It doesn't invalidate the original claim, as they haven't been able to make a pure sample to test with, as I understand it. Or at least that there is some difference between their samples and the others, but still it shows promise.
Here's a lump of 100% Cu and another lump of 50% Cu and 50% "other".
If purity doesn't matter, they must have the same properties? Why not just use 0.0001% copper in all our applications and save millions on material costs.
I know nothing about the theory of superconductors, but purity can affect electronic properties. For example, pure silicon is quite a poor conductor. If you dope it with a little bit of an n- or p-type dopant (an impurity), it conducts better. Add a bit more, and it conducts more. There are other effects, too.
Purity would affect superconduction. Consider grains of superconducting material embedded in a matrix of regular conductive metal.
One of the theories for the behavior of this material explains that the copper atoms preferentially form a structure that does not superconduct. The structure that does superconduct is tricky to achieve due to the energy levels. (Paraphrased summary)
It's likely that a poorly prepared sample will have discontinuous regions of both superconducting and non-superconducting material. If that is the case, you won't observe superconduction.
It may be that the non-superconducting material does superconduct at low temperatures, which would mask the purity problem.
The truth is that quite literally nobody fully understands what's actually happening here. That's kind of the point of all this experimentation
