Does this mean you expect AI to be able to train on one video of one cat from different angles?