Hero image for Why I'm Publishing My AI Playbook in Real Time
By Scott Armbruster

Why I'm Publishing My AI Playbook in Real Time


Today’s Highlights

  • The moment between “I should document this” and “I should publish this” is smaller than you’d think
  • Jumped between five projects today without re-reading a single note
  • Someone asked point blank if I should be publishing this, and I didn’t have a good answer
  • Santiago Guy’s stat about AI-generated words that’s been stuck in my head all week
  • Jordan Peterson, of all people, said something that connected to hitting publish

The Notebook That Wanted to Be a Blog

Somewhere around the third context switch today, hopping from a consulting engagement to a content system build to debugging automation for a financial services client, I caught myself doing something I’ve been doing for weeks without naming it. Taking notes. Not task notes. Observation notes. Patterns I was seeing in how AI changes the cadence of a day. Small realizations that felt like they mattered beyond the specific project.

Then someone asked me point blank: should you be publishing this?

The honest answer was complicated. There’s vulnerability in showing your daily work in real time. Not a polished case study six months later. Not a keynote where you already know the ending. The raw version, including the days where the most important thing you did was fix a broken file path.

Most of what gets written about AI right now falls into two buckets: breathless hype or abstract strategy. Almost nobody is showing the Tuesday. The actual texture of a day where AI is deeply embedded in how you think, build, and move between problems. So I’m going to try.

When Your Tools Remember What You Forgot

The thing that made today’s pace possible wasn’t speed. It was memory. Not mine. My AI assistant’s understanding of every project I’m working across. Where the files live. What decisions I made last week. Which client cares about which constraints.

Early in the day, the assistant needed to navigate a directory structure spanning a dozen active projects. It already knew the layout. Not because someone programmed a map, but because over weeks of working together, it built one. When I jumped from a healthcare-related engagement to tweaking my personal brand site to reviewing automation scripts, the context was already there. No ramp-up. No “where was I?” No twenty minutes of re-reading my own notes.

This changes how many things one person can hold at once. The old constraint wasn’t intelligence or hours. It was cognitive overhead. The tax you pay every time you switch contexts. When that tax drops close to zero, you don’t just work faster. You work wider. Five client projects, personal brand work, infrastructure builds, and a growing collection of ideas I’m curating for later. Not because I’m superhuman. Because the bottleneck moved.

From the Vault

Santiago Guy made an observation this week that landed hard: AI can generate millions of words for a few dollars. Humans speak maybe 16,000 words a day. The math seems to favor machines. But Guy argues the opposite. When words are infinite and cheap, the scarce resource becomes lived experience. Real stories. Genuine perspective that didn’t come from a training dataset.

That’s partly why this journal exists. Not because the world needs more words about AI. Because it might need more honest ones.

Separately, a stat keeps rattling around my head. Someone surveyed 2,000 people at a business event recently. Ten were using agentic AI. About 20% had even heard of Claude. We’re deep enough in this world that it’s easy to forget: for most people, this hasn’t started yet. The tools I leaned on today to juggle five engagements and a content system are invisible to 84% of the market. They don’t want better AI tools. They want specific problems to stop existing.

And Jordan Peterson, of all people, said something that connected to the decision to start publishing. He talked about how your life becomes the sum of whatever you practice most. The easy thing is to keep private notes and never put them in front of anyone. The harder thing is to practice in public, knowing some days the insight is sharp and some days it’s just a guy fixing file paths. But that’s the version I’d actually want to read from someone else.

Today was the first day I stopped taking notes for myself and started taking them for you. Whether that was smart or reckless, I genuinely don’t know yet. But I think we’ll find out together.