I find it frustrating how eager AI supporters are to dismiss legitimate concerns using the reasoning "Oh, well, it's always worked out before, I'm sure it'll work out again". What if it doesn't? And if it doesn't, it'll be too late to make changes and prevent suffering.