Furthermore, does anyone really know what back doors might be baked into pieces of proprietary hardware and software?
How much monitoring and inspection is really done for the source code to ensure it hasn't been compromised? For the hardware boot sequence hardware? For the encryption technology in use?
Personally I suspect that the capabilities of the largest of networked clustered systems are already capable of cracking encryption on behalf of our most technologically advanced nations. And the US is not alone on that list, as much as it galls them. Nor are human individuals above being tricked into clicking an abusive link, responding to a questionable post, or otherwise triggering the trap of some scammer, which may or may not involve a code injection on your system.
Many systems do little to prevent that. Most of us with moderate knowledge of technology trust anybody but RedHat to provide reasonable monitoring and access options for the source code that is used to build the systems; they, alas, have decided to do their best to obfuscate the changes that they make and make them as difficult to replicate and access as possible, completely missing the point and nature of the GPL licenses on which their products depend. IBM's coffers are dwindling; sooner or later someone with a bigger legal team and budget is going to take them on over the issue.
The days of the proprietary silo are dwindling. People expect and demand that the code be open for inspection so that it's security and quality can be monitored and enhanced. It's in the best interests of everyone who uses open source.