This is obviously an absurd overextrapolation, and it's unlikely that a significant number of people would actually change their name to exploit it, but the principle is accurate: If alphabetical is used consistently then someone with the last name Zelenskyy will consistently end up last in every list of coauthors, while Adams will consistently come near the top. Even if people intuitively understand that alphabetical ordering is used because all coauthors are equal, the citations will still be for Adams et al., and it's not hard to see how that would give an unfair non-merit-based leg up to Adams over Zelenskyy.

If applied consistently, random order would be a fairly sound way to ensure that over a whole career no one gets too large a leg up from their surname alone.