Addressing the energy cost of "large" but rare events is a remarkably low leverage way to improve environmental outcomes. Alongside training AI models, Spacefl

Addressing the energy cost of "large" but rare events is a remarkably low leverage way to improve environmental outcomes. Alongside training AI models, Spaceflight is the other issue that comes to mind. These things look bad because of "big" numbers or plumes of soot coming out of engines, but when added up they account for such a small fraction of emissions that they won't be useful focuses for a long time. Focusing on them now detracts from places where we can have real impact: human transport, cargo transport, food production, clothing manufacture, carbon capture, and so on.