A bot that detects logical fallacies would be amazing for journaling. Would be really interesting to stream of consciousness into a text document and have the computer put red underlines under questionable premises or conclusions.