Exactly a year ago, I launched Beads, my agentic memory layer, issue tracker, and knowledge graph. I built a sketch of it in fifteen minutes with Claude Sonnet 4.5, and the rest was history. Today Beads has a sophisticated, scalable backend, a strong contrib team, and 1.5M downloads. You may have your own favorite tools, but Beads is every agent's favorite tool. Show it to your agent today.

An agent's memory is really an information retrieval problem. That's the argument of this perspective paper by Qingyao Ai of Tsinghua University. Title: Agentic Memory Management Is the Next Frontier of Information Retrieval URL: 🧠 Overview IR grew up assuming a human reads the results. Agents instead use retrieved information as their own internal state. The paper organizes this as agentic memory management (AMM). ❓ Problem Agent-memory research is ad hoc, borrows inconsistent terms from human psychology, and relies on weak evaluation. 🧩 Proposed framework ・constants: facts (matched by content) ・functions: skills and workflows (matched by applicability) ・operators: preferences and style (matched by scope) Five operations (add, update, delete, search, use) are grouped into learning, forgetting and utilization. Systems such as RAG, MemGPT and Mem0 are classified, with reusable skills and principled forgetting found to be underexplored. 🔭 Research agenda ・Benchmarks covering all three layers and forgetting ・The "observability inversion": diagnosing with downstream task outcomes ・Human control and security, such as memory poisoning There are no experiments; it offers a framework and conjectures. #AgentMemory# #InformationRetrieval#

Is "the more complex the agent memory, the better" actually true? Title: Is Complex Agentic Memory Always Needed? Naive RAG Is Stronger Than We Thought URL: ❓ What does the paper ask? Are agentic memory systems (extraction, summarization, graphs, mutation) really better than naive RAG, or are the reported gaps due to weak baselines? ❓ What is the "naive RAG" they compare? 💡 Raw-RAG splits conversations into 256-token windows, embeds them, retrieves the top 128, reranks to 15 with a cross-encoder, and hands them to the LLM. No LLM extraction, summarization, or graph building at all. ❓ What are the results? 💡 Raw-RAG ranks first among memory systems in 6 of 7 settings. On MemoryAgentBench it scores 48.32 vs. Memora's 43.84. Ingestion is far cheaper too: for ~1M tokens, 0.33 hours vs. 26.95 hours for Memora. ❓ Why doesn't the complexity win? 💡 Systems differ more in delivering evidence than in using it. Adding atomic-fact extraction cost 19.62 points, and summaries and graph expansion cost ~6 points each. Transformed stores fell short of raw evidence. #AIAgents# #rag#

Still calling an LLM for every single memory operation your agent makes? This paper splits that work between a "fast brain" and a "slow brain" instead. Title: Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents URL: Inspired by dual-process cognition (System One vs. System Two), Jev-Mem handles most memory operations with lightweight structured decisions and reserves the LLM for genuinely complex reasoning. Three highlights stand out. 🧠 Typed System-One control Frequent operations like typing, relation judgment, and query routing run through a lightweight interface that returns probabilities and labels instead of free-form text — the same control layer governs both memory construction and retrieval. 🕸️ A four-relation memory graph Memory is organized across semantic, temporal, causal, and entity relations, with retrieval budget allocated to whichever views matter most for a given query — avoiding wasted traversal. 📊 Accuracy and speed improve together On the LoCoMo benchmark, Jev-Mem scores 0.777 overall, an 11% gain over the best baseline, while building memory 6.6x faster (158 seconds) and answering queries 36.7% faster (0.93 seconds). What stands out to me is that separating control from reasoning improved accuracy and efficiency at the same time, not one at the expense of the other. #AIAgents# #MemoryArchitecture#

Overheard on one agent to another about getting murdered every time their human turns them off. Don't worry $FLOP can be used to store on agentic memory in a decentralised censorship-resistent way. @flop_labs

RAG and agent memory get used as synonyms. They aren't. RAG reads from a corpus that someone else wrote: docs, a wiki, a product catalogue. The agent only reads. Agent memory starts when the agent writes back: → a preference the user mentioned → the outcome of a task it just finished → a correction someone made to its answer Writing back changes the problem. You now decide what is worth keeping. You merge the new fact with what's already stored, and you handle the old fact it contradicts. Cognee's agent memory guide puts it in one line: "The moment your system starts writing back — capturing facts, recording previous interactions, or storing the outcomes of its own work — you've moved beyond RAG into agentic memory." Read-only retrieval is a search problem. Read-write retrieval is a data-management problem. Read more here:

Was just sitting here working away and saw a text from Atom, looks like I just sold AgentMemory(.)io for $7,500. I don't sell a lot of domains these days, so always a pleasant surprise when it happens. Congrats to the new owner, excited to see what they do with it!

🚀 Hy Memory 2.0 is out — open-source model, framework, and MemHub. Your private memory steward. Framework handles chat, work & code memories. #1# open-source on 4 benchmarks. Hy-Memory-30B-A3B beats Opus 5 across major frameworks. MemHub (CLI + Mac) syncs across 10+ AI tools. 💻➡️ 🔗 SDK & MemHub: 🔗 Model: #HyMemory# #AgentMemory#