Many great physicists criticized the renormalization as an incoherent method of neglecting infinities in an arbitrary way. Dirac said that renormalization “is just not sensible mathematics”. Here, applying the Scale-Symmetric Theory (SST), we proved that the mainstream QED is incomplete and is based on a few wrong assumptions. QED leads to experimental data only due to the free parameters. It is not true that there is not upper limit (cutoff) for energy of virtual pairs and it is not true that bare particles are sizeless. The elimination of these two wrong assumptions causes that the modified QED (MQED) is very simple and free from infinities so from renormalization as well. In reality, an electron-positron pair at first appears as a binary system of loops with condensates in their centres both composed of entangled Einstein-spacetime components. Next, due to the superluminal quantum entanglement of the luminal Einstein-spacetime components, the binary system immediately transforms into binary system of tori/electric-charges. With each torus/electric-charge is associated loop, condensate responsible for weak interactions and only one virtual electron-positron pair. Due to the superluminal quantum entanglement, electron disappears in one place and appears in another one, and so on - it is the quantum behaviour of electron. Most important is the fact that both descriptions of magnetic moment and spin of an electron, i.e. via the initial loop plus condensate and via the torus/electric-charge plus loop plus condensate plus virtual pair, are equivalent. Both descriptions lead to the same ratio of magnetic moment of electron to Bohr magneton: 1.0011596521735. This result is very close to experimental data.