"From TFA, the changing cross srction reduces resistance as it stretches. ... they could be designed for no change when stretched"

Well, that's not quite what TFA writes: "As expected, electrical measurements show that the fibers increase resistance as the fiber elongates and the cross sectional area narrows. Fibers with large diameters (~600 [micrometers]) change from a triangular to a more circular cross-section during stretching, which has the appeal of lowering the resistance below that predicted by theory."

The abstract doesn't mention how the circular/triangular transition would affect the resistance - with conservation of volume it shouldn't matter. But I don't read here in any way that this effect would be able to cancel the resistance increase due to stretching.

Note that in first approximation, resistance would scale as L^2 for a wire with length L (both diameter decrease and length increase affect the resistance). With stretching up to a factor 10, i.e. 100x increase in resistance, a small effect due to the shape of the cross section would be negligible.