I see your argument as one against frequent redesigns, not one against redesigns in general. If you change up your UI every 12 months, even if each iteration is technically better, you are causing problems for your users. Wikipedia's last redesign happened 12 years ago.
If we take as given, for the sake of argument, that the new thing is objectively better than the old thing, never changing means never getting those benefits. Stagnate permanently, rather than pay the temporary price of adjusting to change?
Also, userbases are not static. New users would don't need to pay the "cost" of adjusting, because they have nothing to adjust from; and if the new thing is objectively better than the old design, then it will attract more new users than the inferior old thing.
(Note I am not trying to make the case that the new wikipedia design is objectively better or worse than the previous one; that's an argument nobody will win.)