Machine learning is fundamentally transforming financial markets. Where trading strategies were once crafted by human experts--executed manually or through pre-coded rules--firms now build models that generate the strategies themselves. These are not just tools but trading automatons: semi-independent systems designed to learn from markets and act on their own. Drawing on over a decade of fieldwork in financial markets, Christian Borch offers a rare inside look at how these systems are built, the risks they pose, and how they challenge our understanding of markets and decision-making. As trading automatons grow more complex and opaque--even to their designers--new sociological questions emerge: What happens when machines become the primary agents in markets? And how should we understand economic action when human judgment is no longer at the center? Trading Beyond Understanding is a powerful investigation of machine agency, market transformation, and the shifting boundaries between technological systems and social life.