Hacker News Reader

Comment by EGreg | Hacker News Reader

MegaUpload, LastPass and others are provably not able to access your information. Storing the encryption key yourself is the way to go.

But then the govt can capture you and make you give up the key. A whistleblower can threaten to have more incriminating evidence disseminated encrypted somewhere, and if he doesnt check in every 30 days it gets released ... but then the government can just torture him until it makes him give up the key he uses to check in every 30 days. It wouls take a really stubborn guy to persist and let the information be released. Since you dont have any information like that anyway, just assume that if you can access your own data the govt can compel you to do it for them.

View on HN

MegaUpload, LastPass and others are provably not able to access your information

Where can I find the proof you're talking about?

If you want to prove it to yourself, fire up Wireshark.

Sorry, I'm too tired to continue this bullshit conversation in a different tone, so here's a somewhat angry rant.

EGreg said "provably not able to access your information", which means that there's a proof of some kind. What the fuck does "prove it to yourself" in your comment mean? That I'm now responsible for writing the security proof? That you acknowledge that no such proof exists, but I can write it myself?

Oh, you probably didn't mean "prove", you meant "ensure", as in I can ensure that the traffic between Mega and my browser is encrypted. What does this say about security? Absolutely fucking nothing!

It's a known fact that Mega is NOT secure. Even if you store your key locally, and they don't know it today, they can do the fuck they want with their web pages: next time you login, they can send you a piece of JavaScript to decrypt everything, or read your key, or... just do anything! If you're looking for an example, this happened with HushMail (http://en.wikipedia.org/wiki/Hushmail).

Anyway, I didn't really mean to ask for a proof, because I know that there's no security proof that "Mega, LastPass and others" are not able to access my information, I was looking for a way to encourage EGreg to do some research and declare that he was wrong, because his claims are fucking dangerous. I certainly didn't expect replies like yours.

Lastpass does encryption client side. The difference is that would require an active attack (pushing new code in an update) with a chance of being spotted by a reverse engineering attempt.

Who says that the backdoor hasn't been shipped since the first release and just sits there idle waiting for a nudge from the server?

A larger point that security is based on trust. Trust is based on assumptions, especially and exclusively when it's a closed source software. Assumptions that what the company says is true. Now we know that they can be forced to lie and any claims of the security just crumble as a house of cards.

Even if LastPass would've been an open-source, it made a little difference as the company could've been forced to distribute binaries made from altered sources. It wasn't easy to build trust before, but it's going to be nearly impossible to build it now. Though it's not to say there isn't plenty of people who are after pseudo-security and who are easily lulled by cross-my-heart promises squirting out of every second company now.

True, they could have shipped a backdoor since day 1 but it's unclear what their incentive would be to do so and they run the risk of it being found with a reverse engineering attempt.

Is there any basis in law under which the government can force a company to distribute altered binaries? If the government can force you to add a backdoor in your own product, in effect they have the power to demand that you perform free labour for them.

That would seem to form a basis for also demanding that a person work as a spy for them etc.

You haven't been paying attention recently at all, have you?

What is that supposed to mean?

There are "incentives" that are shoveled down companies throats and secured in place by gag orders. Perfectly legal too.

You're right, I used the word "prove" way too loosely.

https://www.grc.com/sn/sn-256.htm

Gibson is about as credible as I am when it comes to security - ie, not very.

Credibility doesn't matter, as I'm not asking for a confirmation from an expert: EGreg claims there is a security proof of some kind.

Sorry, this is the first time I'm seeing a security proof in the form of interview, so I can't find it. Could you point out the relevant part?

Also, what about Mega?

If you use forward secrecy, you shouldn't have any key to give, right? DDG recently announced support for forward secrecy.

Are you sure? And in any case how would you know? Perfect forward secrecy is hard to implement with TLS:

https://www.imperialviolet.org/2013/06/27/botchingpfs.html

"So how do you run forward secrecy with several servers and support session tickets? You need to generate session ticket keys randomly, distribute them to the servers without ever touching persistent storage and rotate them frequently. However, I'm not aware of any open source servers that support anything like that."

Reply on news.ycombinator.com