Basically nobody's goal is to "write a performant web server" either, it's to serve data to customers quickly and efficiently. And that highlights why it may not be worth optimizing the Python web ecosystem. There are so many newer alternatives for that overall goal - Firebase, Amazon Lambda, ditching webapps for native mobile, etc - that it may not make sense to try to optimize an application server unless you work for Google or Amazon, because very few people setup a standalone webserver on bare-metal hardware anymore, and those that do probably aren't going to try a new and untested alternative.