As someone who has yet to do more than do minimal dabbling in the language, this is a very positive message. It expresses that there is some work to learning this so don't be put off if you experience it, it's normal, and that the work required to learn it pays off in the end.
That's probably a more appropriate message than that it's not hard. I don't think it's appropriate to express to people that learning pointers in C and C++ is "easy". It's not "hard", but it's not necessarily easy for some people. It requires a specific mental model, and depending on how they learned to program, it may be more or less easy for them to wrap their head around. Afterwards, it's easy and makes sense. I assume the borrow checker follows a similar learning hurdle. That doesn't mean we should forget what it's like before we've learned it though (and at this point, there's probably a lot of people in the midst of learning rust that haven't quite fully internalized the borrow checker).
That said, I still occasionally have problems where I feel like I'm doing something "dirty" or "hacky" just to satisfy the borrow checker. It's easy to program yourself into a corner and then find yourself calling `clone()` (the situation, I've been told, has gotten much better in recent releases with improvements to the borrow checker, but alas I haven't had a chance to play much with Rust in nearly a year).
Another thing that I still find difficult is dealing with multiple lifetimes in structs, to the point that I usually just say "to hell with it" and wrap everything in an `Rc<T>`. And sometimes there's simply no safe way (afaict) to do some mutation that I want to do without risking a panic at runtime (typically involving a mutable borrow and a recursive call), which leads to a deep re-thinking of some algorithm I'm trying to implement. That's not Rust's fault, though—it's a real, theoretical problem that arises in the face of mutation. In time, I'm sure there will be well-understood patterns for handling such cases.