Whenever I say that one key pillar of a viable strategy to curb global warming ought to be efforts to promote high-density living arrangements, I'm invariably confronted by a kind of circular argument that Ross captures well here, channeling Joel Kotkin but with my emphasis added:
The traditional unipolar urban downtown isn't going to make a comeback: Young couples with families can't afford to live there, and aging Baby Boomers don't want to. The American city of the future will be more of an archipelago of suburbs than the kind of one-downtown organism bred by the Industrial Revolution: "We aren't creating more New Yorks and Chicagos; we're creating more Los Angeleses.
There's the paradox. The urbanist proposal isn't "hey, jerks, why don't you all move to dense downtowns." Rather, the proposal is something like "why don't we impose carbon taxes so that things like driving long distances and heating or cooling large detached structures are priced in accordance with their social cost? Why don't we stop having the federal government heavily subsidize driving cars as the preferred mode of transportation? Why don't we have more areas that allow for high-density zoning, thus reducing the cost of urban housing?" It's not that we urbanists are unaware that many people live in low density areas because its cheaper, it's precisely that we are aware of this fact that makes us believe that the "traditional unipolar downtown" could make a comeback.
Now, will it come all the way back? Of course not. Douglas Rae's City quite brilliantly explains the connection between the high-tide of urbanism and a particular technological moment when the availability of fast, cheap rail and water transit and the total unavailability of cars encouraged very dense settlement patterns. Quite naturally, the combination of cars being invented, cars being massively subsidized, and governments being successfully lobbied by car companies to dismantle mass transit systems led to a massive shift in the direction of sprawl. But by that same token, if we step away from those policies to some extent we'll see a rebalancing in the direction of urbaism.
Neither Ross nor Kotkin cares to deny that the future will entail less driving. Instead, they rely on this aperçu "Telecommuting, not mass transit, is the wave of the future" (combined with the nonsensical observation that "if you take New York out of the equation, there are already more Americans telecommuting today than taking mass transit"; by the same token if you ignore the three percent of Americans who do the most telecommuting very few people telecommute!), but why choose? It seems to me that the thing to say is that the cost of driving should be priced more appropriately and that people will respond to that policy shift in a variety of ways.
This is all, I should say, a bit irrelevant to the issue of whether or not cities should have unipolar downtowns. Contra Kotkin, New York City, for example, is strongly multifocal just like LA. The cities have very different development patterns, but the existence or lack thereof of a unipolar downtown isn't the issue.