Azeem Azhar

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI asks whether the individual is worth amplifying — what it would mean to bring one’s best self to the tools of the AI transition rather than being consumed by them. Azhar supplies the structural complement that the personal question requires. Whether amplification dignifies or degrades the individual depends not only on the individual’s choices but on the institutional arrangements surrounding her. A society that adopts AI tools without closing the exponential gap — without building the regulatory frameworks, labor protections, and distributional mechanisms that a general-purpose technology requires — will amplify some individuals and dispossess many others, regardless of the choices those others make.

His foreword to the cycle’s Azeem Azhar volume frames the essential question: when Edo Segal wrote [YOU] on AI, he asked whether the reader was worth amplifying. Azhar asks the companion question — are our institutions worth keeping, or worth rebuilding, now that the constraints they were designed around have begun to dissolve? The two questions are inseparable. The technology is abundant, genuinely dangerous, and genuinely ours to govern. The gap between the possible and the well-arranged is closed by human choice or not at all.

Azhar also provides the cycle with its clearest account of why concentration is the default trajectory of the exponential age rather than an anomaly. The economics of near-zero marginal cost, the network effects that reward incumbency, the learning curves that compound early advantage, and the platform forms that diffuse activity while concentrating control all point in the same direction. They are not bad actors’ choices; they are structural features of technologies that improve with scale. Understanding this, Azhar argues, is the precondition for any governance response, because countermeasures directed at individual firms will be outpaced by the structural dynamics that produce new dominant firms as fast as old ones are constrained.

His position in the cycle is that of the thinker who supplies the widest frame. Where others address the tools, the psychology, or the ethics of the AI transition, Azhar addresses the civilizational scale: the relationship between what technology makes possible and what our inherited institutional arrangements are equipped to handle. His answer to the question of which outcome we get is unambiguous: it is decided by us. The technology sets tendencies. Institutions set outcomes.

Origin

Azhar’s analytical framework emerged from his observation, working as a journalist and entrepreneur through the rise of the internet and mobile computing, that the same pattern recurred across every major digital technology: capability compounded at rates that observers consistently underestimated, and the social and regulatory adaptations that followed came years or decades later and were never quite adequate to what they were adapting to. He named the physicist Albert Bartlett’s complaint — that the greatest shortcoming of the human race is our inability to understand the exponential function — as the root cause, and built his career around making the exponential intuitive to people who are not mathematicians.

The Exponential View newsletter, which he founded in 2015, became the intellectual home of a community of analysts, policymakers, and technologists who shared the diagnostic premise that linear thinking about exponential change was the central policy failure of the digital age. The newsletter grew into a podcast distributed by Harvard Business Review and a Bloomberg television series, and the framework it developed crystallized in the 2021 book. By the time the AI transition documented in [YOU] on AI accelerated in 2025, Azhar had been mapping its structural features for a decade and possessed the analytical vocabulary to situate the acceleration within the longer pattern of exponential change.

Key Ideas

The exponential gap. The organizing concept of Azhar’s work: the widening distance between the accelerating curve of technological capability and the comparatively flat curve of institutional adaptation. The gap is invisible year-to-year because each increment of exponential change looks small and manageable. Then, without warning, the accumulated steps cross a threshold and the world is qualitatively different, and the institutions that watched each step approvingly discover they are governing a landscape that no longer exists. Closing the gap requires institutional innovation that matches or exceeds the pace of technical change — governance that is itself adaptive and iterative rather than setting rules once and leaving them to ossify.

General-purpose technology. Azhar’s framing of AI as a general-purpose technology — a technology so broadly applicable that it seeps into every sector, reorganizes how unrelated industries operate, and reshapes society for generations — has direct policy implications. You cannot regulate electricity by writing rules for light bulbs. Similarly, narrow interventions aimed at particular AI applications will always lag the diffusion of the underlying capability. The strategic questions change once you accept this framing: they become questions about access, ownership of the layer, and what complementary institutions must exist to ensure that a general-purpose technology produces broadly shared capability rather than narrowly held power.

The abundance paradox. Azhar argues that exponential technologies produce abundance — goods and services approaching near-zero marginal cost — but that the economics of abundance are not the economics of scarcity, and institutions designed for scarcity do not naturally produce fair distribution of abundance. Markets tip toward winner-take-most outcomes when marginal cost approaches zero. The firms that control the cheap thing control a great deal. Abundance at the level of the commodity can coexist with extreme concentration at the level of the firm. This is why abundance is not automatically liberating: it depends entirely on the institutional choices made as the old scarcity dissolves.

Networks eating hierarchies. Azhar observes that exponential technologies do not merely change what we can make; they change the basic shape of how we organize ourselves to make it. Hierarchies, which were the rational answer to a world where communication was expensive and coordination was hard, are being displaced by network forms. But the networks that emerge tend to become extraordinarily centralized — radical decentralization at the edges, where millions act freely, combined with intense centralization at the core, where the platform owner sets the rules. AI intensifies this dynamic because AI systems improve as more people use them and are most powerful when deployed at the hub of vast digital networks.

Conditional optimism. Azhar describes himself as an optimist but carefully specifies the conditions of his optimism. The abundance is possible but not automatic. The better outcome — AI that expands human possibility broadly rather than pooling in a few hands — depends on deliberate institutional innovation: treating dominant platforms as utilities, requiring interoperability, empowering regulators to unwind acquisitions that entrench dominance, decoupling worker security from the particular job, and building international coordination on AI governance. His optimism is an optimism of the will: it generates obligation rather than complacency, because it is a bet on human agency rather than technological benevolence.