The examples of plugging in cables and writing data mining code fall into the former category. I'm not suggesting that those people are to blame. Most of the technology that people in our industry build is like this - such as the Internet and the web.
Where I would draw the line (in terms of me personally agreeing or disagreeing to be part of a given project) would be whether that project was specifically designed to operate in a manner that violate's peoples privacy in the manner so many people are upset about right now. This second category builds on the generic technologies of the first, but I consider agreeing to build something like PRISM to be something requiring ethical judgement, whereas the protocols and storage systems which comprise it would not be.
What I'm opposed to is wholesale collection of everyone's data, regardless of whether or not they are suspected of a crime. People who are innocent should not be subject to government surveillance.
That's my opinion at least; I know that some people disagree on that. But regardless of where any of us stand, the public should have full knowledge of the ways in which the government carries out its activities and this should be decided in an open debate in a democratic manner.
And you wouldn't be reading this, because by an obvious extension of your logic, numerous modern technologies including the Internet wouldn't exist.
The flaw in your fundamental argument is that the technologies and access that can be abused for mass surveillance are exactly the same technologies and access that can be used for legitimate surveillance of actually bad people. The people plugging in the cables and writing the data mining code have no way to know how those systems will be used later.