From my experience C++ programmers talk more about language features and quirks than about real problems — you’ll rather hear about things like “perfect forwarding” or “xvalue references” than about applications. You can learn how revolutionary constexprs are from people who never heard about partial evaluation and Futamura projections.
Bjarne Stroustrup wrote a book for beginners, titled “Programming. Principles and Practice using C++”. Therein, he devotes a lot of content to the development of a command line calculator.
Now compare it with “Structure and Interpretation of Computer Programs”. The book is half the length of Stroustrup’s book, but it covers topics like symbolic differentiation, digital circuit simulation, logic programming and compilers.
I mention that, because from my perspective C++ is a language for creating problems rather than for solving them — and Stroustrup’s book focuses mainly on solving the basic problems that C++ creates, rather than on solving actual problems. And worst of all, this patological situation is perceived by many programmers as a norm rather than abberation.