This phrase is tiring to hear in this form, and your understanding seems to be incomplete here. Signal also rolled its own crypto, but you don’t see anyone saying it’s insecure for that reason. That phrase is used to tell non-cryptographers not to roll their own crypto because of the high chances of vulnerabilities being introduced. In the case of Telegram, the company defends its protocol saying that it’s been created by people with PhD in mathematics (which is related to and is foundational for, but different from, cryptography). Telegram’s encryption protocol (the second version) has not been broken by anyone till date.
It was created by Nikolai Durov who has a PhD in geometry. That's like a gynaecologist performing brain surgery. Specialization matters. Sure both took human anatomy 101 class in college, but somewhere along the way they went and spend their ENTIRE career doing different things. It's easier to get another decree in medical science sure, but in this case the gynaecologist did not, they just started cutting the brain with kitchen knife and just because their patients haven't died yet doesn't mean they have the credentials to abandon best practices.
> The reason why that statement exists is because there are _countless_ examples of teams coming up with their own, new cryptographic mechanisms that either break...
But aren't there _countless_ examples of this in crypto made by cryptographers?
I'm not playing devil's advocate, I don't really have a stake here. :)
1. Actual cryptographers usually design with a set of constraints that make their crypto work: those might be about compute power, or memory bandwidth, or what have you, that make an algorithm difficult to brute force.
2. The algorithm will typically be peer-reviewed to try to weed out mistakes, either fundamental mathematical ones, or in the assumptions.
3. The implementation then needs to be high quality.
There are certainly no shortage of examples where systems which pass 1 & 2 are undermined by failures in 3. All algorithms are susceptible to the context around 1 changing (changes in compute power or whatever).
When you go it alone, you're assuming that you won't make any mistakes any of these. That seems a pretty tall order.
In machine learning, a model that works 90%, 95% of the time is pretty good.
A glitchy video game can still be fun as long as the glitches don't happen too often or cause a loss of too much progress.
Even a filesystem that doesn't lose most people's data, most of the time, will have a lot of adherents.
But if your cryptography implementation isn't completely perfect, it's frequently just 100% useless for its intended purpose.
But a bad crypto implementation will work. For all intents and purposes, it will appear completely fine. Users will get their messages. The bitstream will appear completely random. At least, until somebody with expertise in breaking crypto systems digs into it.
And that applies even if they're using AES but wrong mode of operation. That applies even if they're using best practices like AES-GCM but the CPU doesn't support AES-NI and a cache timing attack allows key exfiltration.
Like Swiftonsecurity wrote:
"Cryptography is nightmare magic math that cares what kind of pen you use. Should math care what kind of pen you use to implement it? No, but Fuck You, this is Cryptography."
The attacks are incredibly subtle for even the best systems, and Telegram is so far away from even adaquate it's difficult to emphasize it so I'll try with my best restraint:
TELEGRAM FUCKING LEAKS EVERY GROUP MESSAGE TO THE SERVER WHICH IS THE EXACT EQUIVALENT OF A FUCKING BACK DOOR.
I hope that didn't leave anything unambiguous.
The argument I've heard is that Telegram uses their own encryption protocol. The rule of thumb in cryptography is "don't roll your own crypto".
The reason why that statement exists is because there are _countless_ examples of teams coming up with their own, new cryptographic mechanisms that either break (intentionally or not) or were written with a backdoor. People get incredibly clever when it comes to breaking encryption.
AFAIK the only way to be on the right side of this argument is to use a time-tested encryption protocol. However, there are even instances where some protocols have been live and in production for x years before discovering that a backdoor has been in the code since day one.