Sounds like someone needs to rigorously define a good subset then write a tool to verify that a program uses only that subset.
https://isocpp.org/blog/2015/09/bjarne-stroustrup-announces-...
Such an effort would be so much easier to advocate for in a C++ shop, and more importantly, very difficult to argue against.
This deserves a proper HN discussion. I've resubmitted it, it missed the front page the first time around:
https://news.ycombinator.com/item?id=10545921
(sorry for the karma stealing, lucozade totally deserves the karma if my resubmission takes off)