just different economics, homeshop is trying to control total costs whereas commercial shop is trying to maximize throughput against overhead.
That strikes me as an insightful observation. And obviously the "total cost" criterion applies, to varying degrees, to shops which are supporting a technical or development effort.
The economics of, say, making a hob to do one wormwheel of a non-standard profile just don't stack up, when you can use a relatively simple flycutter to hob it to the same accuracy - I choose this example to illustrate that flycutters are not restricted to milling machines.
I do a bit of design and methods consulting for machine, mechanism and product development, and I used to get frustrated sometimes when a client's in-house shop would throw up their hands at doing prototypes, because they couldn't shake off the 'maximize throughput against overhead' imperative burned deep into their patterns of thought.
That's one reason I gradually set up my own prototyping shop. The real reason is I love machine tools, always did, always will ....
As well as flycutters vs facemills (or hobs), "McGyver's law" explains why shapers can still be relevant and useful to HS operators, similarly pantographs. (Both the topic of recent discussions, no doubt there are plenty of other examples)
As others have said, tooling cost tends towards trivial when doing it the 'old' way.
Just as importantly, when the inevitable design tweaks crop up, it's often relatively easy to make small changes to that tooling and run the job again.
Having said that, we all know the 'break even' batch size for CNC has not bottomed out yet, and of course there are some types of prototyping work where manual machines do not cut it, especially where the finished part will be made by CNC, or moulded.
However for the small proto shop I quite liked the recent idea from another thread of a rapid proto (3D printed) master in resin, duplicated by a pantograph -- or a copy turning lathe, for turned shapes.