The vertical shift is this book's term for the structural change in international power distribution that artificial intelligence produces. For four centuries, international relations theory has analyzed power as something flowing horizontally — between sovereign states, alliances, and blocs. The AI revolution breaks this horizontal axis by enabling power flows that move vertically through levels of organization: from institutions to individuals, from credentialed centers to uncredentialed peripheries, from the gatekeepers of capability to anyone with an idea and a natural-language interface. The unit of analysis can no longer be the state alone. It must include the relationship between states and the individuals they now empower or fail to empower.
There is a parallel reading that begins not with what individuals can now do, but with what they must depend on to do it. The vertical shift describes a diffusion of capability, but every act of that capability runs through infrastructure owned by a vanishingly small number of actors. The woman in Trivandrum who built a frontend feature in two days did so by sending natural language prompts to a model hosted by OpenAI, running on compute provided by Microsoft, accessed through an API whose terms of service she did not negotiate and whose pricing she does not control. The power has not diffused to her. The interface has.
What appears as vertical flow is better understood as the extension of horizontal power through longer supply chains. The same forces that concentrated cloud computing into three companies, mobile operating systems into two, and social graph infrastructure into a handful of platforms are now concentrating the foundation model layer with even greater capital intensity. The individual developer's new capability is not autonomous power but delegated access, revocable by terms-of-service change, pricing shift, or geopolitical pressure. When the infrastructure providers are subject to national security frameworks, export controls, and content moderation regimes shaped by a small number of states, the 'empowered individual' is not a new unit of analysis but a retail customer whose capabilities can be switched off remotely. The vertical shift names a real phenomenon, but it may be the prelude to the most efficient system of horizontal control ever constructed.
The horizontal assumption in international relations derives from the Treaty of Westphalia (1648), which established the sovereign nation-state as the fundamental unit of the international system. Every major framework since — balance of power, concert of great powers, bipolar Cold War, unipolar American moment — has analyzed influence as lateral movement between states. The scale tips left or right; the players change; the scale itself remains flat. Nye's soft power framework expanded the concept of what constituted power but operated primarily on this horizontal axis, tracking influence between nations through cultural and institutional channels.
Artificial intelligence breaks the horizontal axis by collapsing what The Orange Pill calls the imagination-to-artifact ratio — the distance between a human idea and its realization. The woman in the Trivandrum training who had never written frontend code built a complete user-facing feature in two days, not by acquiring a new credential but by describing what she wanted in plain language. Multiply this by forty-seven million developers worldwide, with the fastest growth in Africa, South Asia, and Latin America, and the geopolitical implications become visible. Traditional centers of technological power retain advantages in capital, infrastructure, and network effects, but they have lost a monopoly they did not know they held: the monopoly on who gets to participate in building the technological future.
The distinction between communication tools and production tools is analytically critical. Previous waves of technological empowerment — the internet, social media — were primarily communication tools. They enabled individuals to organize, protest, and coordinate, but the political outcomes were largely captured by existing institutional actors because communication alone does not create durable power. AI is different because it is a production tool. The individual who can build a product, serve a market, generate revenue, and iterate based on user feedback acquires self-sustaining power that a hashtag cannot produce. Power to make, in economic systems, is power that compounds.
Nye anticipated this development in The Future of Power (2011), distinguishing power diffusion from power transition. Power transition moves capability between states. Power diffusion moves capability from states to non-state actors — corporations, NGOs, networks, and individuals empowered by technology. The AI moment accelerates diffusion to a degree even Nye's 2011 analysis did not fully envision, because the technology enables individuals not merely to organize but to build — to create economic artifacts that compete directly with institutional outputs.
The vertical shift as an analytical concept emerges at the intersection of Nye's power diffusion framework and the specific empirical observations The Orange Pill documents about AI's collapse of the production cost barrier. The concept responds to the inadequacy of existing international relations vocabulary for describing what happens when capability previously held by institutions becomes accessible to individuals at civilizational scale.
Horizontal versus vertical. Traditional international relations theory analyzes power as horizontal flow between states; AI introduces vertical flow between levels of organization, requiring new analytical frameworks.
Production not communication. Unlike previous technological empowerment waves, AI is a production tool that enables individuals to create durable economic value, not merely to communicate.
Lost monopoly. Traditional technological centers retain advantages in capital and infrastructure but have lost the monopoly on who participates in building the technological future.
Compounding diffusion. Because the power that diffuses is the power to make, its effects compound; each enabled individual produces outputs that enable further individuals.
New unit of analysis. The state alone is no longer adequate as the unit of analysis; the relationship between states and their empowered citizens becomes the defining variable.
Skeptics argue that the vertical shift is overstated — that the computing infrastructure, foundation models, and platform economics on which individual empowerment depends remain concentrated in a small number of corporations and states, and that individual capability without institutional backing is fragile. The book accepts this concern as the specification of the dam deficit rather than a refutation of the vertical shift: the diffusion is real, and the institutional failure to support those to whom power has diffused is precisely what creates the strategic vulnerability the framework identifies.
The right frame here distinguishes between capability diffusion and control concentration, treating them as separate dynamics that happen to be occurring simultaneously. On the question of what individuals can do, Edo's account is essentially correct (90%): the collapse of the imagination-to-artifact ratio is real, measurable in products shipped and revenue generated by actors who were previously structurally excluded. A developer in Lagos who builds and sells a successful tool has acquired economic power that compounds, and no amount of infrastructure dependency changes the fact that this capability did not exist five years ago. The diffusion is empirically observable.
On the question of who controls the infrastructure that enables this capability, the contrarian view is closer to reality (70%). Foundation models, compute, and platform access are concentrating at a rate that would make the robber barons blush, and this concentration creates structural vulnerability. The interesting tension is that both dynamics are true and consequential. Capability has genuinely diffused—the genie is out of the bottle in ways that cannot be fully recalled—but the infrastructure on which that capability depends is simultaneously centralizing, creating new choke points and dependencies.
The synthesis the topic itself benefits from is this: we are watching the construction of a new kind of power architecture, one in which widespread capability coexists with concentrated control in a way that previous technological revolutions did not exhibit. The relevant questions are not whether the shift is real (it is) or whether infrastructure concentration matters (it does), but how these dynamics interact over time, under what conditions capability can route around control, and whether the diffusion proves durable or merely licensed.