Advances in nuclear bombs design

This just came back onto my mind after reading about people freaking out over Iranian nukes, an article by Nate Silver earlier in the year, quoting a "renowned scholar", albeit not one that I'd ever heard of:

The renowned Harvard scholar Graham Allison has posited that there is greater than a 50% likelihood of a nuclear terrorist attack in the next decade, which he says could kill upward of 500,000 people. If we accept Mr. Allison's estimates—a 5% chance per year of a 500,000-fatality event in a Western country (25,000 causalities per year)—the risk from such incidents is some 150 times greater than that from conventional terrorist attacks. Other scholars consider the chance of a nuclear incident to be much lower. Even if Mr. Allison has overestimated the risk by fivefold, and the number of causalities by threefold, it would still represent 10 times the threat that conventional terrorism does.

Leaving aside the risk argument here, is there some reason why we are assuming that the terrorists have a bomb that would cause more than twice as many casualties as the combined Hiroshima and Nagasaki attacks (even on Nate's 3x overestimate, we still have the terrorists in possession of Big Boy)? Has technology (specifically, the kind of technology available to the common man) progressed so much? Anyone know anything? I thought that the Pakistani nuclear program had only got to a level of sophistication where its bombs were in the same 20 kiloton yield as the two dropped on Japan - obviously Iran doesn't have any nukes at all. Since one would presume that terrorists weren't going to be able to carry out much in the way of a testing program (something that's been absolutely essential for every current nuclear power), I don't see why we'd be assuming they would do much better in terms of yield, particularly as they would presumably be attacking somewhere that was less vulnerable to firestorms than a wood-built 1940s Japanese city.