I don't know the career histories of these kinds of people, I'm sure its varied, but I just can't imagine them working in very large public facing dynamic sites that update multiple times a day with monoliths thinking "this sure can't be improved! We have reached the peak of computing!"
I've seen microservices go really badly of course - I worked in a place where the devs insisted that two microservices need to go out "at the same time," as if such a concept existed, because they depended on each other. At the end of me working there, there were around 40 microservices, all Java so they needed at least 2GB of memory each, some 8GB, and at least 3 replicas for high availability. Cost a small fortune in servers for what really could have been like, 5 microservices, written in Go or Node and run on a handful of normal servers.
But microservices, like Kubernetes, are not _hype_, they are not a flash in the pan and they're here to stay because they are a good idea conceptually, even if they often aren't executed very well.