x
Adaptive Coherence — LessWrong