Yes, the moving train thought-experiment illustrates why, if each observer assumes that light always moves at c in their own frame, then different observers must disagree about simultaneity, hence slicing the loaf in different ways (each 'slice' represents a set of events which are all simultaneous in a given frame). As a simple example, if I'm at the exact middle of a train car and I set off a flash there, and there are two detectors at either end, then if I assume light moves at c in both directions in my rest frame, I must conclude that the two events of the detectors going "click" happen simultaneously. On the other hand, if you are on the tracks and see the train car moving forward, you'll see the detector at the front of the train moving away from the point where the flash was set off while the detector at the back is moving towards that point, so naturally if you assume light moves at c in both directions in your
rest frame you'll conclude that the light must have caught up with the back detector before the front one, so the click of the back detector happened at an earlier time than the click of the front one.
Of course nothing forces you to assume that light moves at the same speed in all directions in every frame, you'd be free to pick a set of coordinate systems where light had different speeds in different directions. But the physical appeal of this assumption is that when you do
design each observer's coordinate system based on the idea that light moves at c in every coordinate system, you'll find that when you express the equations representing the fundamental laws of physics in terms of each coordinate system, they end up obeying the same
equations in each of these systems--this is a symmetry in the laws of nature known as "Lorentz-invariance". If you made a different assumption about the speed of light in different directions, the equations would be different in different coordinate systems.