Doug Shannon & I recently discussed using multiple agents in a transactional model to create complex #AI systems. Let’s consider whether AI *should* think like humans, or whether they’ll be at their best when they use entirely new forms of cognition.