[hga]

Yeah, that solves it neatly.

Reduces to a protoactinium problem, it decays to U-233 with a half-life of 27 days. The article claims there's so little protoactinium in the total mass of salt and stuff that it's not practical to isolate it, at least not without detection, and failing that, not quickly. This gets into fine details beyond my level of expertise, but I agree the problem is much reduced. Although very possibly still greater than for current LEU designs.