Thoughts on the January CFAR workshop — LessWrong