Many features of the organization of visual cortex develop without the need for normal visual experience and before the beginning of the classical critical period. How such early wiring of the brain takes place, and the degree of synaptic specificity required to do so, remains unknown. I will introduce two dominant proposals, correlation-based learning rules and molecular guidance, and discuss their shortcomings in explaining existing data.  I will then introduce the statistical wiring model, which posits that the early architecture of the cortex in inherited from the spatial structure of retinal ganglion cell mosaics. I will describe how the model is consistent with past data, present some new predictions, and conclude by discussing its current status.