Two inertial observers in motion relative to one another may disagree as to which of two events happened before the other.
They are both drawing reasonably sensible and natural conclusions from their respective observations – assigning what they actually see to a time that is earlier by the light travel time (which they compute from the observed distance and the measured speed of light). But since they disagree one or both must be wrong if they both think their own idea of simultaneity is the only “right” one.
It could be that one of them (or perhaps some other inertial frame) is “right” in some sense and all the rest are “wrong”. But we don’t have any way of exactly identifying the “right” frame, and using that fact as a starting point provides a way of making accurate predictions without needing to make the untestable assumption that some arbitrarily chosen particular frame is in fact the special one.