Big Data has made many data processing tasks significantly easier over the past couple of years. We now have the capability to perform data processing like never before. However, Big Data comes with a Big Assumption: that we can bring lots of data together in one homogeneous dataset. But what if we can’t? There are many practical, legal, and economical reasons that make bringing data an impossibility in several uses cases. In this talk, I will discuss the Solid ecosystem of decentralized data stores, leading to a large number of small datasets. I will explain the challenges this brings, both the broad fundamental ones as well as the very concrete developer problems. We will need to think differently, and finally leverage both the “Web” and “Semantic” components of the “Semantic Web”.