I can’t think of a reply to this that won’t start a game of reference class tennis; but I think there’s a possibility that Aaronson’s list is a more complete set of the relevant experts on the climate than your list is of the relevant experts on the existence of deities. If we grant the existence of deities, and merely wish to learn about their behavior; your list would be analogous to Aaronson’s.
I think “etc.” is a request to the reader to be a good classifier—simply truncating the list at “etc.” is overfitting, and defeats the purpose of the “etc.” Contrariwise, construing “etc.” to mean “everything else, everywhere” is trying to make do with fewer parameters than you actually need. The proper use of “etc.” is to use the training examples to construct a good classifier, and flesh out members of the category by lazy evaluation as needed.
I can’t think of a reply to this that won’t start a game of reference class tennis; but I think there’s a possibility that Aaronson’s list is a more complete set of the relevant experts on the climate than your list is of the relevant experts on the existence of deities. If we grant the existence of deities, and merely wish to learn about their behavior; your list would be analogous to Aaronson’s.
Both lists end with “etc.”, so I have trouble calling either of them incomplete.
I think “etc.” is a request to the reader to be a good classifier—simply truncating the list at “etc.” is overfitting, and defeats the purpose of the “etc.” Contrariwise, construing “etc.” to mean “everything else, everywhere” is trying to make do with fewer parameters than you actually need. The proper use of “etc.” is to use the training examples to construct a good classifier, and flesh out members of the category by lazy evaluation as needed.
It’s not a reasonable presumption that “etc.” will cover “any arbitrary thing that happens to make trouble for your counterargument”.