Jev-Mem: System-One/Two-Inspired Memory Architecture Builds Agent Memory 6.6x Faster
omarsar0 · x · 2026-09-26
A new report proposes Jev-Mem, an agentic memory architecture inspired by System-One/System-Two cognition. A lightweight controller makes memory decisions—assigning types and relations on write, and handling routing, retrieval budget, graph traversal, scoring, and stopping on read—so the LLM is only called for final reasoning.
Results: memory construction is 6.6x faster, query latency drops 36.7% (to 0.93s on average), and it scores 0.777 on LoCoMo with an LLM judge, an 11.0% relative improvement over the strongest baseline. Construction takes 158 seconds. A solid weekend read for anyone building agent memory layers.
More from coding & agent
- Devs clash over retiring Three.js for game dev in the AI vibe-coding era — AIandDesign · 2026-09-26
- Muse AI agent gets its own phone number via Bland to call, book and text — ramagetime · 2026-09-26
- Abstractions were priced for human programmers — a thousand agents break that logic — ccerrato147 · 2026-09-26
- Stop putting chatbots in your app: give agents a CLI instead — ccerrato147 · 2026-09-26
- Shopify rebuilt the Shop app natively in Swift and Kotlin with 6 engineers in 12 weeks — ccerrato147 · 2026-09-26
- After 50+ product videos, a team breaks down Apple's framework for premium-looking video — GCWebDesigner · 2026-09-26