It is very easy to predict the past because the robots have the current state of the environment to verify past predictions. However, the same advantage that allow robots to predict the past is a disadvantage when predicting the future. Future predictions have no benchmarks to determine if a future prediction is right or wrong. All atoms, all objects and all events are interconnected in a massive dependent network, whereby all atoms on earth affect each other one way or another. This is true for the past, present and future. The AI system has to track and understand all events, objects, places and things on the planet first. It uses all the current and past information to predict the future. The best way to do this is through the AI gnosis where “all atoms on earth” are tracked in a timeline. All current atoms and their sub-atomic particles on earth are used as forensic evidence to predict the future. All events on earth must be predicted as a whole, no prediction is done in isolation. Most importantly, the AI system has to predict every single human being on the planet and what they are going to do in the future. The examples in the issue covers how to predict the Powerball numbers 2 weeks into the future and prediction of Super Bowl 2100, frame-by-frame and atom-by atom.