Comment by heisgone

I'm somewhat skeptical of that approach. It's not even clear if "world models" are significant for our subcouscious decision making as human beings. Why assume it's necessary for AI?