A metal block smoothly moves on a glass surface. But a rubber eraser doesn't move very smoothly on a glass surface. In each of the cases, the surfaces are smooth at the macroscopic scale. If the latter is due to friction at the molecular level, how come a metal block can be moved happily on a glass surface?
Glass is smooth but it's not perfectly smooth. Glass has little imperfections. Otherwise it'd be frictionless. The better a material can dig into those imperfections and grip them, the better higher coefficient of friction that material will have when it's pressed up against glass. One reason rubber has such a high coefficient of friction is because rubber is a flexible material, especially when compared to metal. A flexible material can squeeze into those little imperfections in glass like a lock into a key, resulting in a high coefficient of friction. Metal tends to be relatively inflexible. It won't squeeze into those imperfections very much.
Another reason rubber has such a high coefficient of interaction is due to Van Der Walls forces, demonstrated for the first time in only 2013 (Matthias, 2013). Van Der Walls forces come from molecules' electric dipoles. Rubber, involving lots of highly-electronegative carbon atoms bonded to low-electronegativity hydrogen atoms, creates lots of dipoles for other molecules to stick to. Metals, on the other hand, tend to be of middling electronegativity, neither too high nor too low, resulting in few strong dipoles. Many metals are a single element which further reduces the possibility of dipoles.
It all comes down to molecular arrangement. Metal atoms are arranged in a somewhat crystalline configuration called a lattice, and this orderly arrangement is what makes the surface of a buffed and finished metal surface having less friction than a rubber. Looking at the rubber surface microscopically, it isn't as smooth as you think it is, and the molecular arrangement isn't as orderly. On the other hand though, the adhesion force between rubber molecules seems to be way more than that of a metal, I'm not certain about that, but empirically it's highly probable.