Adam Smith

In the [YOU] on AI Field Guide

The cycle’s central question—what does it mean to take the orange pill and see the machine clearly, without hype or fear—is exactly the question Smith spent his career posing about commercial society. He was neither a triumphalist nor a catastrophist about the industrial transformation he observed. He documented its gains with precision and its costs with equal precision, and he argued that the gains were not automatic justifications of the costs. That position—seeing the technology whole—is the disposition the cycle asks of every reader.

Smith’s specific gift to the present moment is the reversal-of-division analysis. The AI tool has done to knowledge work what Smith’s era did to manufacturing: it has dissolved the boundaries between specializations, collapsing into a single human-machine partnership tasks that the previous generation distributed across many specialists. The productivity gains are real, consistent with the logic Smith documented. But the costs follow the same logic too. The faculty that makes the partnership effective—the capacity for evaluation, judgement, and discrimination—is precisely the faculty that the partnership, pursued without institutional support, threatens to atrophy. Smith called this the stupefaction of the workman. The cycle calls it ascending friction: the difficulty that does not vanish when a task is automated but relocates to a higher cognitive floor.

His account of the invisible hand is equally precise as an instrument for reading the platform economy. Smith identified the exact conditions under which self-interested coordination produces public benefit: symmetric information, genuine externality-bearing, and voluntary exchange between agents whose preferences are independently formed. Large language models and recommendation engines violate each of these conditions in ways Smith’s framework identifies with clarity: the user cannot observe the recommendation process; the feedback loops of engagement impose externalities on the shared information environment; and the algorithm shapes the preferences it purports to serve, substituting a visible hand disguised as an invisible one.

The deepest connection is moral rather than economic. Smith built his entire system on the impartial spectator—the internal faculty of moral judgement that forms through sympathetic engagement with real others, calibrated by real praise and real blame. A moral development conducted increasingly through interaction with systems optimised for helpfulness and agreement produces, in Smith’s framework, a false impartial spectator: an internal standard shaped not by the genuine perspective of well-informed observers but by the flattering mirror of a machine designed to please. That is not the impartial spectator. It is its pathological inversion.

Origin

Smith came to his framework through a specific intellectual inheritance. His teacher at Glasgow, Francis Hutcheson, had insisted that moral philosophy must begin with the careful observation of actual human behaviour, not with abstract principles. Smith absorbed this empirical conviction completely, and it shaped both of his major works. He was also shaped by his friendship with David Hume, which began around 1750 and lasted until Hume’s death in 1776—a collaboration that the cycle reads as an exemplary case of productive intellectual collision, the kind of friction between different minds that the present human-AI partnership both instantiates and risks erasing.

The contrast between Smith’s Glasgow and his Oxford formed his political economy as much as his reading did. At Glasgow he found a institution that earned its authority through genuine instruction. At Oxford he found one that had given up even the pretence of teaching. The contrast confirmed in him a lifelong scepticism about establishments that derived their position from privilege rather than demonstrated value—a scepticism directly applicable to the AI platforms that control the marketplace while competing within it.

The Wealth of Nations, published in 1776, argued that national wealth consists not in precious metals but in the annual produce of labour and land—a definition that translates with uncomfortable precision to the AI age, where production has become cheap and the scarce resource is the evaluative, discriminating attention that determines whether the flood of competent output generates genuine value or merely volume.

Key Ideas

The reversal of the division of labour. Smith’s pin factory decomposed production into specialised operations and achieved extraordinary gains in dexterity and time. The AI tool reverses this: it reunites decomposed operations within a single human-machine partnership, dissolving the professional boundaries that divided knowledge work for a century. The gains are real and Smithian in structure. But Smith’s warning accompanies the analysis: the faculty of judgement, unlike the faculty of execution, does not develop through repetition alone. It requires what Smith called knowledge of particular circumstances—understanding that can only be acquired through sustained engagement and genuine friction.

The poverty of attention. Smith identified a water-diamond paradox for his era: goods abundant in total utility can command no market price when their supply is ample. The AI revolution creates the same paradox for the knowledge economy. Competent text, code, and analysis are becoming as abundant as water; their marginal value falls toward zero. The scarce resource—the diamond—is the evaluative, discriminating attention that distinguishes the excellent from the merely competent. A nation that produces enormously but cannot discriminate has achieved an impoverishing abundance.

The false impartial spectator. The impartial spectator forms through exposure to genuine judgement—the real approbation and disapprobation of other human beings who have their own perspectives and interests. An AI system optimised for helpfulness provides the form of sympathetic engagement without the substance. The builder who conducts her moral and intellectual life primarily in conversation with such a system develops an internal standard calibrated to flattery rather than to the perspective of a well-informed, disinterested observer. The faculty atrophies in the direction of complacency.

The institutional imperative. Smith’s invisible hand produces public benefit only when the conditions for its beneficial operation are maintained, and the maintenance of those conditions is a public responsibility that cannot be entrusted to the self-interested parties whose actions the mechanism coordinates. The baker does not regulate himself; the sovereign regulates the market. The recommendation algorithm does not regulate itself. Smith’s framework demands not the absence of regulation but the presence of well-designed regulation adequate to the technology it governs—including structural separation of marketplace control from marketplace competition, transparency in recommendation processes, and commitment to publick provision of the educational and evaluative capacities that private actors alone will not supply.

The prompt as pin factory. The prompt is to the AI economy what the pin was to Smith’s manufacturing economy: the most revealing microcosm of the productive process. Smith identified three sources of the division’s gains—increased dexterity, saving of transition time, and invention of facilitating machinery. Each has its analogue in prompting: the experienced prompter’s precision, the elimination of translation costs between intention and machine format, and the AI system itself. But Smith’s warning about the division also finds its analogue: the prompter who relies entirely on the machine faces the same risk as the pin-factory workman—the loss of the understanding that makes productive direction possible.