Live data from Hacker News

I'm giving up on PGP

blog.filippo.io

121–130 of 350 posts

Re: I'm giving up on PGP

#121
post #94
post #62

Earlier quoted context omitted.

Keybase has clearly moved away from PGP. They want to use Saltpack whenever possible, NaCl based encryption. They want to solve the problem of multiple devices and not having to share the private key between all of them. As far as I know they are working on a messaging app as well.

They also have the KBFS which is very interesting. But yes, very clearly pivoted away from PGP and are working on other problems.

I don't think they are working on other problems, rather they realised that GPG has limitations and they can not solve it with PGP. The problem they are working on, is the same problem they started with.

Re: I'm giving up on PGP

#122

People who use PGP keys, can you give examples of your use? I'm genuinely curious. Who are you contacting, or who is contacting you? The author says he only receives 2 encrypted emails a year. Not only do I not have a PGP key, I don't think I've ever found myself in a situation where it was even an option to use one.

a) signing scala libraries I release b) ordinary emails to a few of my friends c) facebook notification emails

I don't think I've ever used PGP when emailing a stranger, but I very rarely email strangers in the first place.

Re: I'm giving up on PGP

#123
post #113

> Yeah, about that. I never ever ever successfully used the WoT to validate a public key. If you ever installed a Debian package then you did. A long-term identity as "Bob Jones" might not be terribly useful - but that's not the kind of long-term identity we care about a lot in real life either. A long-term identity as "Debian release manager" or "Signatory on bank account xyz" or even "Wikileaks committee member" is…

You understand the difference between a lone hacker as an APT vs a state level threat as an APT.

That distinction is huge and chooing not to defend yourself against one or the other may allow for huge convenience gains at the cost what is to many a purely hypothetical notion of security.

Can we improve the tools and techniques we have enough so they are convenient enough to not have to make such a choice?

Re: I'm giving up on PGP

#125
I use MacGPG to sign commits on GitHub but I have to admit that the E-mail portion has fallen by the wayside. (The last time I really used it was to E-mail a professor of a cryptography course for an assignment!)

While I also don’t know many people that use this for E-mail, it doesn’t help that virtually every OS update in the last 5 years has consistently broken it, taking sometimes months for a fix.

For those reasons, this needs to be baked into the OS to be viable. Only when somebody like Apple can install it by default, and make sure it works between updates, will it have the reliability and widespread availability that is necessary for success.

Re: I'm giving up on PGP

#126
post #64

Earlier quoted context omitted.

The standard for adoptability isn't the average person at their peak hours of attention and focus. It's the drunk teenager at 2 in the morning fumbling around in the dark.

A functional WoT should be no more difficult to use than managing your Facebook friends or contacts on your phone. Neither of those are difficult tasks, and are achieved by normal users every day.

Yeah I'm sure it's ridiculously simple and all the experts here saying it looks like the idea itself is fundamentally flawed are wrong.

Re: I'm giving up on PGP

#127
post #113

> Yeah, about that. I never ever ever successfully used the WoT to validate a public key. If you ever installed a Debian package then you did. A long-term identity as "Bob Jones" might not be terribly useful - but that's not the kind of long-term identity we care about a lot in real life either. A long-term identity as "Debian release manager" or "Signatory on bank account xyz" or even "Wikileaks committee member" is…

> If you're using iOS you've already given up against state-level attackers. Wasn't the recent apple vs FBI debacle evidence to the contrary?

There's also this https://blog.cryptographyengineering.com/2016/11/24/android-...

Re: I'm giving up on PGP

#128

Good points, but also I would like to point out that https://www.usenix.org/system/files/1401_08-12_mickens.pdf linked from the blog post was an entertaining read so for anyone that didn't read said PDF, do.

Very entertaining. Thanks for highlighting this and introducing me to this guy's writing.

Re: I'm giving up on PGP

#129
post #122

People who use PGP keys, can you give examples of your use? I'm genuinely curious. Who are you contacting, or who is contacting you? The author says he only receives 2 encrypted emails a year. Not only do I not have a PGP key, I don't think I've ever found myself in a situation where it was even an option to use one.

a) signing scala libraries I release b) ordinary emails to a few of my friends c) facebook notification emails I don't think I've ever used PGP when emailing a stranger, but I very rarely email strangers in the first place.

I was only thinking about keeping communication private. I didn't think about uses like signing libraries. Great example, thanks.

Re: I'm giving up on PGP

#130
post #91
post #5

After all that, he was only getting two encrypted emails a year! Damn. That's crazy.

Stuff to secure has moved away from email while gpg stays primarily an email project. The concept of a git repo means I don't need to sign anything, I'll just roll back if I pull the wrong thing. Socialization at work hasn't atomized enough that my only human contact with a coworker would be a gpg signed commit anyway. The concept of software distribution being a tar.gz.gpg or verifiable md5 file is obsolete. Behind…

> The concept of software distribution being a tar.gz.gpg or verifiable md5 file is obsolete. Behind the scenes something like apt-get does sign things but how to integrate its list of keys with the end user is a mystery, its essentially magic. Besides it provides no security due to lack of MITM attacks in practice.

The very first bootstrap is impossible to do securely in the general case short of building a computer from scratch, but you can do things that make it difficult to attack e.g. ask a bunch of different friends what the sha1sum of the latest debian release should be.

On the assumption that you manage to get a non-compromised version of debian installed you are secure even against MitM attacks; there's a chain of trust, every package has been signed by a key that has a key fingerprint claimed by a specific human maintainer, and new maintainers can only join after at least one maintainer has confirmed their identity against a government-issue document. Of course this doesn't make attacks impossible (e.g. rubber-hoses against one of the maintainers), but it makes the cost a lot higher.

Post reply on HN