Errors in division occur all the time and can show up in any dividing cell; this is, of course, important for cancer biology. If one of my cells replicates oddly right now it likely won't matter since it's only one out of trillions, but if that happened at a very early age in development it could be present in many if not all of my cells. Identical twins indeed come from the same egg-sperm fusion event but they develop separately from a very early state. It's trivial to imagine a minor replication error in DNA at an early cell stage that results in different genotypes between so-called identical twins. Besides, we've known for years that identical twins can have major epigenetic differences, so this isn't particularly ground-breaking.
If we assume that identical twins are exactly identical, then if we make a clone of a twin, will all three be exactly identical?
Depends on your definition of "clone" but how identical is identical? If you mean exact, then probably not; at the very least they will have different numbers of cells, for example, not to even mention CNV or epigenetics. If you want some SciFi cloning thing then, well, join the club.