Hi Sichkar, thanks for this repo, very helpful to RL beginners like me. In agent-brain's code, i think there should use 'and' of this line: ``` if next_state != 'goal' or next_state != 'obstacle': ``` Otherwise this judgement will always be true.