Comment by cedws

Does agent memory actually work as a concept yet? I've only seen LLMs commit garbage to memory and recall in irrelevant contexts.