I'm not going to worry about it until some data exists. It all seems so speculative right now. Even in that preprint, a small minority of antibodies seem to enhance the binding. Maybe that explains why some small subset of people are vulnerable to reinfection while most people appear to have robust immunity. Anyway it won't matter once a sufficiently large number of people are vaccinated or recovered, because those vulnerable people shouldn't encounter the virus anymore.
Personally, I think that the vaccine trials missed a really good opportunity in their trials. I think they should have been 3 arm trials where you have placebo, 2 doses, and 1 dose followed by a placebo. Would have been really instructive to learn how much immunity a single dose gives. If it's, say, 60%, then it's probably far better mathematically to give twice as many people a single dose than half the people the 95% immunity that a double dose confers. After all, we don't have to worry about mutations if the virus is defeated before a potentially catastrophic one arises (if it hasn't already). Not sure how or why that decision was made, but I guess when thigs are done hastily, it's easy to be a Monday morning quarterback.
As an aside, antigenic drift probably wouldn't destroy the effectiveness of the vaccine, but rather reduce it marginally, because the domain that the vaccine are produced against is too important to the virus to change that radically. The Ab's that are produced by the vaccine will be effective to a certain extent. (And anyway there's even evidence that a recent dose of MMR vaccine reduces severe disease just by ramping up innate immunity.) So even if we can't achieve 95% immunity, it's still worth taking as soon as possible. This isn't a doom scenario, as far as I can tell.