In a word: noise.
But let me explain:
In the example above, I follow hnshah. He's a good tweeter, someone who posts a lot of good links that are relevant to me. He retweeted @pricing, and it comes into my feed as @pricing. However, I don't know @pricing, they don't have any "brand" cachet with me, and therefore I don't place any value on their tweet. I have to read through the entire tweet to find out where it came from, which was @hnshah.
Given the amount of crap I sift through every day, this extra step makes me more likely to dismiss the tweet. Hence, it's noise.
When I was at Bleacher Report, we did a lot of data analytics on our version of the feed, and found that no one gave a crap about it because it was full of people you don't know. If I don't know you, your actions on the site mean little to me. Even on a site like Facebook, where you're connected to the people in your feed, the chances that I read and process your action are greatly increased if I care more about you. (Not surprisingly, I slow down when I see my wife's name. People I don't know well either get hidden or skipped over.)
Now, Twitter's new retweet function is not all bad - it exposes me to more people, and if I see a new person crop up constantly, I might be more inclined to follow him/her. From Twitter's point of view, the more connections between users, the better.
To get the best of both worlds I would make a small suggestion: swap the order of the data in the retweets. So, it would look something like:
hnshah rewtweeted pricing: Set yourself and your team up for the best chance to win and keep winning.
That way, it leads with hnshah, the guy I follow. I would maintain his icon as well, for scannability.
It'd be interesting to see how these different versions might test against one another, and which one creates the most new connections. Maybe Twitter's already on it, and I'm seeing only one variation.
