Weird. Does this mean they predicted GPT-4′s performance in advance (and also didn’t let that update them toward doom)!?
Perhaps it’s less surprising given who counted as ‘superforecasters’, cf magic9mushroom’s comment here? I’m not sure how much their personal anecdote as participant generalizes though.
Weird. Does this mean they predicted GPT-4′s performance in advance (and also didn’t let that update them toward doom)!?
Perhaps it’s less surprising given who counted as ‘superforecasters’, cf magic9mushroom’s comment here? I’m not sure how much their personal anecdote as participant generalizes though.