Dacyn comments on Imitative Reinforcement Learning as an AGI Approach