PROWL-2 introduced as a recursive-learning approach for agents and their world model

Odyssey introduced PROWL-2 on October 1, describing a process in which agents and their world model improve through recursive learning. It says agents expose errors in imagination that can be repaired to enable further learning. Odyssey believes this open-ended learning is a critical step toward superintelligence.