For a given protein sequence, many, up to sometimes hundreds of different amyloid fibril folds, can be formed in vitro. Yet, fibrils extracted from, or found in, human tissue, usually at the end of a long disease process, are often structurally homogeneous. Through monitoring of amyloid assembly reactions in vitro, the scientific community has gained a detailed understanding of the kinetic mechanisms of fibril assembly and the rates at which the different processes involved occur. However, how this kinetic information relates to the structural changes as a protein transforms from its initial, native structure to the canonical cross-β structure of amyloid remain obscure. While cryoEM has yielded a plethora of high-resolution information that portrays a vast variety of fibril structures, there remains little knowledge of how and why each particular structure resulted. Recent work has demonstrated that fibril structures can change over an assembly time course, despite the different fibril types having similar thermodynamic stability. This points to kinetic control of the fibrils formed, with structures that initiate or elongate faster becoming the dominant products of assembly. Annotating kinetic assembly mechanisms alongside structural analysis of the fibrils formed is required to truly understand the molecular mechanisms of amyloid formation. However, this is a complicated task. In this review, we discuss how embracing this challenge could open new frontiers in amyloid research and new opportunities for therapy.