You are claiming they are statistical parrots, which I don’t think the parent poster meant.
The “statistical parrots” argument might have been compelling with GPT-3, but not with today’s models and the results of mechanistic interpretability research, which show internal representations and rudimentary world models.
The issue here is that, even with a lot of VRAM, you may be able to run the model, but with a large context, it will still be too slow. (For example, running LLaMA 70B with a 30k+ context prompt takes minutes to process.)
Do we know how human understanding works? It could be just statistical mapping as you have framed it. You can’t say llms don’t understand when you don’t have a measurable definition for understanding.
Also, humans hallucinate/confabulate all the time. Llms even forget in the same way humans do (strong recall in the start and end of the text but weaker in the middle)
I don't care much for the concept of AGI since it's ill defined.
However, out of distribution generation and generalization are a better more useful metric. In these Yann Lecun has argued that interpolation is meaningless in the high dimensional spaces these "curves" are embedded in.(https://arxiv.org/abs/2110.09485)
ARC has proven they can generalize, and Alpha Go (not an llm but a deep network) has proven it can generate novel/creative solutions. We don't need AI to have a sense of self "I" for them to beat us at every human skill and activity. Infact it might be detrimental and not useful for us if AI developed a sense of self.