Friendly and Unfriendly AGI are Indistinguishable 2022-12-29T22:13:00.434Z

Comment by ErgoEcho on Friendly and Unfriendly AGI are Indistinguishable · 2022-12-30T07:36:32.631Z · LW · GW

Exactly! This leads to sociopolitical chaos.

Comment by ErgoEcho on Where is human level on text prediction? (GPTs task) · 2022-03-14T10:10:22.884Z · LW · GW

From that paper:
> A new improved method for evaluation of both lower and upper bounds of the entropy of printed texts is developed.

"Printed texts" probably falls a standard deviation or three above the median human's performance. It's subject to some fairly severe sampling bias.