This is becoming common belief among people in the AI field. I'm convinced it is a case of a metaphor overtaking reality. With the LLM model in mind, we can see how Humans are "kind of like" LLMs, therefore we start to believe humans are just LLMs.

Of course this forgets the enormous ways that humans are nothing like LLMs. For example: Humans don't have to read the entire internet to learn how to write complete sentences.