[arg01]

I don't think that's a given (possibly even after the leaks). If your giving up very real technical capabilities and increasing costs to protect your self from an ally it seems like a reasonable argument could be made either way. Of course after it's shown that that ally is no longer acting like an ally that changes the weighting of the argument significantly.

I'll also point out a lot of people expect the US would be able to compromise a system if they were actively targeting it even if it was reasonably hardened and in house. They also didn't expect (even if they should of) that the bulk collection was so widespread that it meant common practice implementations were compromised even without serious targeting. So the ROI for in house didn't look as good as it does today.