I'll agree that it's inevitable, but how much time we have is the question. The replacement of human labor is likely to be incremental, because AIG is such a tough nut to crack. So certain areas will be replaced before others, so we'll have a few dry runs before the whole thing gets finalized. I'm hopeful we can figure it out, but big changes come with a lot of unforeseen, like imagine trying to predict internet culture wars based just on DARPA researchers making the first versions of the internet. "Wait so everyone has access to every piece of human knowledge, and that somehow created more flat earthers?".

I think that it will take generations to get to full automation, because there are challenges unique to most specializations that are hard to master completely. That is a lot of time for a society about to lose necessary employment to figure out how and if to replace it. UBI might be the way, but given how incompletely we see the future, it could be that there are circumstances that make it not the best solution. I think we should keep an open mind, and not despair because our favored solution looks unlikely.

*edit* An example is UBI is predicated on capitalism, and we may find capitalism doesn't work with hyper-unemployment, which would be a very likely step along the path to full automation. There are a lot of corners we can't see around, and a lot of corners we have yet to see at all.