The concept that you can melt glass in a microwave oven - as long as you heat the area to near-melting with a blow torch first - seems like a bit of a cheat.
Steve Mould does go on to explain how microwave radiation works to heat food - causing the water molecules in the liquid state to vibrate in the microwave field. This is based on the polarity of the water molecule as well as its ability to freely move.
If the water were locked in a solid lattice (like in ice), those molecules aren't able to vibrate back and forth nearly as easily, hence why ice doesn't heat - or defrost - well in a microwave.
And Mould uses that to his advantage to take glass to the nearly-molten stage, allowing the molecules to move freely and absorb the microwave radiation causing them to heat more and move more freely. He doesn't, sadly, explain why the glass molecules - a network of SiO2 as far as I know - respond to the microwave field similarly to how water does.