What Annoys Subbarao the Most?
From the viewpoint of his decades of experience as a computer science teacher and researcher, Subbarao Kambhampati, a professor in the School of Computing and Augmented Intelligence, part of the Fulton Schools, sees things he likes and does not like about paths the field is taking, particularly in regard to the rapidly emerging use of artificial intelligence, or AI, technology and Large Language Models, or LLMs. First of all, AI and LLMs are not actually intelligent or knowledgeable in the basic human sense. AI and LLMs are instead merely retrievers of information, although very prolific retrievers, Kambhampati says. LLMs can only guess about what is correct, and cannot verify information they gather, only accumulate it, he adds. One of his current focuses in the field is finding ways to develop and encourage better human-AI interaction that might actually aid human reasoning.
See Also, LLMs are just like Toothpaste, Analytics India Magazine/AI Origins & Evolution, January 3