Jev-Mem Uses a System-One Control Plane to Cut Agent Memory Latency 36.7%
Dongming Jiang · hf · 2026-09-22
Most agentic memory systems lean on autoregressive LLMs to organize, retrieve, and use memories, putting expensive generation on the critical path. Jev-Mem borrows System-One/System-Two cognition: a dedicated System-One control plane handles memory typing and relational organization at build time, plus query routing, retrieval-budget allocation, graph traversal, candidate scoring, and adaptive stopping at retrieval time; System-Two is invoked only for complex reasoning and answer synthesis.
Results: on LoCoMo it scores 0.777 overall LLM-as-a-Judge (+11.0% relative over the strongest baseline), cuts memory construction to 158 s (6.6× faster than the fastest competing system), and lowers average query latency to 0.93 s (−36.7%).
More from coding & agent
- Halo open-source post-training framework hits 2.8x TRL throughput — prasenx · 2026-09-22
- Jev from a ChatGPT co-creator auto-routes tasks to the cheapest capable model — fsharpman · 2026-09-22
- Prof. Tom Yeh releases printable hand-calc agentic AI cost math problems — ProfTomYeh · 2026-09-22
- apibase unifies 327 tools from 92 providers behind one pay-per-call MCP endpoint — modelcontextprotocol · 2026-09-22
- mcp-server-s3 ships MCP server letting agents browse, upload and share S3 files — modelcontextprotocol · 2026-09-22
- Dev built an AI game-playing plugin but shelved it: vision models too slow and costly — ezshine · 2026-09-22