The problem with that statement is you have to in a very small time period pick enough corners to find what they know and not the many parts you think are not e

The problem with that statement is you have to in a very small time period pick enough corners to find what they know and not the many parts you think are not esoteric but are unused by many people. What isn't esoteric? Shared pointers, move semantics, atomics, inline assembler, CRTP... If you quiz someone on your particular known C++ language details you can far too easily think everyone is pretending.