They do not in fact do it just fine. That's why it's utterly laughable to suggest that they can lead to AGI. They can't even do the one thing they are supposed to be good at, even after years of effort.
I didn't say anything about AGI, nor will I except to say it's an incoherent quasi-religious topic that has next to no engineering relevance, at the very least until somebody can empirically test for it. Stop coping.