mesozoic; it waited out the chaos, proving that resilience often looks like standing still while the world changes around you.

Key Takeaway

Survival is not always about rapid adaptation; sometimes it is about being an outlier that the world cannot easily erase.

04 — The Marginalian

The Art of Growing Old

Henry Miller's rejection of the 'living corpse' and the pursuit of a fruitful spirit.

By Maria Popova · 11 min read

Editor's note: A sharp, unsentimental look at how to maintain agency and curiosity as the body fails.

Henry Miller's reflections on aging are notably devoid of the usual sentimentality. He does not speak of the 'golden years' or the wisdom of age. Instead, he focuses on the maintenance of the spirit. For Miller, the true tragedy of aging is not the loss of physical strength, but the descent into becoming an automaton. He identifies the 'living corpses' as those middle-aged individuals who have retreated into mental bomb shelters, terrified of a status quo that is inevitably shifting.

The Trap of Public Success

Miller offers a warning about the corrosive nature of success. For a person with something to say, a successful career can become a cage. The prestige that follows a lifetime of work often brings with it a legion of people looking to exploit one's name. The struggle in later life, therefore, shifts from the struggle to achieve to the struggle to remain free. The challenge is to avoid letting the 'false deity of prestige' dictate one's remaining time.

Success, from the worldly standpoint, is like the plague for a writer who still has something to say.

He argues that character is remarkably stable. We often hope that life will refine us, but Miller suggests that success usually just accentuates our existing faults. The brilliance seen in youth often curdles into arrogance in maturity. To avoid this, one must remain capable of the most difficult human tasks: forgiving parents, falling in love again, and maintaining the ability to be surprised by the world. Youth is glorious, but it is also painful; the goal of old age is to achieve a cheerfulness that is earned rather than accidental.

The Markers of a Fruitful Life

Physical agency: The ability to walk, eat, and sleep without chemical assistance.
Emotional fluidity: The capacity to forgive and to love repeatedly.
Intellectual curiosity: The refusal to become a 'middle-class automaton'.
Freedom from prestige: The ability to do only what one wants to do.

Ultimately, Miller's philosophy is one of orientation. How one directs their attention to the present moment determines whether a life is fruitful or merely endured. It is a call to stay engaged with the 'birds and flowers, mountains and sea' rather than retreating into the safety of a comfortable groove. To grow old well is to refuse to let the world own your spirit, even as it claims your body.

Key Takeaway

Aging is not a decline in capacity, but a test of whether you can maintain your agency and curiosity.

05 — The Marginalian

The Quantum of Us: Relationships as Third Entities

Moving beyond the individual to understand the reality of connection.

By Maria Popova · 13 min read

Editor's note: An exploration of how interaction, rather than individual properties, defines reality.

We tend to view ourselves as self-contained units—containers of personality, memory, and intent. We approach relationships as the meeting of two distinct objects. However, both physics and pottery suggest a different reality. In the extreme heat of a kiln, two different glazes do not simply sit next to each other; they react to create a third, entirely unpredictable substance. This 'third entity' is not the sum of its parts, but something of a different order entirely.

The Physics of Interaction

This observation finds a scientific parallel in quantum field theory. As physicist Carlo Rovelli argues, the fundamental reality of the universe is not composed of entities with fixed properties, but of interactions. An object's properties are simply the way it acts upon other objects. There are no 'things' in the traditional sense, only nodes in a vast web of relationships. To describe a thing is to describe how it relates to everything else. Reality is not a collection of objects; it is a play of perspectives.

The self is not the container of your interactions with the rest of the world, but the contents.

When applied to human connection, this perspective shifts everything. If the 'self' is actually the sum of our interactions, then we are not static beings moving through a world of others. Instead, we are constantly being reshaped by the people we encounter. A relationship is not something two people *have*; it is a distinct, living dynamic that exists between them. This dynamic can activate dormant parts of our character and expose impurities that were invisible in solitude.

Lessons from the Quantum/Kiln Model

Relational Reality: The connection is a third entity, not a property of the individuals.
The End of Objectification: Reducing a person to a set of traits misses their true nature.
Perspective as Truth: Reality is a web of interactions, not a collection of absolute facts.

To live with this understanding is to stop trying to 'manage' relationships as if they were tools and to start participating in them as if they were the very fabric of existence. If we accept that we are defined by our connections, then the quality of our interactions becomes the primary measure of our own reality. We are not isolated observers; we are the bridges where the world meets itself.

Key Takeaway

Identity is not a fixed property of the individual, but a dynamic result of our interactions with the world.

06 — Simon Willison

The Debugging Mindset: Lessons from the Smallest Fixes

Why the difference between a working tool and a broken one often lies in a single, invisible dependency.

By Simon Willison · 4 min read

Editor's note: A brief but vital look at the fragility of modern software ecosystems.

In the world of software development, complexity is often an invisible tax. A tool can work perfectly on a developer's machine for months, only to crash the moment it is deployed in a clean environment. This isn't usually due to a failure in the core logic, but due to a missing dependency—a tiny piece of code that was present by accident but never explicitly required. This is the reality of modern, modular software: we are all building on top of layers of other people's work, often without knowing exactly what those layers are.

The Invisible Dependency

The recent crash in `sqlite-utils` is a textbook example. The introduction of a specific type-hinting feature relied on a package that was installed as a side effect of other tools, rather than being listed as a direct requirement. When the tool was run in an isolated environment, the 'invisible' support vanished, and the system failed. This highlights a fundamental truth about scaling technology: as we move from bespoke setups to standardized, automated environments, the things we take for granted become our greatest vulnerabilities.

Reliability is not just about how your code works, but about how it fails when the environment changes.

Solving this required more than just adding a line to a configuration file. It required a new way of testing: running 'smoke tests' in deliberately stripped-down, isolated environments to ensure the tool could survive without its accidental crutches. This is the essence of robust engineering. It is the move from 'it works on my machine' to 'it works in the void.' For anyone managing complex systems, the lesson is to treat every convenience as a potential point of failure.

Principles of Robust Tooling

Explicit over Implicit: Never rely on a dependency that isn't explicitly declared.
Isolation Testing: Test your tools in environments that lack all non-essential components.
The Smoke Test: Verify core functionality immediately after any environment change.

As we continue to build increasingly complex stacks—from AI models to global financial networks—the importance of these small, disciplined checks grows. The more layers we add, the more we must obsess over the integrity of the foundations. A single missing link in a chain of a thousand can still bring the entire structure down.

Key Takeaway

True reliability comes from testing your systems in the absence of accidental conveniences.