Large Language Models (LLMs) have shown impressive capabilities in tasks requiring instant prediction or in-context learning, but they lack the ability to continually learn and effectively transfer temporal insights. Researchers have identified a crucial limitation: the inability of LLMs to self-modify and consolidate memories, akin to the process of sleep in humans. This deficiency hinders their capacity to learn from experiences and adapt to new information over time. To address this, a new approach has been proposed, enabling LLMs to learn from their own experiences and modify their internal state accordingly1. This breakthrough has significant implications for the development of more advanced and autonomous AI systems. The ability of LLMs to self-modify and consolidate memories is essential for their continued improvement and adaptability, making this research a crucial step forward in the field of machine learning. So what matters to practitioners is that this innovation could lead to more efficient and effective LLMs, capable of learning and improving without extensive human intervention.