The Algorithm Was Always a Foreman

Matteo Pasquinelli's central provocation — that labour is the first algorithm — landed in 2023 as a corrective to the breathless mythologizing of large language models. Three years later, it reads less like a corrective and more like a diagnosis that arrived just before the fever broke. The book traces a lineage from ancient Hindu ritual mathematics through Babbage's Difference Engine, Marx's abandoned "general intellect," and Hayek's connectionist mind, arriving at the thesis that artificial intelligence is not an imitation of biological cognition but an encoding of social relations, specifically the division of labour. This is not a history of inventions. It is a history of how work was watched, parsed, and fed back into machines that then disciplined the workers. The eye of the master, the title's governing metaphor, is the gaze that decomposes skilled labour into reproducible steps — and that gaze, Pasquinelli argues, is what we now call AI.

What the book got right is almost uncomfortable. In 2024 and 2025, the explosion of AI-driven productivity tools did not eliminate cognitive labour so much as Taylorize it: breaking complex intellectual work into microtasks, routing it through algorithmic management systems, and extracting value from the cooperative intelligence of distributed human workers who increasingly cannot distinguish their own contributions from the model's. Pasquinelli's insistence that AI encodes "collective human knowledge and praxis" anticipated the legal and political battles over training data, the revolt of creative workers against generative models, and the dawning realization that the datasets are not raw materials but congealed labour. His excavation of Marx's shift from the utopian "general intellect" to the material "general worker" — the super-organism of humans and machines — now reads as a precise description of the hybrid production systems that dominate platform economies. The chapter on Hayek is particularly sharp in retrospect: the connectionist mind serving not industrial automation but market autonomy maps neatly onto the way AI has been deployed less to automate jobs than to automate market logic itself, optimizing pricing, surveillance, and the sorting of populations into tiers of economic utility.

The blind spots are structural rather than incidental. Pasquinelli's genealogy is resolutely European and Anglophone; the ancient Hindu mathematics of the Agnicayana serves as a gesture toward universality but remains an exception in a narrative dominated by Babbage, Marx, Hayek, McCulloch, and Pitts. The book could not have anticipated the speed at which Chinese AI development would complicate any social history rooted primarily in Western industrial capitalism and neoliberalism, nor the degree to which state-directed AI programs in the Gulf, India, and Southeast Asia would scramble the neat lineage from Enlightenment calculation to Silicon Valley connectionism. More pointedly, the book has little to say about the environmental costs of computation — an omission that felt minor in 2023 but now, after two years of data center buildouts straining electrical grids and water supplies across three continents, registers as a gap in the materialist analysis Pasquinelli otherwise champions. If labour is the first algorithm, energy is the first constraint, and the book does not reckon with it.

Certain passages hit with altered force. The discussion of driving as a cooperative cognitive task that resists full automation was, in 2023, a theoretical observation. By 2026, after multiple high-profile autonomous vehicle programs have scaled back their ambitions or pivoted to narrower domains, it reads as prophecy grounded in social theory rather than engineering analysis. The argument that AI perpetuates class, gender, and racial biases through the very structure of intelligence metrics — not as a bug but as a feature of its genealogy — has been borne out so consistently that it risks becoming a cliché, which is precisely the danger Pasquinelli's historical depth was meant to guard against. The book's value now is not that it tells us bias exists but that it shows us *why* bias is architecturally embedded: because the machines were built to see like masters, not like workers.

In the broader intellectual landscape, this book stands as a bridge between the Italian operaismo tradition and the Anglo-American critical AI studies that have proliferated since 2020, drawing from Marx and Paolo Virno while speaking to the concerns of Kate Crawford and Meredith Whittaker. It gave subsequent work a usable past — a way to talk about AI that does not begin in 1956 or 2012 but in the moment a foreman first watched a worker's hands and thought, *I can break that down*. What it could not give, and what the intervening years now demand, is a theory of what happens when the workers internalize the master's eye — when the algorithmic decomposition of cognition is not imposed from above but adopted voluntarily, even eagerly, by those whose labour is being encoded. If the algorithm was always a foreman, what does it mean that millions of people now invite the foreman into their thinking, their writing, their art, and call it collaboration?