Indeed; it's definitely on topic and interesting work, and I expect that simple people-reading models like this will do tremendous amounts of good and make significant progress in parallel to the first-principles work that SI appears to be doing.