Telegram added auto-detection of the programming language used when monospacing a code snippet.

It does a pretty nice job, but there are issues with this approach due to similarity in syntaxes across a huge number of languages.
The guess can be different (wrong) from what the user actually intends.

For instance, language for this snippet is detected as "C":

#include<stdio.h>
printf("Hello World");

Syntax wise, it is perfectly fine, but what about the actual code?
Is it a valid executable C code? No, it lacks the entrypoint, the main() method and hence would not even compile, let alone execute.

But the same piece of code is perfectly valid and executable in Ruby.

The first line #include<stdio.h> is ignored in Ruby because it is treated as a comment, the next line is executed and the program outputs "Hello World" on the screen!

Conclusion: While the auto-detection of the snippet language is a very cool feature, you may not always see it detect what you intend, for such cases you can always explicitly mention the language you want and TG's auto-detection algorithm won't override it!