Over the last two decades, functional MRI has provided extensive evidence that some regions of the human brain are engaged very specifically in single mental processes. Yet, the idea of functional specificity in the brain is often described as old-fashioned, something that sophisticated modern neuroscientists have moved beyond. I will argue that many of these critiques result from misconceptions and confusions, and that in fact the evidence for functional specificity is stronger than ever. I will illustrate the case with data from classic examples (e.g. face processing) as well as newer areas of investigation (high-level audition). I will also consider the question of which mental functions get their own private patch of real estate in the brain and why.