The post discusses Large Language Models (LLMs) and the complexity of their explanations. While many users appreciate the simplifications, they also express frustration over the lack of concrete details on how to actually build LLMs. Key points raised include the prediction capabilities of LLMs, parallels drawn with human cognition, and a demand for a more practical guide that combines theory with actionable implementation steps. Several comments highlight a perceived vagueness in existing explanations, calling for more specific information on LLM architecture and training processes, as well as practical guidance for building such models.