Until quite recently it has been generally believed that the observed tertiary structure of a protein is controlled by thermodynamic and not kinetic processes. In this essay we review several recent results which call into question the universality of the thermodynamic hypothesis and discuss their implications for the understanding of protein folding.