IANAL but it's not immediately obvious to me that this ruling covers bulk scraping and republishing untransformed. I'm genuinely curious about this personally. I presumably can't just grab anything I feel like off the web, curate it, and sell it.
It's only after they had become truly huge that they started actually serving data directly (news snippets, song lyrics, factual answers, maps, reviews etc).
Of course. The amount of the original work is one of the key criteria on whether something is fair use or not. And there have been debates over whether Google crosses that line or not.
But no one except the most ardent anti-copyright maximalists would argue that so long as something is available on a public web page, anyone can just reproduce it verbatim and distribute it on their own site. Scrapers do a lot of this anyway but doing anything about it is a wack-a-mole game of publications often don't bother.