In contemporary social epistemology, the claim has been made that there is a traditional “neglect of testimonial knowledge,” and that in the history of epistemology, first-hand self-knowledge was invariably prioritised over secondary knowledge. While this paper acknowledges some truth in these statements, it challenges the given explanations: the mentioned neglect of testimonial knowledge is based not so much on a primacy of self-knowledge, but that of self-agency. This article retraces some crucial chapters of this ‘do-it-yourself’ paradigm: it considers the imperative of autopsia in early Greek epistemology, history and medicine, and the early modern refashioning of the privilege of self-generated and self-taught (autodidactic) knowledge. A new picture emerges of how the emphasis on (self-)agency progressively shifted towards a focus on the self as the source of ultimate knowledge.