Superconductivity and magnetism are often considered as antagonistic phenomena. Indeed the phenomenon of superconductivity is usually observed in materials showing no instabilities toward magnetic order of any kind. However, in recent decades several new classes of superconductors have been discovered where superconductivity occurs on the brink of magnetic order, and in a limited number of examples coexistence of superconductivity and ferromagnetism has even been claimed. In many of these new materials the Cooper-pairs were observed and/or predicted to have finite angular momentum, and sometimes even finite spin. The field of superconductivity has, so to speak, become very 'rich' and ever so more fascinating. In this Colloquium I will, based on the state of affairs in unconventional superconductivity, argue in favour of the scenario where high temperature superconductivity arises from pairing mediated through the virtual exchange of spin fluctuations. This is born out by a vast body of experimental data obtained worldwide using photo-electron spectroscopy, tunnelling and optics. These experiments have enabled the community to establish the temperature and energy dependence of the dissipation and the mass renormalization of the charge carriers in various materials including the high Tc superconductors. Based on the energy range observed for the inelastic scattering, it is now possible to eliminate phonon mediated pairing as 'the' mechanism of superconductivity in the cuprates, and provide quantitative arguments in favour of the thesis that superconductivity in these materials is driven by proximity to magnetic (or stripe) instability.