Theoretical Neuroscience For Alignment Theory — LessWrong