Coronagraphs are highly sensitive to wavefront errors, with performance degrading rapidly in the presence of low-order aberrations. Correcting these aberrations at the coronagraphic focal plane is key to optimal performance. We present two new methods based on the sequential phase diversity approach of the "Fast and Furious" algorithm that can correct low-order aberrations through a coronagraph. The first, called "2 Fast 2 Furious," is an extension of Fast and Furious to all coronagraphs with even symmetry. The second, "Tokyo Drift," uses a deep learning approach and works with general coronagraphic systems, including those with complex phase masks. Both algorithms have 100% science uptime and require effectively no diversity frames or additional hardware beyond the deformable mirror and science camera, making them suitable for many high contrast imaging systems. We present theory, simulations, and preliminary lab results demonstrating their performance.