Humans are special in that we have achieved more than any other species on this planet in every aspect you can experience. That does not mean that most people are smart or that they are utilizing their potential. LLMs, as they are currently marketed, are not AI as most people would define it. Without aggressive marketing no-one would refer to them as "AI". LLMs are currently still dumber than most animals or insects. On the other hand, most people are dumber than even this current "AI" because they refuse to think for themselves and instead let themselves be manipulated into doing whatever someone else wants as long as that someone is even slightly "persuasive". So, if we let LLMs "rule" our future then some people will be better off for a little while and it will all fall apart when there is no longer anyone capable of maintaining our new LLM based world or when something new occurs that LLMs have not been trained on so they cannot draw on multiple prior examples of correct resolution. And no, LLMs do not "think" or "reason", they produce output that make it seem like they do but deviations from training examples will trip them up.