You're assuming the only value Google gets is the raw data.

Does Google seem like the sort of company that stops with raw data?

Or do they seem like a company that will squeeze every tiny little bit of information out of that raw data?

There's MUCH more value in Twitter's firehose than the literal tweets.

Google will extract it.

Can you give some examples. How does anyone know 'stephenfry' is the real Stephen Fry, and an 'authority', and not a spammer.

Well, you can argue RT is more prone to spam because there is less (recent) data and therefore spam filtering is harder. However, with Twitter/social networks the accounts have reliable authorship/history and you can exploit that for filtering... basically rank higher the pieces of data from old accounts that didn't spam before. I wouldn't say spammers have inherent advantage for real time.

Google has been pretty good in figuring out what to include in their "universal" search, I guess their quality control will notice if RT results are not useful for most people.