The observations being made are correct in that what he's describing is actually and actively happening. The spot where things stop working is in the very last paragraph in that it very neatly wrecks the argument:
Slapping unbreakable crypto onto more and more packets is just going to make matters worse. The only way to retain any amount of electronic privacy is through political engagement.
The words "any amount of electronic privacy" are where we have big problems. When it comes to protecting privacy from the prying eyes of a hacker, the only amount of electronic privacy that is acceptable is total privacy through unbreakable crypto. This is the problem our politicians have a difficult time grokking but most of us understand: what the government is able to break into (i.e. a lesser amount of privacy from total) is also able to be broken into by a hacker. And, in fact, it becomes more easy if it is government mandated since a hack only need to attack the specific method used by the government to inspect those packets.
I agree that political engagement is necessary, but so too is slapping unbreakable crypto onto ore and more packets. The political engagement needs to come in the form of educating politicians on the realities of today's internet-connected world. We either have solid, unbreakable encryption[0] which necessarily includes protection for people performing financial and otherwise sensitive operations over a public network or we have limited and government-breakable encryption that offers less security than "in the clear" transmission[1].
[0] To the extent that such a thing is possible since this isn't a "set it and it's done" sort of situation but an ever increasing arms race of privacy vs. attacker. This includes such things as ensuring researchers are allowed to test/expose and validate said technologies and not be hindered by legal requirements that prevent them from notifying people of the dangers of broken crypto.
[1] At least, potentially less. Assuming the "backdoor" method is used and protected as well as TSA master keys, we now have a scenario where limited use results in easy to identify "high value" traffic and less security because the hacker need only attack the government for the masters. Even if a much more "well thought out" method is used for providing government access, the problems multiply since now we're not only having to create sufficiently attack-resistant protocols but also manage sufficiently complicated escrow or other processes to provide access. Complexity is the enemy of reliability.