In principle we do not know the long-term impact in advance, but we can often make a probabilistic estimate.
I'm not sure this is true. We can't make a probabilistic estimate without a model. And to design an accurate model, you need to describe precisely what it means to ask, "is this activity worth pursuing?" Is there a way to do this without involving all of human society in your model?
I'm not sure this is true. We can't make a probabilistic estimate without a model. And to design an accurate model, you need to describe precisely what it means to ask, "is this activity worth pursuing?" Is there a way to do this without involving all of human society in your model?