Machine learning embeds the idea that the way to solve a problem is to find an objective to optimize on. Optimization is a particular kind of rationality, one that requires the context to be datafied and asserts that condensing its complexity into a calculation provides a superior kind of solution. Machine learning’s optimizations are a kind of abstract utilitarianism, a mode of calculative ordering that results in particular ways of structuring systems. The logic of optimization, which has deep Cold War roots, already underpins our systems of logistics and planning, and the combination of granular data and machine learning opens up the opportunity for it to be used for social problems. The new era of machine learning means that a similar overarching logic to that which revolutionized global supply chains, through the abstraction and datafication made possible by containerization, can now be applied directly to everyday life.