Abstract : Is extreme modeling so extreme? We advocate that natural modeling might be a better term. After all, the ultimate goal is to enable modelers to perform their job naturally. In the century of the "disappearing computer", it definitively makes sense to search for non invasive and flexible modeling technologies. This paper considers modeling from an anthropological point of view. A retrospective starting back to the Prehistoric Age leads to new perspectives for natural modeling in the Information Age. It is shown (1) that the need for compromises between flexibility and formality is "natural" rather than "extreme", (2) that the languages are emergent by nature, and (3) that natural interfaces should be provided to all stakeholders. We advocate that surface computing, tangible user-interfaces, collaborative modeling and emergent (meta)modeling are future research directions to be investigated in order to make "extreme" modeling just "natural". Just as it should be.