Explore & listen
Learning
Interactive surveys of the literature, and generated deep-dive podcasts, ways into the topics I work on.
Literature explorers
Navigable, themed maps of recent research, built with SciX MCP and code-intel sources.
-
Agentic Memory Systems
A curated survey of how agents manage long-horizon memory across architectures, retrieval, and consolidation.
-
Enterprise Multi-Agent Reliability
Practices for running multi-agent systems dependably in production environments.
-
Agent Memory — Design Considerations
Engineering perspectives on memory architecture from industry implementations and community knowledge.
Podcasts
The Agentic Memory Reading Path
- 1. Foundations
The field had a vocabulary problem: everyone was building agents, nobody agreed on what the parts were called. A shared framework for what agent memory actually is.
- 2. The Memory Stack
Why a plain baseline can beat a fancy memory system on its own benchmark, and what a real memory stack actually needs.
- 3. Procedural Memory & Skills
Agents that write, store, and reuse their own skills, from Voyager's self-taught Minecraft tech tree onward.
- 4. The Measurement Crisis
How a single choice in the scoring script can flip which memory system wins, and why agent-memory evaluation is in crisis.
- 5. Forgetting & the Frontier
Why forgetting matters as much as remembering, and where agent memory goes next.