PGP, if used incorrectly, can actually be worse than if you had communicated in plain text (assuming you have some preshared phrases that make sense in context to use). PGP can reveal not only to whom your talking but without a doubt that it is you and only you and them and only them no matter what email addresses you use.

PGP doesn't necessarily protect against replay attacks either and when sent via email, the email the headers and subject lines are, of course, unencrypted and unprotected. Accordingly never put sensitive information in mail headers and, by extension, the subject line.

You can use a well-known strongset key for rendezvous but once contact is established always rotate PGP identities with every communication and discard target keys with the throw-keyid(s) option. PGP/GPG can keep you safe but it takes a significant amount of work to use correctly and one screw up can bring it all crashing down.

If you need it to keep you alive you must practice, practice and practice some more. Develop your own tools and filters to catch mistakes and always consider the metadata trail you might be leaving. Metadata can convict. Remember, the United States Government targets people for execution based on metadata alone.

Consider posting crypted/signed messages to Usenet instead of using email to obscure your communications. Sign and encrypt your messages on an offline system and distribute them online using a different and preferably public computer you don't own and only use once. Distribute and collect your correspondence via ToR if at all possible. Keep your keys offline on encrypted storage. Set up a duress key, always chain different keys (never thread messages with the same key) and always plan for failure, because if it can fail it will.

Only then will PGP/GPG have a chance at helping you stay alive and free.

Good tips but GPG gives a nice default. We know laypeople following basic advice successfully stopped NSA collection from reading their messages per Snowden leaks. So, the pitfalls are far lower than the benefits. Sure, they might get hit with a 0-day, be profiled or whatever. Sure they can improve.

You're post just came off as unnecessary negative on what was one of only two recommendations that worked consistently per the leaks. Anyone reading it wouldn't have known about its successes and would've assumed people just got busted all the time with it.

You're right. I apologise for the negative tone. PGP/GPG are some of the best tools we have right now for protecting privacy on a hostile network. Increasing and normalising its usage is vital to protecting privacy. I wanted to underscore the fact that the system is like a very strong lock - its level of protection depends very much on the surrounding weak links. PGP/GPG alone will keep you safe from a voyeur, but alone it will not keep you safe from an attacker. PGP/GPG does not, in my opinion, sufficiently make this distinction to the lay person. This is very dangerous because a false sense of safety is far worse for all parties involved. It is far too easy to go from 'this would be embarrassing if people knew' to 'they might kill me' information using the same keypair, because people think 'nobody can read this, it's encrypted'. That said, GPG/PGP is a great tool but in the end it's the skill with which you use that tool that determines how much it protects you and GPG/PGP makes it exceptionally easy to shoot yourself in the foot once you've learned how to use it in a basic way.

Also, burn keys after each message. Never reuse non-rendezvous keys and preferably set up new rendezvous keys for each group, groups or entities with which you collaborate and rotate rendezvous keys regularly. Finally, never respond on a regular schedule or too quickly. Delay is your friend. Delay adds noise and makes it harder to connect the metadata you're generating to your data trail thus revealing the connection and breaking one layer of your hopefully multilayered defence.

Technical means are not enough. We need political change.

Historically technology has been more successful in affecting political change.

I believe that we have an example of such "success". In Stasi era total surveillance was too expensive. Today technology made it cheap, hence we do total surveillance.

We could try to bet on technology-first approach. That means rewriting of all the stacks, create automated solutions that hack and/or patch systems (see DARPA Cyber Grand Challenge), hardening systems and protocols, etc. Now, this leaves us with dumb users that need to be retrained. And that can't be done.

And there is other side of things. While we wait for thing that may not happen (dark internet),

a) public money are being wasted

b) surveillance is creating chilling effect TODAY

c) wars with US involvement are raging, fuelling ranks of radicals in Arab world

>> We could try to bet on technology-first approach.

Since we don't control all the stack(processors, maps ,cellular towers, etc ), and we'll probably won't be allowed to control all the stack , the technology-first approach is just wishful thinking.

I disagree. In the end, all encryption is breakable with a rubber hose. Nothing stops people in a position of great power from abusing that power except social and political pressures. The key is to create a social and political environment that stigmatizes these behaviors.

Unfortunately, we're not there in the US yet. I believe it will come, in the same way that the search warrant became necessary for tapping someone's phone, say, but it will take time for society to catch up.

Can you cite an example? It's hard to come up with one where you can't find examples of use or abuse, with the direction depending on social factors.

Yes, and Tor :)

Good that you two were wise enough to mention the only two that slowed and stopped NSA in the leaks. Keep it up at every opportunity. ;)