is this really a contrarian opinion? i feel like this is a pretty obvious fact that it's easier to read network requests to scrape a website than parsing html
In my experience internal layout changes seem to happen way more often than changes to the AJAX handlers.