Re: Two Questions
2. If telegram is truly a secure E2E platform, how can they possibly determine the content of a message in order to cooperate with authorities?
Because it isn't -- it's a proprietary, effectively closed source* app on a phone. That's about as far from secure as you can get. I am assuming that the open source desktop variants are not relevant here, for the sake of argument, since the vast majority of users are on some kind of cell phone and the open source version should be verifiable to actually do what it says on the tin regarding encryption.
If they're not bothering to MITM the traffic directly, it's probably some proprietary local filter in the app detecting keywords. Where it gets interesting is what happens when that filter detects unsanctioned content -- does it send out a trace to the authorities and pretend to send encrypted? What exactly happens in that case?
Other possibilities could be metadata checks -- better hope you're not within some degrees of a known bad apple, or you'll be tasting Putin's special polonium dessert...
* If you can't compile the app yourself, it could have any special modifications required for the app store / local authorities, and you can't check it or fix it.