> If you hire the best engineers....
Guess what, there is no broad consensus on what "best engineer" means. I bet your org is rejecting really good engineers right now because they don't know what Kubernetes is. Same goes for literally any other technology that has been part of a hype cycle (Java in 2001, Ruby on Rails in 2011, ML in 2011; no the precise years don't matter).
> ...trust them to make the best decisions.
A lot of work encapsulated there in less than ten words. If you hire a bunch of people and tell them "you are the best", you think they are going to sit around and run the Raft protocol for consensus on deciding how to architect the system? No, each of them is going to reinvent Kubernetes, and likely not in an amazing way.
Microservices are often best deployed when there is a mixture of cultural and engineering factors that makes it easy to split up a system into different parts that have interfaces with each other. It has little to do with the superiority of the technical architecture.
----------------------------------------
[1] Looks like the article was written by the CTO of Amazon, which...surprises me a bit. Then again, from all accounts, Amazon's not exactly known as the best place to work; so maybe I'm right? In any case, anything written by Amazon is not directly applicable to the vast majority of small-to-medium companies.