Hawking showed many years ago that pair production near black holes would violate quantum mechanics. But while many relativists were convinced about information loss, string theorists hoped that small subleading corrections to Hawking's computation would invalidate his result. Recently an inequality was derived that shows that such small corrections do not in fact change Hawing's argument. What happens instead, however, is that the structure of the black hole gets altered at the horizon due to the emergence of a new length scale for quantum effects, creating states called fuzzballs. In this talk I will give an overview of these developments, which taken together give us a resolution of the information paradox.