Yeah, these language models might be one building block, but their output will for example need to be parsed and related to world models and specific knowledge graphs. Also, people have an individual memory and many other elements to them.