Re:Telegram has launched a pretty intense campaign to malign...
by kaba0 on Monday May 13, 2024 @08:30AM
(#40342552)
“Rolling your own crypto” is discouraged for programmers, not for field experts. It’s not your average joe’s first try at encryption writing a caesar cypher…
Re:Telegram has launched a pretty intense campaign to malign...
by ants_everywhere on Monday May 13, 2024 @09:14AM
(#40342980)
It's discouraged for field experts too. In practice, real crypto schemes go through several rounds of analysis by multiple teams of experts, often working against each other. It's unusual these days for a single company to come up with a custom crypto scheme. It was probably more usual toward the beginning of cryptography.
For example of this sort of vetting, take a look at the standardization around AES or the post-quantum schemes.
In crypto you're almost always relying on hardness assumptions that aren't provable yet. So you need to guard against things like accidentally haven chosen the wrong constants that collapse a problem's hardness. Or, more mundanely, making a seemingly reasonable engineering choice that is known to weaken the protocol and which would be caught by a big org with a thorough review, but a startup may not catch.
Researchers of Telegram's protocol have said in some ways it's weaker than TLS.
E.g.
- https://www.wired.com/story/the-kremlin-has-entered-the-chat...
- https://therecord.media/telegram-blocks-chatbots-used-by-ukr...
- https://www.oporaua.org/en/polit_ad/and-telega-is-still-ther...
- https://www.pravda.com.ua/eng/news/2024/05/8/7454849/
- https://time.com/6280190/cia-recruit-russian-spies-telegram/
- https://www.cia.gov/stories/story/cia-launches-telegram-chan...
- https://nordvpn.com/blog/is-telegram-safe/
- https://portswigger.net/daily-swig/multiple-encryption-flaws...