I don't think that casting a range of bits as some other arbitrary type "is a bug nobody sees coming".
C++ compilers also warn you that this is likely an issue and will fail to compile if configured to do so. But it will let you do it if you really want to.
That's why I love C++
Well as long you know what you’re doing and weigh the risks with the benefits you’re probably ok.
In my experience in the industry, there’s little benefit in pretending you’re John Carmack writing fast inverse square root. Understanding what you wrote 6 months ago outweighs most else.