Eye of the Master
Matteo Pasquinelli's social history of AI argues that the field's foundational logic comes not from cognitive science or neural networks but from labour organization -- from Babbage's factory-inspired calculating engines through to modern surveillance algorithms. AI, in this telling, is a technology for automating the extraction of workers' knowledge and crystallizing it into capital-owned machines. The book aims to demystify AI by reframing it as a tool of industrial management rather than a step toward sentience.
Key Quotes
"The inner code of AI is shaped not by the imitation of biological intelligence, but the intelligence of labour and social relations."
Key Themes
Pasquinelli's argument is a useful corrective to the "AI = brain" metaphor that dominates popular discussion. By tracing the lineage from Babbage's observation of pin factories to modern algorithmic management, he makes visible the continuity between industrial automation and machine learning. The book sits alongside Why We Fear AI in treating AI anxiety as fundamentally about capitalism rather than technology, though Pasquinelli is doing historical scholarship while Blix and Glimmer are doing political argument. How to Get Rich in America gives Pasquinelli's history its empirical counterpart: the capital-labor transfer the book traces conceptually — from Babbage's factory to the modern "AI wealth effect" — is measurable today in labor's share of output sliding from 58% to 51% while corporate profits climb.
The "eye of the master" metaphor itself -- Babbage watching workers to extract and codify their knowledge -- maps uncomfortably well onto modern RLHF, where human labelers' judgment gets compressed into model weights.
Critical Analysis
Strong: the historical scholarship connecting Babbage's factory observations to modern ML is genuinely illuminating. Reframing AI as labour technology rather than cognitive technology opens productive lines of inquiry.
Weak: the Marxist framework, while generative, can become a straitjacket -- not every development in ML is best understood through the lens of capital-labour relations. Neural network research genuinely did draw from neuroscience, even if the results diverged. The book risks overcorrecting from "AI = brain" to "AI = factory."
Most useful for people building agents (see Agent Architecture) as a reminder that the systems we build inherit the power dynamics of the training data and deployment context, not just the technical architecture.
Sources: Eye of the Master Last updated: 2026-05-14