Is cancer one or many? If many, how many diseases is cancer, exactly? I argue that this question makes a false assumption; there is no single “natural” classificatory scheme for cancer. Rather, there are many ways to classify cancers, which serve different predictive and explanatory goals. I consider two philosophers’ views concerning whether cancer is a natural kind, that of Khalidi, who argues that cancer is the closest any scientific kind comes to a homeostatic property cluster kind, and that of Lange, whose conclusion is the opposite of Khalidi’s; he argues that cancer is at best a “kludge” and that advances in molecular subtyping of cancer hail the “end of diseases” as natural kinds. I consider several alternative accounts of natural or “scientific” kinds, the “simple causal view,” the “stable property cluster” view, and “scientific kinds,” and argue that the diverse aims of cancer research require us to embrace a much more pluralistic view.