Comment by taneq

It might not even be the leadership at this stage. It’s entirely possible that “rounds of conversation” is a metric that their reinforcement learning has been told to optimise.