if there are 894 candidates for a position, there's going to be somewhat arbitrary weeding out somewhere down the line
They could have used any number of other criteria for "arbitrary" weeding. Academic or research record, number and impact of publications, years of experience...
Regardless of how arbitrary you claim the filtering is, the effect is the same - it makes commitment to diversity the most important, primary criterion.
Sure enough, a report on Berkeley's diversity initiative—recently publicized by Jerry Coyne and John Cochrane—shows that eight different departments affiliated with the life sciences used a diversity rubric to weed out applicants for positions. This was the first step: In one example, of a pool of 894 candidates was narrowed down to 214 based solely on how convincing their plans to spread diversity were. - https://reason.com/2020/02/03/university-of-california-diver...
I'll let you decide if using this criterion as the first step, that disqualifies 76% of candidates, is a "token gesture that takes 0.05% of your time".