The academics say the ability to reform jihadis is complicated. And the programs don't really seem to work. The Saudis, for example, seem to define success as graduates not killing inside Saudi Arabia. Exporting their jihadis is not a success for the world, as September 11, 2001 already showed us.
My view is that academics can overly and needlessly complicate problems. I think that killing jihadis (or imprisoning them until they die, if you want to pay for that) is the only sure way to stop them from killing again.
This bleeds into the online radicalization and recruitment issue--which that initial MIT Technology Review article spends a lot of time on--where I think LeadOps should take priority over WebOps:
I'd focus on killing jihadis and smashing up jihadi organizations and sanctuaries, but that's just me. Did we really have a RadioOps information war in World War II to persuade potential Nazis to turn away from Hitler? Or did we pound the Nazis into the ground and delegitimize their ideology with our victory and their defeat? WebOps, indeed. Sounds like it is more funding-magnet jargon than a real part of the war. Be the strong horse and they'll slink away into the shadows.
One problem the MIT article notes is that a lot of emphasis is on online efforts to fight radicalism because it is easier to quantify. Not that the online method is better--but that it is easier to measure. That's a problem for the non-academics among us.
I think the basic problem is seemingly treating jihadis like victims of their surroundings rather than as mad dogs who will kill us all if left to run wild.
"Curing" jihadis is a sucker's game. Killing jihadis is a better way to prevent jihadi recruitment rather than giving them the idea that if caught in the jihad they might spend time in a facility to persuade them of the evil of their ways.