I think you've articulated your point quite succinctly.
> Might as well poll dentists on the best way to build software.
A programmer with hubris would take VC money to "make better dental software" never having met a dentist or having done any dental work, perhaps only ever taken a scant look over at the monitor and deciding it's not fast enough or whatever. Then they'd build an MVP, try and sell it to some clinics, and upon being laughed out the door, would take more VC money to try and reinvent dentistry because their parents never told them "no"
Plenty of leaders in charge of SWEs will screw up software by itself, let alone a city. I've seen it a dozen times, someone comes in who thinks everything is simple (or can be) and fails to understand that software mirrors the org chart more than anything. They they force some overcomplicated in-house solution that's supposed to unify everything, and it fails less than halfway.
What bugs me is that the programmers of today seem to think they can optimize everything. Not only are they incapable, they're simply unqualified. At least professional urban planners have studied the topic in depth. Google throwing an engineering team at cities is just ridiculous. Might as well poll dentists on the best way to build software.