This paper discusses how central banking is evolving in light of recent experience, with particular emphasis on the incorporation of uncertainty into policy decision-making. The sort of post-crisis uncertainty that central banks are dealing with today is more profound than that which is typically subjected to rigorous analysis and does not lend itself easily to formal modelling. As a practical matter, the policy-maker is dependent on macro models to develop a coherent monetary policy plan, and this burden of coherence means that fundamental uncertainty must be incorporated explicitly into the policy formulation process. As suggested here, doing so transforms policy formulation from an exercise in reverse engineering to one of risk management, one consequence of which is to inject a little more realism about uncertainty into the policy narrative, while trusting markets to wrestle with the data flow and deliver two-way trading. The evolution is likely to be a long one - researchers are encouraged to keep focusing on developing a practical understanding of how the economy works, one that admits that rules around economic behaviour are not cast in stone, but are almost certainly subject to variation through time and events.