This article explores the epistemological trade-offs that practical and technology design fields make by exploring past philosophical discussions of design, practitioner research, and pragmatism. It argues that as technologists apply Artificial Intelligence (AI) and machine learning (ML) to more domains, the technology brings this same set of epistemological trade-offs with it. The basis of the technology becomes the basis of what it finds. There are correlations between questions that designers face in sampling and gathering data that is rich with context, and those that large-scale machine learning faces in how it approaches the rich context and subjectivity within its training data. AI, however, processes enormous amounts of data and produces models that can be explored. This makes its form of pragmatic inquiry that is amenable to optimisation. Finally, the paper explores implications for education that stem from how we apply AI to pedagogy and explanation, suggesting that the availability of AI-generated explanations and materials may also push pedagogy in directions of pragmatism: the evidence that explanations are effective may precede explorations of why they should be.