x
subfunctional overlaps in attentional selection history implies momentum for decision-trajectories — LessWrong