Browser automation and RL

For a long time, automating website interaction focused on the very practical tasks of website testing and web scraping. Website testing is especially critical when you have a complicated website that you (or other people) have developed and you want to ensure that it does what it is supposed to do. For example, if you have a login page that has been redesigned and is ready to be deployed on a live website, then you will want to be sure that this new design does sane things in case a wrong password is entered, the user clicks on I forgot my password, and so on. A complex website could potentially include hundreds or thousands of use cases that should be tested on every release, so all such functions should be automated.

Web scraping solves the problem of extracting data from websites at scale. For example, if you want to build a system that aggregates all prices for all the pizza places in your town, you will potentially need to deal with hundreds...