I tried this and after being unhappy about how much longer the agent seemed to take, I wonder if there’s some tradeoff between a model’s performance (and task cost) and the amount of principles / rules you’ve invoked so far in the session

Just fyi the rquest availability button is a smaller grey button that shows its misfit with the background (on iPhone 16) You might want to have a look

i think people say that thinking that only training to produce the next likely word would end up producing some local minimum word that generally fits but doesn't actually lead to intelligent thought.

that feels like a misunderstanding of how the loss function behaves when used within a sequence