Such a program is way too complex to model out and reliably predict what the impact will be. We simply don't know how much UBI is the "right" amount for specific outcomes, and we definitely don't know what the impact will be on any subset of the economy.
According to the OP author, we can't even test UBI programs before a full rollout. If we can't reliably model or predict the outcome, and we can test it, how are we supposed to actually implement it?