I haven't used iOS in a while but hasn't it always been extremely obvious when an app was using the microphone in the background?
* Firmware is patched after the fact either by inserting a backdoor into the build process, or binary patching the resulting firmware image. Such an operation is likely to be compartmentalised to avoid employees knowing about it.
* Intelligence agencies seed online discussions with false assurances (note that I'm not suggesting any wrongdoing on your part - merely pointing out that an anonymous forum post is hardly verification that can be relied upon).
That said, I do agree that exploits are likely sufficient, and that planting a mass backdoor is probably not worth the risk of it being detected.
everyone is ignoring what was pointed out by "kennywinker"
https://news.ycombinator.com/item?id=20884384
the fact that both for android and for iphone seperately, there is a correlation with sound / silence and data usage, even if in the opposite sense. why did this not at least prompt the investigator to dig deeper?
There's also hardware exploits. The firmware doesn't need to have a back door if a certain sequence of packets sent to the network controller will cause key memory to be overwritten and live patch in a backdoor.
That said: monkey-patching is much more easily concealed than hardware or source-code changes, and is more in line with how US spooks have been seen to operate.
Do Android and baseband firmware run on one and the same processor(s)? Or are we calling ARM Secure World / TrustZone baseband these days?
That or you and I are just a lot more predictable than we think (which is actually what I think).
It’s a distinction without a difference.