China’s pursuit of global leadership in the research and development of artificial intelligence (AI) has been extensively documented. AI is widely discussed in Chinese media, addressed by national policy documents, and implemented in growing numbers of digital platforms and consumer products. Driven by advancements in both optical devices and deep learning, machine vision is one of the main applications of AI, and a key component through which users interact with automated systems. In China, these systems include digital payments, epidemic control, interactive entertainment, industrial manufacturing and police surveillance. As a broad domain of computation bridging AI and optical media, machine vision is increasingly central in determining how states, platforms and users see each other across scales. Understanding how machine vision is used in everyday life ‒ from the few bits of information encoded in a barcode to large technological systems like biometric surveillance ‒ allows researchers to probe into new articulations of mediated agency and optical power. Drawing on preparatory research for an ethnographic study of these technologies, this talk will discuss the role of machine vision in Chinese everyday life through three cases studies: the success of QR codes and other data encoding patterns as infrastructural gateways; the popularization and regulation of deepfakes; and the controversies around the deployment of increasingly pervasive biometric identification systems.