Don't get your point. I was referring to this statement from OP: Google used to have a reputation for being an "engineer led" organization, I guess that's slowly changing over time.
In enterprise applications (which solves for workflows - business, IT etc.) one needs lot of PM thinking for feature identifications, what to prioritize, gathering requirements etc.
In Search defined Google one did not need PMs as requirement is pretty straight forward from feature perspective, but one needs deep engineering and computing science knowledge for making the best search algorithm. Whereas in Cloud IT world the focus shifts more towards product & roadmap planning.
I think you're over-inflating the value of people fetching specs back and forth between a pool of customers and design / engineering.
If you're a software or hardware company, ultimately the people who design and build the software / hardware have to design and build the thing, everything else is ancillary to those tasks.
In enterprise applications (which solves for workflows - business, IT etc.) one needs lot of PM thinking for feature identifications, what to prioritize, gathering requirements etc.
In Search defined Google one did not need PMs as requirement is pretty straight forward from feature perspective, but one needs deep engineering and computing science knowledge for making the best search algorithm. Whereas in Cloud IT world the focus shifts more towards product & roadmap planning.