3dAutomask and 3dSkullStrip are very different programs. If you use them on EPI datasets, you may get very similar results. If you use them on anatomical (T1-weighted) datasets, you will almost certainly get different results. 3dAutomask makes no attempt to strip the skull -- it just finds the largest connected set of high-intensity voxels. 3dSS is much more complicated, using geometrical information about the skull to try to locate and remove it. If the skull is very dark, as is common in fat-suppressed EPI, then the two programs may end up the same. But that is not guaranteed.