This text critically examines the principle of replicability, which serves as a cornerstone of modern exact sciences, including biology and medicine. The author posits that a scientific proof is only deemed valid if it can be independently reproduced in another laboratory with identical results. If a repetition fails, it is traditionally interpreted either as scientific fraud or as the result of unrecorded external variables that escaped the original researcher's attention. While this principle is an essential defense against charlatans and scientific dishonesty, the article highlights the risks of its absolute and dogmatic application. The author argues that an uncritical insistence on this method can lead to a flawed intellectual approach to science. The text thus calls for a deeper reflection on the limits of experimental proof and on how to approach unique scientific discoveries that are difficult to reproduce, without immediately dismissing them as unscientific.