The Blank Slate Problem

Module 2, Skills & Memory | Essay 1 of 8

Your AI agent doesn't get smarter the more you use it. It gets better when you teach it about you: and memory and skills are how you do that.

Without them, you're stuck in a loop. Every session starts from nothing. The agent doesn't know your name, your role, how you communicate, or what you've been working on. Every conversation begins with a blank slate, and you're responsible for filling it in.

That's the core problem this module solves.

Here's what the loop looks like in practice. The novelty of a new AI tool wears off after a few weeks. Then the repetition sets in. You find yourself writing the same background every single time: who you are, what you're working on, how you like things formatted. You're not getting smarter at using the tool. You're just getting better at explaining yourself from scratch.

None of that re-explaining is useful work. It's overhead. Time spent briefing the agent on things it already knew yesterday.

Memory changes this. Instead of re-explaining yourself every session, you teach the agent once. You write down the things that are always true about you and your work, and the agent reads them before it responds. No re-briefing. No context lost between sessions. Your agent starts every conversation already knowing you.

Skills take this further. If memory is what your agent knows about you, skills are what it knows how to do for you. A skill is a saved set of instructions for a task you do repeatedly. Instead of figuring out the best way to run your weekly summary every week, you figure it out once, write it down as a skill, and your agent follows it every time.

Together, memory and skills solve the blank slate problem. Not by making the agent smarter, by making it informed.

What you'll leave this module able to do

The rest of this module teaches you how to build both. Here's what you'll walk away with.

You'll know what memory actually is, not some invisible system running in the background, but a set of plain-text notes your agent reads before it responds. That's it. Once you understand that, writing and updating memory notes becomes straightforward. By Essay 3, you'll have written your first ones, and you'll feel the difference immediately in how your agent responds.

You'll know the difference between personal context and task context, and why mixing them up causes problems. Personal context: your role, how you communicate, what you care about, belongs in memory. Task context, this project, this deadline, this specific audience, belongs in the conversation. Getting that distinction right keeps your memory notes useful and your conversations clean.

You'll have built your first skill. Not a template, not a prompt you've saved somewhere: an actual skill your agent knows how to run on demand. You'll test it on a real task from your work, refine it based on what comes back, and leave with something you can use on Monday morning.

You'll understand how skills compound. One skill saves you five minutes once. A library of ten skills, each covering a task you do every week, changes how much time you spend on repetitive work. By Essay 7, you'll have a plan for building that library over 30 days.

And you'll understand what it feels like to work with an agent that actually knows you. Not just a smart tool you're using, but something configured to your work, your style, your situation. That's what memory and skills build, not a smarter agent, an informed one.

Start with memory. It takes 10 minutes and changes how every conversation feels. Then build your first skill, because that's where the real time savings compound.