In joint work with Andrew Marks, we gave a constructive solution to Tarski's circle squaring problem. In particular, we showed that a disk and a square with the same area are equidecomposible using translations. One important innovation of the proof was to construct a real valued flow from the disk to the square. The notion of flow that we use comes from the study of networks and is related to max flow-min cut. In this talk, I will sketch a simpler construction of a real-valued flow from the disk to the square, which is joint work with Andrew Marks. Using discrepancy estimates due to Laczkovich, this argument works for sets whose boundary has small upper Minkowski dimension. I will also mention ongoing work with Anton Bernshteyn and Anush Tserunyan where we construct a large and diverse collection of flows under the same assumptions.