Key flaw with this is the presumption that the required mutations are not already present in the population. Even if they are rare, it is a facile process for natural selection to make said set of mutations dominant. If a small amount of recombination (I don't know if this is occurring with COVID-19) or novel mutation is required, again, it's a fairly straight forward affair unless we're close to an eradication scenario. "Life finds a way", as they say.

I think the only reasonable response to the parent query is to address the core issue ... are there non-overlapping capabilities between "artificial" and "natural" antibodies apart from the speed and health-risk of their development/inoculation. In other words, is "artificial" immunity a better form of immunity in a way that "natural" immunity can never match? If so, seems like a pretty obvious case of "lock in".

Without being an expert, I imagine it isn't a trivial question to answer, as the evolutionary capabilities of a pathogen are a key factor, and well, we're still wrestling with predicting protein folding.