This is a linkpost for the original article hosted on my old blog: Visualising L1 and L2 regularisation.