> Do you feel there might be information that they can't publicize? I skeptical and sympathetic to that at the same time.
XKeyscore is a (at a minimum) 6 years old.
You are telling me in 6 years they can't provide the broad strokes of a reasonable number of success stories?
The fact they can't stand up and say "Terrorist plot X was stopped by XKeyscore" from 4-5 years ago pretty much proves its a failure as far as I am concerned.
> I really would say that it's a question of risk aversion & utitility. Even targetting a whole class of people the odds are probably astronomical of finding someone planning harm (1M:1? more than that?). To me it comes down to the negative utility of privacy invasion * the number of people targeted ??? the probability of detecting & thwarting the one malactor, where ??? is an inequality.
Given that with sufficient information a person is able to force another person to do quite a few things via blackmail, its simply too dangerous to trust human beings with this tool. Especially an organization like the NSA where their only knowledge of Snowden's actions came after he released all of the information publicly.