I wasn't expecting this reaction at all, but having read the article I think Matt Bai really nails it, writing about Howard Dean, his "50 State Strategy," his conflicts with the DSCC and DCCC, and related matters:

Most analysts in both parties now believe that Democrats have better-than-even odds of winning at least the House. But if they don’t, rather than dissect the mechanical failures that cost them a few thousand votes here or there, Democrats might be forced to admit, at long last, that there is a structural flaw in their theory of party-building. Even a near miss, at a time of such overwhelming opportunity, would suggest that a national party may not, in fact, be able to win over the long term by fixating on a select group of industrial states while condemning entire regions of the country to what amounts to one-party rule. Which would mean that Howard Dean is right to replant his party’s flag in the towns and counties along America’s less-traveled highways, even if his plan isn’t perfect, and even if he isn’t the best messenger to carry it out. As another flawed visionary, the filmmaker Woody Allen, once put it, 80 percent of success is just showing up.

Another thing I would add, that Bai doesn't really get into, is that just about every election looks uniquely crucial. Which isn't to deny that 2006 does, in fact, look uniquely crucial. But I recall 2004 and 2002 as having looked the same way. This is just a natural human bias toward over-emphasizing the present, but there it is none the less. Everyone agrees that the long term needs to be addressed sometime or other and, really, there's no time like the present.