Some authors have recently argued in favor of anti-exceptionalism about logic. The general idea is that logic is not different from the other sciences, and its principles are as revisable as scientific principles. This paper has three sections. In section 1, I discuss the meaning of anti-exceptionalism and its place in contemporary logic. In section 2, I analyze some recent developments on this topic by Williamson (2017) and Hjortland (2017), which will motivate my view. In section 3, I propose a puzzle-solving perspective on logical practice. According to my view, there is a common methodology, in which scientists may use non-classical in order to solve some specific puzzles, but classical logic stays in a privileged position, as a common language and as a general theory of reasoning. This role cannot be fulfilled by other logics, and therefore the comparison between classical and non-classical logic is not like a regular comparison between competing hypotheses in science. The methodology of logical practice is therefore not abductive, at least in many important cases. Classical logic is not the “best available theory”, but the fundamental piece of our scientific methodology. My position is still anti-exceptionalist: logic is like any other science, or at least like any other science which can be characterized by a puzzle-solving methodology.