You only need a few sympathetic hardware engineers to put in hardware back doors, however, and it's practically impossible to verify that the hardware you have

You only need a few sympathetic hardware engineers to put in hardware back doors, however, and it's practically impossible to verify that the hardware you have is trustworthy by the time it's in your hands. A million engineers all making sure they don't implement surveillance functionality can't undo the work done by the handful of bad guys unless they're somehow able to inspect every piece of hardware down to the lowest level and verify manually that it's safe. The user has to do the same thing when they receive their hardware for obvious reasons. And software is no different. There are lots of FOSS devs who would refuse to help implement surveillance systems, yet there's no practical way to prevent one small team of bad actors from injecting code somewhere between software-conception and delivery-to-end-user. The only possible way to be confident about your technology would be to build it yourself from the ground up. You would need to design and fabricate the hardware yourself, from passive components up, and re-implement the entire software stack by hand. Even if you did all that, if you're a desirable target then you still face a very high risk that one of the people working somewhere in your project will be employed by your opponent and will compromise your system. The conclusion I have come to is that if you have anything you don't want intercepted, you simply can't use any modern device.