ie great people don't want to be competing on cost for work, unless they can capture some of the value of their greatness.
eg price for work completed rather than by the hour - so if they finish it in the half the time then they keep that value.
On the other hand, people currently average skill or below ( half... ) are more likely to be better off being paid by the hour.
So if you take a pure economic view, and everyone was behaving rationally, then a cost based hourly recruitment model gets you average and below workers.
ie you get what you pay for at equilibrium.
Of course what they are hoping is to exploit not being at equilibrium - countries with relatively high skills and low costs.
However they are going to be competing with companies based in those lower cost countries ( eg India ), who have both the low cost environment advantage and the ability to recruit locally and better select and retain employees.
My experience of working with even locally based companies that compete on price is high staff churn ( as the good ones find better employment ) - which really damages productivity.
The phenomenon where of Googles/Facebooks/etc. offering mostly very attractive comp in spite of Bay Area CoL is actually relatively recent. Going back a couple of decades I looked at some employment options in Silicon Valley and the comp uplift wouldn't have covered the higher CoL relative to Massachusetts. And the companies freely acknowledged this.
Then you have to ask why do these clusters exist - with high skills, high salaries and high costs?
In my view, that's because there is value created by proximity.
It is more productive to have people co-located in the same office.
Also if you live in a major university town, or high tech area then that accelerates the exchange of ideas, particularly outside peoples direct areas of interest. This makes for better networked, more creative employees.
Co-location of companies also increases the chances of collaborations and deals.