>
People are seemingly scared away from the language by a terrible stigma: the notion that it is a legacy program.Probably more like, they are turned off by how large and complex the language is, and how it's accumulating features like a giant rolling ball of mud.
The discouraging legacy lies in the code bases.
"You can just write code with new features" is a hollow sentiment, because a lot of work is maintaining legacy code, which might using all the features, from C++98 to C++17 and beyond. Old code that successive generations of workers partially modernized here and there.
There is a constant increase in how much you have to know to be a competent C++ maintainer of a legacy code base, under a modernization mandate, as long as the old parts of the code never completely go away.
Also, in the engineering disciplines we have many people with a personality that wants to know something completely. You don't know C++ if you don't know all the old parts too, and that's a turn off for someone who takes pride in being competent. You'd rather master something smaller and simpler completely, than to partially know something huge and complex.
You also don't know C++ if you don't know the new parts, and that's a turn off for long time C++ programmers who have switched to something that isn't constantly adding new unknowns.
Speaking of which, some of the C++ job ads turn away competent long-time C++ programmers by requiring proficiency in modern C++. The C++ community has harmed itself by creating a negative perception of older C++ and older C++ skills, even though those are needed in C++ work. Newcomers pick up on that and get the impression that C++ is a language that they not only cannot master, but which has parts that are frowned upon. The engineering mindset is repelled by something like that.