It understands everything, yet it knows nothing at all.

I hope I haven't misunderstood. Back in March or April, I tried running local AI on my crappy computer with an RTX 3060 12GB, and it gave me a profound realization of the difference between an LLM and AI. To put it simply, an LLM is a language prediction machine; it isn't truly capable of foresight. Instead, it guesses our intent based on the words we input and retrieves useful information! Calling it AI might be a bit of a stretch. But if you add automated engineering processes or logic to it, perhaps it could be called AI!

My current assessment of this kind of LLM is: it knows everything, yet it knows nothing at all.