Hacker News new | past | comments | ask | show | jobs | submit login

Get news to see other services following in the footsteps of WhatsApp!



Telegram had e2e private chats for over a year before WhatsApp added e2e


The problem with Telegram is that it doesn't default to E2E; using it means to go through that additional step of starting a secret chat which I venture a lot of people won't do unless they're very conscious of it.


Does bad crypto counts? I don't think so.


I keep seeing this accusation get flung around whenever someone mentions telegram but never proof. I remember telegram had a vulnerability once and then it was patched just like any other security software.

You shouldn't throw around accusations without proof.



Is it cryptographicly broken or not? Can I read the plain text of the traffic I capture?


https://tonyarcieri.com/all-the-crypto-code-youve-ever-writt...

The word "broken" means susceptible to practical attack, and attacks aren't always of the "cryptanalyze the ciphertext and read the plaintext because you're a clever mathematician" variety.

For example: Padding Oracle Attacks. This is the most accessible explanation on-hand: https://twitter.com/SoatokDhole/status/720435675401744385

A padding oracle attack lets you decrypt a message by studying how the cryptosystem responds to garbage input. Without recovering the key.

iMessage had a compression oracle attack recently: http://blog.cryptographyengineering.com/2016/03/attack-of-we...

They didn't merely "read the plain text of the traffic [they] capture[d]". But these systems were still, quite badly, broken.

So what's my point? Telegram's protocol is susceptible to the same class of active attack. Thus, it is broken.


To add, Telegram's crypto is completely and totally off the walls crazy in terms of design. Add to that the fact that there are cryptographic breaks (though not we can read your ciphertext breaks), and you should be careful.

iMessage would have been reasonably secure had they used AEC-GCM or a MAC. The design at least made sense: compose a scheme out of known primitives. They just missed (very important) details. Telegram is just turtles all the way down.


Yes, but large, established services switching is another ball game. Even is Viper is not nearly as big as it used to be.


signal?


Signal is a niche service. Its selling point is security. Viper and WhatsApp are general services and their selling point isn't security.

I'm not trying to imply that WhatsApp are the first secure chat, but I am saying it is the first major platform to switch to end-to-end security and this looks like the first step in a trend of switching.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: