Deconstructing the riddle of experience vs. memory — LessWrong