Ask HN: Why is web architecture optimized for writes?

Why is modern web architecture optimized for "writes" (anything can cheaply write to the database) instead of reads (must build the world on every request, while the user waits)?

There are obvious answers (like it is easier), but for a technology as dominant as the web, these are surprisingly superficial.

An analogous example is indexes in database technologies. They add a lot of implementation complexity, but are paramount for optimizing for "reads". What would a web framework look like that optimizes for "reads" instead of "writes"?

Inspired by https://github.com/craigmichaelmartin/evanesce#ways-to-think-about-evanesce