Memory Is All You Need: An Overview of Compute-in-Memory Architectures for LLMs | Hacker News Reader