Why do people use c ++

What is C ++ still used for today?

-portability
I think it is also used when portability is important, because the software or even just libs from small 8-bit chips (e.g. ATMega8), via mainstream cpus (cell phones, tablets, desktops, laptops) to HPC (NVidia Titan, Intel XeonPhi) runs everywhere and almost every OS supports them.

-future security
Microsoft had relied on C # for a while (e.g. XBox XNA, Windows Phone 7) and was rejected by many developers because of this. They can't let it die because so many companies rely on it and would sue MS, that's why it will now be "community driven". I think this uncertainty about the future is also what makes many new languages ​​risky for companies. Languages ​​die, developers die, software lives unknown for a long time and who knows whether C #, Rust, GO, Java etc. will be around in 10 years. Of course, nobody in c ++ can guarantee that, but if I should rely on something, it would definitely not be D, E, F, Go, Rust, ...

-quality
since c ++ is more complex to master, it has the disadvantage (for managers) of being more expensive than higher level languages ​​(don't say you don't have "can't we hire a junior from the university to do that in java / c #?" programmed, it doesn't have to be able to / have xyz "). but at the same time, if something is to be really good and money plays less of a role, then c ++ is called for, even if it doesn't really bring any benefit. This is similar to how some think that they have to write their software in assembler so that it is faster, and with their 0 experience beat c ++ compilers in -O0. but the call is there and that's why some UI intensive applications are written in c ++. I think that since new languages ​​are used more by younger people and c ++ more by old hands, your impression tendency goal is correct, which of course is unfair to the bad c ++ and outstanding c # / java ... coders.

-flexibility
if it's not politically driven, then almost everyone offers a c ++ interface for their APIs. Of course, wrappers etc. are developed for other languages ​​at some point, but often there is the best support for c ++. the alternative languages ​​themselves often have no intrinsic support for interoperability with more than themselves and c ++. (e.g. java <-> C #, python, lua) and support for the latest latest language features is often only available in the c ++ based libs (e.g. the latest ECMAScript features).

-userbase
Hiring new developers is often not easy. if you are looking for a senior or principal who has 10 years of experience in his field and knows c ++ and is good, then you can find that java, c # etc. will be tight.
Even specialists (by that I mean the highly paid people, not something like C # -UI 'specialist') can usually be found with c ++ experience. Even people from research e.g. universities who are specialized in e.g. compression, database optimization, network traffic, AI, computer vision, ... are now mostly c ++ coders (in the past, surprisingly, these were often pascal people). java / c # is often taught as an introduction, but at some point the doctoral students work with c ++.
The reason why e.g. 3D engines are often 'given away' to non-profit people, it is not really so much the will to do good, but to have a 'hire base', because the best language / software is of no use if companies cannot hire someone to handle it.
that's why c ++ is often better from a company perspective. From the developer's point of view, it is of course a job that secures the job if you develop software that couples pearl, make, python, java and c ++.

I think we developers like to fall into the line of writing something new and "better" from scratch instead of maintaining, fixing and improving existing things, although we know that it is more or less an "antipattern". that's why we like to develop a new language right away, often it makes one thing "better", but limited to 100 other places.

I think c ++ alone would be an advantage for everyone. Juniors would not be seen as code monkeys (as with other languages), nobody would have to explain to non-programmers why the buz about a new language would not save the world, all platforms would be usable for all, because the focus of all would concentrate one can expect better compilers, more libs, etc.