The person I was responding to said "bombs", so I was talking about that. A fission bomb's chain reaction shuts down as soon as the assembly expands enough to become a bit subcritical. The thermonuclear bomb's hot burning fuel, on the other hand, continues to react (albeit at a slower rate) as it expands; there is no such sharp cutoff. The latter fact can be used to improve fission bombs some by "boosting": including a small amount of DT gas in the fission bomb, to provide a source of neutrons that induce some fission even after k drops below 1. If k is (say) 0.9, then each DT neutron will produce 10 fission events, for a large energy return, roughly 100x the energy from the fusion itself.
I'll add another observation: the secondary in a thermonuclear bomb is a lot like the target in inertial confinement fusion, only much larger. The expansion time of these systems is proportional to their radius, so if (say) the bomb has a million times the fuel, it takes 100 times longer to expand. So for a given density and temperature of fuel, there is 100x as much time for the fusion reactions to occur. In a real sense, thermonuclear bombs become easier to build as they get bigger.