Abstract
In this talk, I will question if there can be possible impossibilities of large language models (i.e., the fundamental limits of transformers, if any) and the impossible possibilities of language models (i.e., seemingly impossible alternative paths beyond scale, if at all).