Back almost 10 years ago, Facebook ran negative tests about what showed up on your news feed. To the point they had a statistical model of how many percentage p

Back almost 10 years ago, Facebook ran negative tests about what showed up on your news feed. To the point they had a statistical model of how many percentage points of increased rate of suicide it caused. And I'm not talking euphemistically; Facebook has a field for "is this user probably dead" and "cause of death" and the reports on experiments about mood-filtering on news feed items included a line item on "rate of suicide" and nobody on that team cared it was whole -digit percentages up. Facebook has long employed unit 731 style experiments. This is no surprise.