It's a long read, but it's an interesting discussion on how to test API services. It's definitely counterintuitive for me being more familiar with testing pyramid.
I found interesting the section about full-stack boxes, something that I have used in the last 2 jobs I worked on and I confirm the authors consideration about maintenance.
Also, the idea that testing is spectrum between pre-production and production is mind blowing to me. Especially, the metrics provided about how low value pre-production tests are. I wonder if it's a chicken/egg situation where the lack of pre-production testing would increase the number of bugs that would have been found pre-production.
Is this something you are applying on your workplace, please share your experiences and any relevant reading material.