It did really well in comparison to ChatGPT4 in my opinion, at my standard way of testing AI’s which is to quiz it about characters and plot points about the bo

It did really well in comparison to ChatGPT4 in my opinion, at my standard way of testing AI’s which is to quiz it about characters and plot points about the book Cryptonomicon, ha. But it still hallucinates up junk when you reach the edges of what it could learn from internet hosted plot summaries. I want an AI that has read the book, not the Cliff’s Notes!