What Does Your AI Agent Actually Remember About You? Give It a Memory, Then Make It Roast You

Ask a fresh coding agent about your project and it will answer with total confidence and get almost everything wrong. Agents don’t remember, so every session is a first date with your codebase. That forgetting has real costs: 1. Hard-won context disappears when the terminal closes. The reasoning behind your code is scattered across sessions nobody will reopen. 2. Every new session starts with expensive rediscovery. Agents spend tokens re-learning structure, conventions, and past decisions instead of shipping. 3. Handoffs break down. When a teammate or a different agent picks up the work, the “why” doesn’t come with it, so reviews turn into guesswork. In this hands-on workshop, we’ll put agent memory to the test. Fresh agents take a Memory IQ quiz every round, and you’ll watch their scores climb as you build continuous memory with Entire. You’ll learn how sessions and checkpoints become reusable project memory, how to brief a brand-new agent in seconds, how to verify memory after the code moves under it, and how to inherit a stranger’s repo with nothing but their agent’s memory to guide you. You’ll finish by reviewing your changes in Trails and turning everything your agent learned into a roast-style “Wrapped” card you’ll want to share. This workshop is for engineers, tech leads, and reviewers of all experience levels who want coding agents that get smarter with every session instead of starting over.