This question is an upshot of a previous question asked by me.
The FRW metric of the Universe is based on homogeneity and isotropy of the universe on a length scale of 100 Mpc or larger. If we go back in time, do we expect the assumption of homogeneity and isotropy to be valid on even smaller and smaller length scales (i.e., the assumption expected to become better) or even larger and larger scales (i.e., the assumption expected to become poorer)?
I guess the assumption must become better because it is assumed to be valid close to inflation when the universe was tiny. But I'm not sure and cannot convince myself.