Gromov-Hausdorff convergence is a standard way to make sense of weak limits, but alternative weak notions are possible. Kontsevich and Soibelman [49][48] aim to establish non-archimedean geometry as a suitable framework for studying limits of Calabi-Yau manifolds and the consequences for mirror symmetry. Kontsevich and Tscinkel [50] initiated the attempt to build up non-archimedean pluripotential theory by imitating Kähler geometry, a task taken much further by Boucksom et al. [4][3][6][5] (cf. section 5). Kontsevich and Soibelman may have anticipated long before any rigorous definitions, that the Calabi-Yau metrics should converge in some potential theoretic sense to a non-archimedean object, and the limiting information should be read off purely in terms of data on the essential skeleton.