Every gen normally lets call it 150mm^2 die 230-300mm^2 die, 400-500mm^2 die, low, middle, high end. Then you get a new node, high end performance in the new node midrange card. That's what we have a ~255mm^2 core having the performance of a last gen what 469mm^2 core. That's what we expect, that is normal. When the economics and yields work out that we can get a 350-450mm^2 die, we'll have a leap in performance.
Now with Nvidia, rather than going 450-500mm^2 16nm die to a 450mm^2 7nm die, they went to a 800mm^2 12nm die. Shrink that to 7nm and you'll have a 450mm^2 die. This gen isn't normal in that sense in that due to the cost, difficulty of 7nm they went HUGE on 16/12nm.
If Nvidia stopped at 500-600mm^2 on 16nm, then the 2070 would be a much slower card than it is being probably a 350-400mm^2 die. Comparing it to a normal gen and normal shrink would be silly. If and when we got a ~800mm^2 card on 7nm it should be WAY faster than a 2080ti, but it will be a long time before that would be viable yield/cost wise and maybe not ever.