The Limits of Large Language Models like Chat GPT

During different periods of history, there have been various explanations for why an apple falls to the earth./nLanguage models like Chachi BT, trained on historical data, may not predict improbable future hypotheses./nArtificial intelligence models can suffer from under generation or over generation of responses./nThese models lack a real understanding of concepts and physical reality./nLarge language models are limited in their ability to distinguish possible from impossible./nMoral reasoning and AGI require the ability to distinguish between what ought to be and what ought not to be./nLarge language models are currently akin to glorified auto complete rather than true intelligence.