Flexibility and the Singularity — LessWrong