Alright so first off, 150k will give you a comfortable upper middle class lifestyle in SF/SV (assuming you have no dependents, in which case it's just middle class). It's really not the astronomical salary that you seem to be imagining.
Secondly, it is completely impractical to pay people based on the "difficulty" of their work. Difficulty, even if we can agree on a definition, does not correlate with revenue. Revenue is traditionally where salaries come from.
Intuitively, if person X in a team of n people writes an app that produces profit Z, then it would make sense for person X to get something on the order of Z/n -- not 100k because it's just a chat app. This is not actually how it works because of some artificial constraints. I believe the calculation in most cases looks more like pay = 120k + 20k if in expensive city + MIN(.01 * profit, 150k).
If we're going to say we shouldn't limit ourselves to 5% of workers then perhaps we shouldn't limit ourselves to 5% of the land. There's plenty of places in the country where 150k would be a very comfortable salary.
$150k isn't astronomical in San Francisco (because rents have inflated to match the salaries) but startup employees who are prudent and don't have dependents will likely save money faster than comparable roles in most if not all other parts of the world. And a substantial proportion of the companies in the Bay Area don't generate revenue in excess of what they pay their programmers; many never will, even if they are blessed with unusually productive programmers
The potential profitability of a startup is even more difficult to gauge than the difficulty of a programming role, and the fondness of VCs for disbursing their money in the Bay Area is probably more to do with the distribution of potential acquirers than the distribution of productive programmers.
Every year 100,000 H1-Bs can come in. To believe this article, we'd have to pretend that we have a big wall up preventing these dozens or hundreds of exceptional programmers from immigrating to the US, and completely ignore the existence of the H1-B visa and the hundreds of thousands pouring in every year.
http://travel.state.gov/content/visas/english/law-and-policy...
The US Citizenship and Immigration Services web page has numbers as well.
153,223 visas were issued in 2013, although some of that is to continue the presence of people who were already here in previous years... It's not all new arrivals.
With remote working, geography is somewhat irrelevant. Perhaps instead if changing immigration instead PG promoted companies embracing a remote culture -- then you could have the best of the best with minimal beareaucratic entanglements. After all, given that, according to PG, it's about 'exceptional' talent, working remotely shouldn't require the same level of management overhead as would a lower quality dev 'sweatshop.' However the fact that isn't even being discussed makes one believe that this isn't about finding 'exceptional' talent as much as it's about lowering the scarcity of developers willing to work for lower wages than they might have received otherwise.
Under the PG logic, we ought to simply open the floodgates to all people, because inevitably from the multitudes there's bound to be a few geniuses. But, does the cost of those geniuses outweigh the costs to everyone else? Should we strip mine the Rocky Mountains to find an ounce of gold?
I don't know the answer to the question and I am not advocating either way, however when a company like Facebook begins asking for something, I question the motives -- I don't think Facebook has a talent shortage at all, they have a desire to pay less for their inputs. Which is a fair desire; however we shouldn't just blindly allow policy decisions to happen just because some tech 'leaders' ask for it. It is and always has been about the bottom line.
http://www.quora.com/What-is-the-current-relationship-betwee...