Conformity is useful in terms of (illusionary) efficiency: they look the same, they think the same, so you can run the same algo on them and they will produce a

Conformity is useful in terms of (illusionary) efficiency: they look the same, they think the same, so you can run the same algo on them and they will produce a curve-bell result which you can use to single out and either eliminate or boost outliers. Problem is, diversity runs deeper than appearances; but few people and even fewer _orgs_ grok that.