Unsafe in C# is much more dangerous than unsafe in Rust, precisely because it doesn’t actually color a function. It just allows its body to use pointers. This is why you have methods in the CLR called “DangerousFoo()”, and the compiler does nothing to prevent you from calling them.
This one is more subjective, but I also think C# has a more mature and painless web stack.
I love both languages but for me they each fill a different role.
C# also has a pretty steep learning curve if you have to care about the things where Rust excels, like correctness or maximum efficiency (which typically does not include web stuff). I would even say that Rust is the easiest language in which to approach that level of correctness and efficiency.