[P] Applying the Ebbinghaus forgetting curve to AI agent retrieval -- a biologically-inspired memory system

Most retrieval systems for AI agents treat all indexed content as equally available regardless of age, access frequency, or contextual importance. This doesn't reflect how effective memory systems actually work. I built claude-memory, an open-source Python package that layers a biological memory model on top of hybrid retrieval (vector similarity via ChromaDB + BM25 keyword scoring