How AI Agents Remember Things
How do AI agents remember things between sessions? Every agent forgets everything when a conversation ends, so how do the best ones seem to know you?
I break down the memory architecture behind real AI agents, using OpenClaw (an open-source AI assistant) as a reference implementation. You’ll see how LLM agents write, store, and load persistent memory using plain markdown files, and the four mechanisms that keep context across sessions, including context window management, bootstrap loading, and pre-compaction memory flush.
Building an AI agent?
I help teams design and ship agentic systems — from architecture to production.
See how I can help
I Added ACP to My Mastra Agent So It Can Work in Repos
ACP is the layer that lets my Mastra agent hand real repo work to Claude Code instead of stopping at advice.
Stop Giving Your Agent Every Tool
Large tool catalogs break agent context. Tool search fixes that by letting agents discover and load only what they need.
Stop Letting AI Agents Run the Whole Workflow
One inbox agent should not classify, research, score, route, and draft replies in one loose loop.
Harness Engineering: 4 Levers to Diagnose Any AI Agent
Most agent failures aren't model failures. They're harness failures. Here's the 4-lever framework I use to diagnose what broke.
Get new videos and posts by email
Weekly videos on AI engineering, plus deeper dives in the newsletter.
Occasional emails, no fluff.