While I agree that these large language models are not 'artificially intelligent' - I call them machine learning models, no more and no less - the question does

While I agree that these large language models are not 'artificially intelligent' - I call them machine learning models, no more and no less - the question does rise what intelligence actually is. In what way does our mental meandering differ from that of a deep model, and what makes our mental processes be described as 'thoughts' while those of the server-under-the-stairs are mere 'calculations' producing 'shaped noise'?