Emphasis in the above is mine. Only problem with the latter bit is that it's precisely the opposite of what we ought to be doing to help stabilize and legitimize the Karzai government. (A case can be made, too, that rapid expansion of the ANSF is similarly counterproductive to this aim.)
The White House is developing “clear targets” for both the Afghan and Pakistani
governments, possibly with specific timelines, as a way to signal that the American military presence will not last indefinitely, American officials said. It is not yet clear what the administration is willing to do if the targets are not met.
Among other things, the officials said, the administration will insist that Afghanistan fight corruption, speed up troop training and retention, and funnel development assistance to areas the Taliban dominate.
So why is it wrong? Well, it's pretty simple: pouring money and development assistance into areas dominated by the Taliban means that 1) everything we do will be much more expensive, 2) the prospects of failure are much higher, imperiling the government's overall legitimacy and control over areas previously deemed "quiet" and "safe," and 3) the enemy will gain from our efforts to the extent that any of them are successful in delivering benefits to insurgent-controlled (or insurgent-influenced) areas.
Control comes before collaboration. The support of individuals and groups is contested by the insurgent and the counterinsurgent through the provision of services and the suggestion of legitimacy, sure, but that only happens after one party is able to largely prevent the other from contesting territory and/or a subject population through force and security. The Taliban doesn't run sharia courts for the local nationals who work at Bagram; why? Because it's senseless to spend resources pitching a guy who cannot plausibly shift his support to the group that's unable to access or protect him. Pouring money and bridges and wells and so on into places that coalition or government forces cannot consistently and safely access decouples those resources from the counterinsurgent's most important tool: presence.
Kilcullen's been on this point lately, too: why spend all our resources in "red areas" when we've got a lot of light green areas we could be shoring up with those efforts? Why contest the hardest spots first? (And further, why work on connecting more of the population to a government that as yet doesn't seem to be competent enough to reap any benefit in legitimacy or support from being more closely connected to more of its citizens?)
There's a whole lot more to say about this -- it speaks to the "ink spots strategy" issues that Bernard and Christian have recently highlighted, and to which I've yet to respond -- and I hope to cover a lot more ground in a comprehensive Afghanistan "path-forward" post in the coming days.