Thinking about the language models and what we can do with them

Abstract

Abstract Massive developments in technology, ICT, and artificial intelligence have been witnessed in recent years, with various projects emerging showing the apparent superiority of artificial intelligence over human intelligence, such as Deep Blue, AlphaGo, OpenAI Five, and GPT 3. In connection with them and with the favoured method of machine learning with deep learning, the term language model has appeared, which has the ambition to become the basis of general artificial intelligence (AGI). Critical responses, however, have been claiming that this is a dead end. An alternative view to these critical responses is attempted to be shown in this article, with consideration given as to whether the language model is just what critics consider it to be or whether something more can be sought behind it.