Yeah, I find the notion that human judgement is superior to algorithmic judgement to be pretty suspect… look at Jens Ludwig’s work on judges, it’s not like involving a human improves outcomes somehow