This paper examines two influential texts from the ‘factory debate’ of 1830s Britain, Charles Babbage’s On the Economy of Machinery and Manufactures (1832) and Andrew Ure’s The Philosophy of Manufactures (1835). In both treatises, the ‘intelligence’ of machinery plays a dual role, on the one hand replacing skilled workers with automated processes that are not subject to the human failings of indolence, fatigue, inattention, error, dishonesty, drunkenness or ‘horseplay’; and on the other, the surveillance and discipline of that same workforce, in Ure’s words, ‘training human beings to renounce their desultory habits of work, and to identify themselves with the unvarying regularity of the complex automaton’. This paper seeks to examine two key terms in these analyses of the mechanization of physical and mental labour—skill and intelligence—and the way definitions of these terms shift between human and machine models. It also seeks to show how, in the factory system, as Simon Schaffer notes, ‘to make machines look intelligent it was necessary that the sources of their power, the labour force which surrounded and ran them, be rendered invisible’. The last section of the paper will then briefly draw out continuities between these early treatises on automation and the contemporary AI industry, which relies on a largely invisible workforce of low-paid micro-workers manually tagging datasets for training machine learning algorithms, filtering for offensive content, reviewing the accuracy and quality of algorithmic outputs, and even impersonating AI assistants, an instance, as Tubaro, Casilli and Coville wryly note, of humans taking away computers’ jobs. The comparison seeks to illuminate the 200-year history of the so-called ‘paradox of automation’s last mile’, the incessant creation of low-level residual human tasks as a by-product of technological progress.