Sometimes an article comes out that is so wrong headed, even delusional, that someone has to respond. In a recent piece posted on the Planetizen website, software guru Paul Graham addressed the issue of "How to be Silicon Valley," which alleged that to become a high-tech hotbed "you only need two kinds of people'¦ rich people and nerds."

This reflects the conceit of a small, but influential group of elite techies and their supporting claque of pundits. According to these worthies, the "right people" only congregate in certain places with "well preserved old neighborhoods instead of cookie cutter suburbs, and locally-owned shops and restaurants instead of national chains". Unlike the rest of us, the chosen will cluster in places "with personality."

Places, in other words, like Cambridge, Massachusetts, where Graham got his Ph.d, as well as San Francisco, Palo Alto, and other favored burgs. Towns where, he asserts, "people are walking around smiling".

In many ways this is a retread of Richard Florida's "creative class" theory. But at least Richard attempts to apply some social science to his construct, albeit using questionable barometers like numbers of gay people, artists, and overall ethnic diversity to prop up his theory. Graham's insight is instead pure assertion and, from my point of view, it only gets the present, the past, and likely the future wrong.

The Present. In terms of the present, our work with Inc.'s "Best Places" looks not only at job growth, but also at key sectors like information and professional business services. Our numbers, taken from Bureau of Labor statistics data by economist Mike Shires, shows that rather than booming, most of Graham's places with "personality" are stagnating. If you don't like our take, you can see a remarkably similar list at the Milken Institute site.

Certainly, Graham's metro role models have not had a very good decade so far. In fact, Boston, San Francisco and San Jose have been among the worst performers in both overall employment and high-end jobs for the past five years.. These places also, according to the most recent Census numbers, have among the highest rates of domestic out-migration in the nation. The cities of Boston and San Francisco, those places with the most "personality," have been losing population.

Maybe I am missing something here. Is a successful place one that actually loses residents? Perhaps if people in these places seem "happy" -- which I have not noticed particularly on my own visits -- it may be because many are still living off past stock windfalls or trust funds, or perhaps it's just that smug contentment that tends to cling to the "anointed" ones.

Indeed even with the current bubble around firms like Google, the San Jose area's information sector has grown only marginally over the past year or so, and is not close to making up the enormous losses of the post-1999 period. Whose region has been growing? Places like Provo, Utah, Boise, Idaho, and, in professional business services, cities like Orlando, Las Vegas, and Reno. Maybe it's just that they are finally getting a personality. Can the happiness bug be far behind?

There may be another explanation. For one thing, places like the Bay Area and Boston are now too expensive for young people, particularly those with families. A recent McKinsey Bay Area economic forum reported that the median regional professional salary there, $70.000, leaves no surplus. In contrast, the median $62,000 salary in Denver will leave the same worker a nifty $22,000 in discretionary spending.

Like Florida and other creative-class theorists, Graham also doesn't make much of children in his argument. But most engineers I know want to have kids eventually. As my friend Tory Gattis at Houston Strategies suggests, they may put it off for another decade than in the past, allowing the "hip" towns a little more breathing room, but at some point, most people want to start a family.

I agree with Graham that most techies may well prefer places "with personality" (who doesn't?) but they also like good schools, affordable housing, and a lot of other things as well. If they can get both -- which is sometimes possible -- they will take it. If the choice is between a nice, albeit somewhat non-descript suburb good for families or a "funky" mega-priced urban center that is not, most will opt for outer ring Nerdistan. At least that's what they have been doing for the past 50 years.

As for the startups, Graham really talks about one kind of startup -- the ones backed by venture capital, which represent a tiny percentage of startups. Yet, even these kind of companies are produced in more than a couple of regions, as new capital networks develop in new places, from the Sierra Nevadas to St. Louis and south Florida. And given the costs, even when a classic Silicon Valley-type company does get launched, there's a growing tendency to shift much of their employment quickly to other places such as India, China, or even, as Yahoo just did, to pleasant small towns like Wenatchee, Washington.

(Later this week: The Next Silicon Valley? Part 2)