What scares me most is that now that storage is so cheap, everything can be stored for later. So even though no one may be watching now (likely), they keep a dossier with all your calls, mails, visited sites, google searches, for any time in the future to look at at their leisure.
It always reminds me of a shocking fact in Dutch history:
In the Netherlands, the Germans managed to exterminate a relatively large proportion of the Jews. The main reason was that before the war, the Dutch authorities had required citizens to register their religion so that church taxes could be distributed among the various religious organizations (https://en.wikipedia.org/wiki/Dutch_resistance)
And this looks so quaint now. We can only imagine what potentially damning information a totalitarian government can now find about every citizen retroactively.
Digital networks are increasingly an intermediate in every little communication and transaction between people. And with the internet of things, in everything we do, in the future maybe even inside our bodies (what's after Google Glass?).
We really need a way to prevent rampant data collection, otherwise the internet is a large threat to civilization. I didn't go into technology to facilitate some 1984-ish world government :(
And of course the prime example:
https://en.wikipedia.org/wiki/IBM_and_the_Holocaust
Putting aside questions about IBM's complicity, it's certain that Jews were indeed identified and recorded in their millions using the Hollerith punched card technology.
So is your plan to move technology back in time to 1938 so that people can't be tracked automatically?
Given that I don't see that as being feasible we should ask ourselves instead what can be done to avoid a homicidal state, even under the assumption that they have more computing power available than in your iPhone.
It's also a poster child of secure F/OSS comms - easy to get the source code, and crazy difficult to have a secure conversation with a verified buddy.
There's no obvious getting started guide, no plain English explanation stating that Pidgin must be installed first, nothing about configuring for first use, and nothing about starting a verified conversation.
Will that solve things? As we are seeing, much of the value is in the metadata, who is speaking to whom. Cryptography won't necessarily hide that, as ultimately the network itself needs to know where messages are going to.
The design of the network has a role to play. Is it possible to design a network that doesn't expose where information is flowing, or better yet, doesn't even need to know where information is flowing (it can't leak what it doesn't know)? Such a network would presumably not require an address space.
Freenet does something like this, exchanging messages by a process akin to a dead drop and restricting each node's view of the network to its immediate neighbours. I'm thinking something like Freenet, but operating as a physical network rather than an overlay network. Does such a thing already exist?
"Is it possible to design a network that doesn't expose where information is flowing, or better yet, doesn't even need to know where information is flowing (it can't leak what it doesn't know)?"
Post encrypted messages to Usenet; since anyone can receive them, there is no need for a destination address. Post the messages through anonymous remailers (mix-nets) if you want to avoid revealing that you sent them.
Strong crypto is only one step... but it's the no brainer softball step, it's _just code_ and it doesn't have to have any gnarly UI impact.
We won the important battle getting the regulations largely out of the way. If we can't achieve pervasive always on encryption— and at least kill passive dragnet content collection dead— then can we achieve anything at all?
