[yepyoukno]

There is a mind trap which makes perspectives constrain themselves to the “one ultimate” for instance or solution.

Despite what is mainstream or widely favorable or regulated, all manner of divergent side tracks will also be happening, particularly anywhere there are incentivized interests.

All that said, I do not think LLMs will ever achieve “AGI”, however useful in animating compendiums of human understandings.

LLMs have trouble disagreeing with established human knowledge, bringing back something Bertrand Russell was famous for saying, regarding much of what we believe we know is merely what we agree to among ourselves. This is as true today. LLMs and humans alike favor agreeability over technical correctness.

True AGI will be more evident when there is more deductive reasoning proficiency. Finding and disputing examples of established knowledge may be such a sign.