Great post!
I've never been a data scientist but I am an entrepreneur from an engineering background and I've managed teams with data scientists and data engineers onboard.
I totally feel the pain delivered through the points you outlined. In many teams, honestly it's a constant struggle to empower data scientists while keeping things pragmatic.
Good data scientists require deep industrial experience (sometimes people classify that as "non-technical" or "business acumen", not sure why) and robust background in theories (both scientific and humanistic). In the tech industry we often value engineering expertise over the skill sets and knowledge that make good data scientists. Thus more often than not, data scientists are not empowered enough to do well - either their voices and decisions get overridden by engineering, or they simply weren't given enough leeway or resources to make real impact within the organization.
Anyway, lovely post!