Over a period of more than 20 years, the CIA and Pentagon spent approximately $20 million to study and employ numerous "psychics." They were supposed to help track down terrorists, find hostages, help anti-drug activities, etc. Experiments were conducted on precognition, clairvoyance, and remote viewing.
You are correct that the Stanford Research Institute (SRI) was involved. Apparently, this project began with help from Russell Targ and Harold Putthoff, who had previously "tested" Uri Geller--that should tell you something right off the bat. Anyway, our tax dollars supported this nonsense while they came up with stories much like ones we are used to hearing from proponents of "psychics"--tales which could not easily be verified or falsified, and which underwent changes in the telling over time.
The CIA asked two reviewers to evaluate the studies. One was Ray Hyman, a psychology professor at the University of Oregon and a well-known skeptic. The other was Jessica Utts, a statistics professor at the University of California at Davis and an advocate of parapsychology. Indeed,
the Nature article noted that Utts had participated in some of the studies--which in my mind raises the question of why she was selected to review those same studies.
As we would expect, Hyman and Utts disagreed on how the studies rated. While both agreed that the first "era of research was problematic," Utts said there was "a statistically robust effect," while Hyman noted that "there's no evidence these people have done anything helpful for the government."
So where does this leave us? Let's look more closely at the studies. Utts said the "psychics" were accurate about 15% of the time when they were helping the CIA. Fifteeen percent? Is this supposed to convince us to pay them to help the United States government? Utts says she thinks "they would be effective if used in conjunction with other intelligence." My intelligence tells me that 15% accuracy isn't much help no matter what it's used in conjunction with--that's an 85% failure rate! So 85% of the time, spies would be wasting their time and resources on incorrect information. We're supposed to be happy with that? And that's presuming she's right about the 15%.
http://www.straightdope.com/mailbag/mpsychicfed.html