I don't think the fact that our 90th percentile capabilities have been known to decrease at times in history refutes the idea that the broader trend is dramatic

I don't think the fact that our 90th percentile capabilities have been known to decrease at times in history refutes the idea that the broader trend is dramatically upward, in such a way that any absolute measure of skill is very likely to give you ridiculous results if you pick it based on the standards of year X and then try to use it in year X + 1000 or X - 1000. Was Pythagoras a first-rate mathematician? By what measure that wouldn't also capture millions of people--most of them not even self-described mathematicians--alive today?