I always enjoy reading some of your comments, they ameliorate the hype about LLM and give a critical review. Anyway, I think a stronger model than GPT-4 could improve the way to use tools, so that the model is able to self-improve using tools. For example using all kind of solvers and heuristics to guide the model. I don't know how to estimate that risk just now.
Edited: Don't know if is a good thing to study the weak points of closed LLMs. Even asking LLMs can give hints about possible ways to improve. In my case I am happy I am certainly old and my mind is a lot weaker than before, but even in this case I prefer not to use LLMs for gaining insight because she will someday get a better insight than myself. But the lust of knowledge is a mortal sin.
Edited: Don't know if is a good thing to study the weak points of closed LLMs. Even asking LLMs can give hints about possible ways to improve. In my case I am happy I am certainly old and my mind is a lot weaker than before, but even in this case I prefer not to use LLMs for gaining insight because she will someday get a better insight than myself. But the lust of knowledge is a mortal sin.