We are witnessing a major achievement of machine learning - appearance of several chatbots which are able to produce human-quality conversation. Appearance of such ‘fluid conversationalists’ has aroused big interest especially among teachers, since they have already passed several exams designed for humans and have been used by students in winter fall 2022 exams. Their performance is not based on linguistic research, but is an achievement of data science. This performance cannot yet be rationally explained (it is based on interplay of billions of variables), thus in the following are used conversations with the currently most visible member of this family - the chatGPT to clear some myths around these programs.