3 ms·

"I use proton mail just for the privacy guarantees enforced by social pressure on their brand." There are no such guarantees. Social pressure on their brand d

"I use proton mail just for the privacy guarantees enforced by social pressure on their brand." There are no such guarantees. Social pressure on their brand did not stop Enron or Madoff from committing fraud. Nor did it stop millions of others from committing various crimes, atrocities, and other unethical acts throughout history. In the realm of email providers, the case of Hushmail[1] serves as an instructive example. Hushmail is an email provider that provides a service similar to ProtonMail, but: "Developments in November 2007 led to doubts, amongst security-conscious users, about Hushmail's security, specifically, concern over a backdoor. The issue originated with the non-Java version of the Hush system. It performed the encrypt/decrypt steps on Hush's servers, and then used SSL to transmit the data to the user. The data is available as cleartext during this small window of time; the passphrase can be captured at this point, facilitating the decryption of all stored messages and future messages using this passphrase. Hushmail stated that the Java version is also vulnerable, in that they may be compelled to deliver a compromised java applet to a user." and "Hushmail supplied cleartext copies of private email messages associated with several addresses at the request of law enforcement agencies under a Mutual Legal Assistance Treaty with the United States.; e.g. in the case of United States v. Stumbo. In addition, the contents of emails between Hushmail addresses were analyzed, and 12 CDs were supplied to U.S. authorities." Incidentally, despite all this, and what you'd expect to be "damage to their brand", Hushmail is still around, and I'd expect many of their users have never even heard of any of this. [1] - https://en.wikipedia.org/wiki/Hushmail#Compromises_to_email_privacy https://en.wikipedia.org/wiki/Hushmail#Compromises_to_email_...

bitdizzy 6y ago
To be clear, I am connected enough to security-conscious social media that I think I would hear about it if someone found it out. I agree with you that someone who isn't at least a little bit diligent would miss news like this. Let me also reemphasize that I don't consider email to be a secure or confidential medium of communication at all, even with PGP. I only want that my inbox is not sold to advertisers and the security practices of my provider aren't utter garbage. Maybe the fact that they're in Switzerland helps me in some ways, but if I had a state adversary I wouldn't bet on it.
- pmoriarty 6y ago
  The problem is that ProtonMail could violate your privacy by sending you Javascript that gave them your keys and you'd never know about it. Relying on hearing about compromises in the news only sort-of works when: 1 - such compromises are revealed 2 - they're big enough to make news in the first place 3 - your own data hasn't yet been stolen, so you have time to change services after you hear about the compromise None of these is guaranteed to happen ever. And even if you did hear about some compromise in the news, it could be far too late for you, as your data (the data ProtonMail is supposed to protect) could already be in somebody else's hands.
  - bitdizzy 6y ago
    > The problem is that ProtonMail could violate your privacy by sending you Javascript that gave them your keys and you'd never know about it. I think we don't agree on how proton mail works. As I understand it, they already have my keys. You can't even give them just a subkey, it only works if you upload a set of PGP keys including the master secret key. What is your understanding of how it works? As for your other concerns, unless I am conversing only with myself, the attack vector for my data is the entire e-mail ecosystem. Even if I only talk to people who use encrypted email, they are also part of my threat model, even if I don't trust a provider with my keypair. e-mail is simply not secure. It wasn't meant to be secure; security cannot be bolted on top. What is your threat model and how do other providers or self hosting achieve your desired level of security?
    - pmoriarty 6y ago
      "As I understand it, they already have my keys." From: https://protonmail.com/security-details https://protonmail.com/security-details "ProtonMail's zero access architecture means that your data is encrypted in a way that makes it inaccessible to us. Data is encrypted on the client side using an encryption key that we do not have access to."
      - bitdizzy 6y ago
        That is a very interesting claim on their part considering that they hold the private PGP keys used by their service. I'll have to figure out what they mean by this. Edit: Ok I see. They store the PGP keys encrypted with your password. Like you said, they could just as well inject javascript to phish your password from your session. But this does seem to mean that if one uses their API directly it would be possible to securely use their service. Thanks for the heads up. There is a third party open source bridge that reverse engineers their API. I think I will look into it to see how authentication is done. https://github.com/emersion/hydroxide https://github.com/emersion/hydroxide

bitdizzy 6y ago

To be clear, I am connected enough to security-conscious social media that I think I would hear about it if someone found it out. I agree with you that someone who isn't at least a little bit diligent would miss news like this. Let me also reemphasize that I don't consider email to be a secure or confidential medium of communication at all, even with PGP. I only want that my inbox is not sold to advertisers and the security practices of my provider aren't utter garbage. Maybe the fact that they're in Switzerland helps me in some ways, but if I had a state adversary I wouldn't bet on it.

pmoriarty 6y ago
The problem is that ProtonMail could violate your privacy by sending you Javascript that gave them your keys and you'd never know about it. Relying on hearing about compromises in the news only sort-of works when: 1 - such compromises are revealed 2 - they're big enough to make news in the first place 3 - your own data hasn't yet been stolen, so you have time to change services after you hear about the compromise None of these is guaranteed to happen ever. And even if you did hear about some compromise in the news, it could be far too late for you, as your data (the data ProtonMail is supposed to protect) could already be in somebody else's hands.
- bitdizzy 6y ago
  > The problem is that ProtonMail could violate your privacy by sending you Javascript that gave them your keys and you'd never know about it. I think we don't agree on how proton mail works. As I understand it, they already have my keys. You can't even give them just a subkey, it only works if you upload a set of PGP keys including the master secret key. What is your understanding of how it works? As for your other concerns, unless I am conversing only with myself, the attack vector for my data is the entire e-mail ecosystem. Even if I only talk to people who use encrypted email, they are also part of my threat model, even if I don't trust a provider with my keypair. e-mail is simply not secure. It wasn't meant to be secure; security cannot be bolted on top. What is your threat model and how do other providers or self hosting achieve your desired level of security?
  - pmoriarty 6y ago
    "As I understand it, they already have my keys." From: https://protonmail.com/security-details https://protonmail.com/security-details "ProtonMail's zero access architecture means that your data is encrypted in a way that makes it inaccessible to us. Data is encrypted on the client side using an encryption key that we do not have access to."
    - bitdizzy 6y ago
      That is a very interesting claim on their part considering that they hold the private PGP keys used by their service. I'll have to figure out what they mean by this. Edit: Ok I see. They store the PGP keys encrypted with your password. Like you said, they could just as well inject javascript to phish your password from your session. But this does seem to mean that if one uses their API directly it would be possible to securely use their service. Thanks for the heads up. There is a third party open source bridge that reverse engineers their API. I think I will look into it to see how authentication is done. https://github.com/emersion/hydroxide https://github.com/emersion/hydroxide