Here’s the document President Obama mentioned in his statement yesterday, detailing the administration’s legal rationale for surveillance efforts.

A fascinating section that shows the extremely limited scope of the analysis done on the massive amounts of metadata collected:

Thus, critically, although a large amount of metadata is consolidated and preserved by the Government, the vast majority of that information is never seen by any person. Only information responsive to the limited queries that are authorized for counterterrorism purposes is extracted and reviewed by analysts. Although the number of unique identifiers has varied substantially over the years, in 2012, fewer than 300 met the “reasonable, articulable suspicion” standard and were used as seeds to query the data after meeting the standard. Because the same seed identifier can be queried more than once over time, can generate multiple responsive records, and can be used to obtain contact numbers up to three “hops” from the seed identifier, the number of metadata records responsive to such queries is substantially larger than 300, but it is still a tiny fraction of the total volume of metadata records. It would be impossible to conduct these queries effectively without a large pool of telephony metadata to search, as there is no way to know in advance which numbers will be responsive to the authorized queries.