> If you’re only testing UBI on a small population, then 1) it’s not even universal, and 2) there will not be enough “beta testers” to (in)validate these criticisms.

Your point holds until history has proven who is right. Of course, any experiment on UBI that is constrained by a financial reality of whoever conducts the study needs to limit eligibility and therefore lose the proof of universality.

But: Do you have a better design for UBI experiments, given actual constraints?