Tuesday 15 September 2026Paper No. 1,386

Evening EditionRSS

ZOTPAPER

News without the noise

Memory & Planning

Memory architectures, long-horizon reasoning

28 articles

All Agent Harnesses Tools & MCP Multi-Agent Systems Creative Agents Memory & Planning Safety & Evals RL for Agents AI & ML

Viewgrid list compact

Agent-side memory steers stateless robots through long manipulation tasks

Kernel-managed memory personalizes multiple AI agents with shorter prompts

Agents can retain old facts without using them indiscriminately

Hierarchical memory trees improve long-context reasoning without model training

Game-theoretic filtering helps agents suppress harmful memories during long tasks

Memory guidelines help language-model agents succeed consistently across repeated runs

Smarter merging and packing improve LLM memory under tight budgets

Structured agent memories withstand model upgrades better than compressed notes

Streaming video memory works better when internalized as evolving latent tokens

A Vestigial Attention Branch Signals Which Cache Entries to Keep

Memory lets traffic-simulation agents build reusable skills without retraining

Rule-based memory improves question answering across long conversation histories

Agents benefit from representing what expands future planning capacity

Coding Agents Need Semantic Measures of Working Memory Performance

LLM agents do not reliably improve from their own experience

Persistent Wikis Help AI Agents Build Better Transferable Skills

Vector memory helps recommendation agents use full interaction histories

Query-aware evidence forests improve long-term memory for multimodal agents

Dual-brain memory keeps real-time speech agents accurate and emotionally aware

Sliding reasoning windows make long test-time scaling 3x faster

Reusable latent trajectories improve world-model control while cutting planning time

Memory consolidation outperforms retention under tight budgets, but retention wins when budgets are loose.

Skill evolution improves agent performance in image generation workflows

Treating memory management as a trainable cognitive skill boosts LLM performance 2x–4x in long-horizon tasks

Three-stage training instills world model planning into LLM agents

Shared agent memories boost performance without coordination

First systematic characterization of agent memory systems across long-horizon tasks

Memory as state management instead of semantic retrieval improves long-horizon agents