simon comments on Are top existential risk estimates 50,000 times too high? An optimizer’s curse model and analysis

[simon]

To be clear, it need not be deliberate and they need not benefit personally!