Foreign Policy ran this article some time ago, and I only noticed it now. In it, the author is arguing that since small terrorist groups still tend to turn towards the Big Terrorist Group which is al-Qaida, network theory should tell us that killing the Big Terrorist Group would make the small terrorist groups more dangerous. If only the small terrorist groups would realise this and make themselves more dangerous by not turning to the Big Terrorist Group for a change! Lucky for us they don't! So let them aggregate! (Essentially as opposed to the disaggregation that David Kilcullen is arguing for as the right global strategy - hence the title of my post).
Things like this are argued there, at least implicitly. But do read the article to see if I am right with this interpretation, anyway.
I have just a final remark in criticism. If the only reason Afghanistan's fate is acknowledged to matter from a U.S. perspective is that al-Qaida could return there one day, then why exactly would a U.S. decision-maker opt for following las Casas' guidance and let it be, provided there is a chance to finish it off?
(P.S. Some probably do not realise the effect carelessly recycled second-hand ideas can have on the ideological aspect of taking on terrorism and on global public support for it. Consider what thoughts people get reading dailies in their own languages with articles titled "U.S. foreign policy expert says the U.S. should keep alive al-Qaida.")