I would say the hype has started to fall off, as it's becoming increasingly obvious that AGI is not around the corner, but meanwhile practical use cases keep ge

I would say the hype has started to fall off, as it's becoming increasingly obvious that AGI is not around the corner, but meanwhile practical use cases keep getting better and better. The more we understand the strengths and weaknesses, the better we can exploit them, and even if the models themselves have hit the scaling wall, I think tooling around them is far from done.