Researchers typically do not have the time to keep up with the literature outside their subject domain. Major breakthroughs often occur when someone is in the right place at the right time to pull together the disparate threads of seemingly unrelated research. If Google can address this problem algorithmically, it could have profound impact. These thoughts have been rolling around my head since my young brain was warped by the esteemed James Burke many years ago.
In my 146 updates the first third is younger than 1 month, and about 4 in the 146 I have already found by ususal ways of following the literature. But the list contains a higher 'concentration' of interesting sources, than any issue of my favorite journals. That's good support for any researcher.