They don't, actually. While the engineering population of SF is growing fast, the increase is a tiny fraction of the total number of software engineers in the U.S, let alone in the world. Remember Joy's Law: "No matter who you are, most of the smartest people work for someone else." [1]
What the Bay Area does have is a greater concentration of talented engineers than any other major municipality. When you're trying to make a concrete decision where to locate your startup, that matters. And most of the people who argue "Don't move to SF" aren't actually arguing "Move somewhere else", it's usually "Adopt a distributed team where you pick the best engineers regardless of location." The latter approach has a bunch of its own problems, so it becomes a tradeoff of getting the biggest possible talent pool possible vs. reducing communication costs and building a team locally.
It works if you substitute "willing and able" for "willing", which is the usual context of this discussion. If you are willing and able to date a supermodel and you don't, well, why the hell not?
(FWIW, I personally have little interest in supermodels...but then, most people would consider that lack of interest = "not willing", so the logic still holds. Like all logic, it obscures subtleties - I'd be quite willing if she were a good match beyond being a supermodel and I weren't already attached - but then, this is a problem generally with imposing words on the world, and not with the specific instance.)
You can't honestly believe that. Consider the # of professional software developers in the US alone is much higher than the total population of San Francisco.
If the opposite was true, and more were not willing to go to San Francisco, don't you think most of the larger companies there would have moved elsewhere?
Keep in mind, I said "willing to go," not that they were already there.
The numbers indicate that many more are willing than those who are not.