News for hackers. Stuff that matters.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by isatty on Sunday July 18, 2021 @01:21PM (#27874669)
While bashing on Telegram is basically HNs favorite past time, there aren't many other alternatives that work just as well in the area that I care the most about: being a great chat app. Telegram is fast, has native clients instead of a garbage web ui (or electron based app), has better stickers than the alternatives, is easy to write bots for and has a larger user base where I'm at than any of the alternatives other than WhatsApp.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by maqp on Sunday July 18, 2021 @03:14PM (#27875669)
>Telegram is fast

It's fast because it's insecure. Would they deploy end-to-end encryption for their groups, it would be as slow as the competition.

>has native clients instead of a garbage web ui

The "native clients" are used to manage data stored on remote server over an encrypted tunnel. The data that sits on Telegram servers is plaintext. The security benefits of the native clients (e.g. handling E2EE in a secure way) don't really play out when the desktop application doesn't feature end-to-end encryption at all, not even for 1:1 chats.

>has better stickers than the alternatives,

My Signal client has 1:1 match of my Telegram sticker packs. This is a non-issue. Porting them is trivial.

>has a larger user base

Sure, which means, larger amount of your data is hoarded by the Mark Zuckerberg of Russia who will take zero responsibility if his servers are hacked.

Just a thought: You should probably refrain from advertising features of an app the glaring security issues of which you can ignore out of your privilege. There's a LOT of oppressive regimes in the world, and people living there shouldn't take your advice.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by isatty on Sunday July 18, 2021 @05:04PM (#27876655)
I don't understand your first two rebuttals. I didn't want to keep talking about the fact that Telegram is not secure by default (there are plenty of other comments that go into this); I understand this, thank you.

It doesn't have anything to do with the competition being unable to provide a fast _native client_ that I can use. I dislike web based apps, and I refuse to use them whenever a native alternative is available.

> It's fast because it's insecure. Would they deploy end-to-end encryption for their groups, it would be as slow as the competition.

Doubtful - it would increase the amount of compute required on their side, sure, but if done properly it will have close to 0 impact on the end users in terms of latency.

> Mark Zuckerberg of Russia who will take zero responsibility if his servers are hacked.

As opposed to _x_ from country _y_ that WILL take responsibility if their servers are hacked? Please.

I will continue to talk about what I like about the user interface and what it does better than the competition because I want to.

[dead]
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by dijit on Sunday July 18, 2021 @07:13PM (#27877564)
> I'd use native Signal desktop client written on Rust if there was an option. But there isn't, thus I'm using the Electron app.

I’m not going to say anything else other than that this is exactly what Signal wants. They are actively hostile towards efforts to make third-party clients.

You can argue that this is for security reasons, and maybe I’d buy that.

But I would also bring up the fact that they implicitly control both ends in a binary distributed or web(thus- backdoorable) fashion most of the time. I’m not saying they do that, but since they have control of the clients, the servers and the protocol it is not impossible.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by maqp on Sunday July 18, 2021 @08:23PM (#27877982)
Reproducible builds help here. Not all clients are supported yet I give you that, but perhaps you can see how far injecting backdoored clients on mass scale is from "user's sending plaintext copy to server every time". Both achieve the same thing, one will fill the headline of every news paper in the world if it happens once, the other happens right now, billions of times a day, and nobody cares. There's no perfect verification process, but there's a metric fuckton of low hanging fruits we need to address first.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by isatty on Sunday July 18, 2021 @07:10PM (#27877537)
HN is not the place where you get to tell people to check their privilege or to warn people. Write a blogpost, or write to dang@ if you feel strongly about it.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by Andrew_nenakhov on Sunday July 18, 2021 @03:21PM (#27875768)
Most of Telegram's advantages come from it not being end to end encrypted by default. They've proven that simply claiming to be the most secure app is better for most users than actually being the most secure. Because true security comes with drawbacks, and users don't like drawbacks.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by t-writescode on Sunday July 18, 2021 @04:50PM (#27876553)
Most of Telegram's advantages come from:

* anonymous accounts. You need a phone number to set up an account, but you don't need to share your phone to talk to people on it

* network effects. The sticker game is great on Telegram because there are so many people on it making stickers.

* the secure chat *option*. It has an option of being secure and so therefore people think it's secure and when they find out it's not secure all the time, they're still happy that it has the option of being secure.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by Siira on Sunday July 18, 2021 @05:56PM (#27877045)
A good app offers options. While most HN folk don't like Telegram's default, it at least gives you somewhat of a choice (and the user can force the E2E option even if the other party doesn't like it)(though they do not give the option to make all chats by default E2e, which is indeed a dark pattern). The other apps just shove possibly backdoored E2E down people's throats with no choice at all.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by maqp on Sunday July 18, 2021 @06:31PM (#27877264)
>The other apps just shove possibly backdoored E2E down people's throats with no choice at all.

You can check e.g. Signal is not backdoored by reading the source code. You can find it here https://github.com/signalapp/

You can vefify the client you downloaded from Play Store hasn't been tampered with. Instructions for that are here: https://github.com/signalapp/Signal-Android/tree/master/repr...

Your post is slightly ironic, considering Telegram doesn't give you any choice on using E2EE for groups or any chats for that matter, on majority of desktop clients. Know this: when you're not using end-to-end encryption, i.e. when you're using Telegram cloud chats, those chats are by definition as private as a backdoored E2EE chat would be. So one could argue they are front-doored by design.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by MikusR on Sunday July 18, 2021 @07:07PM (#27877519)
Signal just spent a year being closed source.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by decrypt on Monday July 19, 2021 @03:12PM (#27886229)
This is an incorrect take. It's true that Signal's server code had not been updated for many months, but that doesn't have an effect on security as messages are end-to-end encrypted locally on the clients/apps. The client/app code had been consistently updated during this period.

Signal responded to the server code being outdated as well: https://github.com/signalapp/Signal-Android/issues/11101#iss...

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by einpoklum on Sunday July 18, 2021 @06:52PM (#27877410)
But Signal won't allow federation and is hostile to independent client development.

Also, what about the server source code?

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by maqp on Sunday July 18, 2021 @07:10PM (#27877538)
Signal choosing not to federate isn't in the scope of this thread. IMO not federating is a guarantee of client quality. The last time I had a look at Matrix E2EE implementations, only Riot was barely usable. There was a ton of clients that didn't support it, didn't intend to, didn't have the confidence to go about implementing etc.

>Also, what about the server source code?

https://github.com/signalapp/Signal-Server There you go, last commit two days ago.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by tffgg on Monday July 19, 2021 @02:21AM (#27879786)
It's not a quality guarantee for clients in practice since you can't control Client Side with open source

Regarding Riot, which is called Element by now, you may want to update your knowledge

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by maqp on Monday July 19, 2021 @10:45AM (#27883213)
Yeah I'm familiar with Riot being called Element now, and you don't need to look further than this chat for me to state that.

"It's not a quality guarantee for clients in practice since you can't control Client Side with open source"

What? Open source native client with reproducible builds is literally the gold standard of individual control over software, it's even GPL licenced.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by int_19h on Monday July 19, 2021 @01:23AM (#27879538)
> IMO not federating is a guarantee of client quality.

That's not how it worked out for email long-term; why should IM be any different?

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by Siira on Monday July 19, 2021 @06:20PM (#27887962)
I know Telegram’s default is to give all the data to the server. But being honest about this makes them able to provide significant UX benefits. Backdoored E2E gives a false sense of security, and, backdoored or not, E2E’s UX sucks. (Also note that the backdoor/malware could be on any level of the stack; From WhatsApp’s plaintext backups, to NSO’s zero days.)
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by jeroenhd on Sunday July 18, 2021 @09:54PM (#27878469)
This, seriously. I've switched to mostly Matrix chat (with bridges for backwards compatibility) and I seriously miss the Telegram experience. The official client is your run-of-the-mill Electron chat app, which comes with the standard lag and slowness. Alternative clients either lack features (notably, E2EE), are based on other terrible frameworks (Flutter, QML+Python, etc.) or don't work on the platforms I want to run them on.

Spectral looks promising, and so does Fractal in a sense. I've had trouble getting E2EE to work on Nheko, and everything that does work is slow and laggy compared to native chat clients.

One day I will probably get frustrated enough to hack the Matrix protocol into the Telegram app, if someone hasn't done that already by then. The lack of anything better annoys me to no end!

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by baby on Sunday July 18, 2021 @08:45PM (#27878105)
I use WhatsApp and it's pretty good for messaging friends, for large group chats I don't find telegram useful and I find discord much better.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by Laarlf on Sunday July 18, 2021 @02:32PM (#27875252)
Matrix or IRC if you hate yourself.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by isatty on Sunday July 18, 2021 @02:37PM (#27875294)
I'm a long time user of IRC, but none of my irl friends are on it (anymore).
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by Laarlf on Sunday July 18, 2021 @02:51PM (#27875411)
IRC for me is a way to seek support for OSS projects and nothing else. It is not really designed to be a IM for phones...
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by theshrike79 on Sunday July 18, 2021 @03:15PM (#27875687)
IRC is really hard to do well, since it requires a consistent TCP connection to really be useful :/
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by joecool1029 on Sunday July 18, 2021 @04:17PM (#27876295)
Combine the parent's suggestions and use IRC over a Matrix bridge. It checks all the boxes when it works and as their comment mentioned you'll learn to hate things when it doesn't!
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by dijit on Sunday July 18, 2021 @07:15PM (#27877579)
FWIW there are efforts to prevent this being required and they’re quite close to mature.

I currently run an irc network with these extensions.

Re:Cryptographers unearth vulnerabilities in Telegram's encr... by Siira on Sunday July 18, 2021 @05:58PM (#27877053)
I use a self-hosted web client (thelounge). It's still not good for casual IM usage IMO, but it's pretty painless.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by mehdix on Monday July 19, 2021 @03:26AM (#27880067)
I run znc on my server and Revolution IRC on my android phone and couldn't be happier about it.
Re:Cryptographers unearth vulnerabilities in Telegram's encr... by shapefrog on Monday July 19, 2021 @02:59AM (#27879970)
> has better stickers than the alternatives do stickers count as a feature or a bug?