I just don't get the huge benefit of using the GO over the subway. Current travel time from Union to Finch is 28 minutes, and TTC has estimated the extension will take 12 minutes, so Union to Richmond Hill is 40 minutes. The GO service is currently 35-37 minutes; it could be speeded up some, but there's also talk of adding stations on the non-stop piece from Union to Sheppard. Let's assume that it takes 30-minutes (which I think is very optimistic), and it comes frequently.
So from Langstaff directly to Union it would save 10 minutes? But what about from Royal Orchard to King? Clark to Dundas? 16th Avenue to UT?
I'm just not convinced that it's going to relieve Yonge much; it may even add to Yonge, if passengers from Richmond Hill GO (which is on Major MacKenzie) start taking GO to Langstaff, and then changing to subway.
Not that I don't think that it either shouldn't be built (though the names have to change; having Richmond Hill Centre subway station at Langstaff GO station with another station called Richmond Hill GO station, is going to cause confusion).
It all depends on where one's point of origin and end-destination are located. I'd imagine that new stations at John/Bayview, York Mills, Concorde and possibly the West Don Lands will not add significently onto travel commute times if one's heading for a trip-generaror within proximity of a stop or one that is easily commutable to via a short bus or LRT/subway feeder. If we were to assume those truly seeking the downtown core in a hurry would prefer the GO train which can run at headways of every five minutes, the justification for a RHC subway extension quickly loses all steam. South of Steeles ridership levels are expected to be 9,100pphpd by 2017. North of Steeles that figure is only 7,100pphpd even by the year 2031 or far well below the minimum threshold for subway operations. The Eglinton corridor meanwhile based on current-day ridership levels would have a minimum of 9,375pphpd if built as a subway line.
Overanalysing where a handful of commuters at Royal Orchard need to go so badly that only a multibillion dollar subway extension can appease them, does the transit-using community at large a huge disservice. A more practical solution for the majority is a minor subway extension to Steeles, met by dedicated BRT (or LRT) lanes down the median of Yonge through Thornhill. This reserved corridor would then intravenously link not only RHCC to the subway line (via bus lanes with queue jumping signal priority, Steeles to RHC wouldn't exceed 15 minutes, if that); but also continue north along Yonge to directly serve Hillcrest Mall, Major Mackenzie and Bernard Terminal.
Given that the subway extension would cover the distance in 12 minutes as you yourself have pointed out, and the headways north of Finch Stn would be every 5 minutes rather than every 2'35"; it begs the question, why promote the most expensive option when the combined carrying capacity of GO trains and LRT can more than supplant its purpose? GO trains can carry 1,800 seated passengers per trip. Bi-articulated buses can carry loads of 270. And 4-car LRT trainsets around 750 per trip. Combinations of the three modes would best cater to local regular, local express, and long-haul interregional transit clientele in ways a single subway line never will.
In summation, if RHC is truly meant to become the next big thing in urban planning, it will become just that with or without the promise of cost-prohibitive subways through low-density, car-oriented sprawl! And before I'm called a hypocrite, at least every other subway line that I've advocated for has preexisting density or popular connecting feeders translating into Day One heavy usage at all stops en route, even through Richview. The TTC itself, meanwhile, readily admits that 4 out of 6 of the new RHC stops will have under 2,000 customers per day.