The headline makes it sound like they did work to improve the performance of count(*) in the postgres code, when that's not at all what the article is about. T

The headline makes it sound like they did work to improve the performance of count(*) in the postgres code, when that's not at all what the article is about. This post from Citus is far more informative: https://www.citusdata.com/blog/2016/10/12/count-performance/ https://www.citusdata.com/blog/2016/10/12/count-performance/