You are 100% wrong. The ad copy on craigslist IS copyright protected so it doesn't matter where you scraped it from. TOSs do govern the data whether or not you are viewing a cached copy or not. By your logic all TOSs are rendered useless if you browse the net via proxy servers.
You can scrape anything off the internet you like and do as you please...but you can't create a business out of it.
Site TOS don't really matter too much or else Google could never exist. They're mostly to limit the liability of the publisher and not to prevent anything with the accessor. Googlebot is not a lawyer, it can't decide the legal implications of scraping or what exact activity can be done with the data it finds.
Google also overlays facts it finds by crawling the web in maps (Google Places and the One Box results) and uses the creative contribution of other people to provide things like reviews inline on SERP. Google is a bad target to sue though because they will punch you right back in the face.
You can scrape anything off the internet you like and do as you please...but you can't create a business out of it.