Recent spatial econometric contributions call for empirical models to be more often derived from spatial theory and W matrices to be more closely related to actual inter-regional linkages. This manuscript answers this call by reviewing some of the latest developments and suggesting future research venues along these lines. All examples are based on the regional knowledge production function literature as enormous advances focusing on the spatial nature of the dynamics at work have taken place sinces Griliches (Bell J Econ 10(1):92–116, 1979) seminal but aspatial contribution. Furthermore, this literature offers several examples of spatial weight matrices that offer innovative ways to account for the nature, magnitude, asymmetry and directionality of inter-regional (knowledge) spillovers. We foresee that other exciting regional science topics will follow this path.