[bionhoward]

Emotions project reward and loss over time. That’s straight up reinforcement learning…all the major LMs are built with it