x
Perceptual dexterity: a case study — LessWrong