lifelonglearner comments on Training our humans on the wrong dataset