The winner of the prestigious Turing Prize has openly expressed his skepticism about this before. He has previously written that before we can worry about controlling artificial superintelligence, "we need to have at least a hint of the design of a system smarter than a domestic cat."
According to the publication's observations, Jan Lekun is fond of "cat" analogies. According to him, cats have a certain ability to reason and plan. But these qualities are not inherent in modern large language models (LLMs).
In his opinion, LLMs only demonstrate that "you can manipulate language and not be smart," and they will never lead to true general artificial intelligence (AGI).
By the way, back in the summer of 2023, Jan Lekun stated that large language models of artificial intelligence are not truly intelligent because they cannot understand, interact, or comprehend reality and rely on language training to achieve results. Then he said that large language models are not even as smart as dogs.