I Started a Garden Experiment. AI Is Involved. It's Going Well (Mostly).
Happy Friday. I have a confession and a new project and they are related.
The confession: I have been quietly building a cut flower garden since February using Claude as a genuine thinking partner and not to generate a plant list and copy it, but to actually learn. To push back. To send photos of seedlings and get specific feedback. To document every order, every mistake, every seed start, every course correction.
The project: I’m writing all of it up. In public. On my blog.
Two posts are already live and I want you to know about them because they’re about more than flowers.
The Two Posts So Far
Post 1 is the origin story of why 2025 was a quiet failure, $400+ of carefully planned bulbs that mostly didn’t bloom, and the decision to start over with a better approach. There’s also a dog named Mabel who has strong opinions about freshly turned soil.
→ I Killed My Garden Last Year. Here’s What I Did Differently. | coffeeforthebrain.com/🌻-growing-with-ai-i-killed-my-garden-last-year-heres-what-i-did-differently/
Post 2 is where it gets interesting. While writing it I discovered I had the wrong hardiness zone for Bettendorf using Zone 6a everywhere in my plan, when the 2023 USDA update shows Zone 5b. I put the correction at the top of the post rather than hiding it, and the error itself became the best AI teaching moment in the series. None of my plants fail because of it. But the how and why of that mistake is worth reading.
→ Zone 5b Is Not Forgiving And That’s What Makes It Interesting. | coffeeforthebrain.com/zone-5b-is-not-forgiving-and-thats-what-makes-it-interesting/
What the Series Actually Is
It’s called Growing With AI. Six threads running through a full growing season from February planning through October when the last dahlia tubers go into storage.
The Garden — What I planted, what happened, what failed, what exploded. Photo-heavy, chronological, honest.
The AI Sessions — Behind the conversations — what I asked, how AI responded, where it was wrong, and what I learned.
The Learning — Concepts that clicked. Written as a learner, not an expert.
The System — Tools and frameworks you can steal. The spreadsheets, the prompts, the planning templates.
The Science — Why flowers do what they do. The biology behind what’s happening in the dirt.
Rooted — What happens when technology reconnects you to the natural world instead of pulling you away from it.
Each post is primarily one of those threads but woven with elements of the others. You don’t have to read all of them, but if you’re someone who thinks about AI in education, posts in The AI Sessions and The System are going to land differently than they would on a straight gardening blog.
Why I’m Doing This Publicly
I talk to teachers and administrators about AI collaboration constantly. I have a philosophy about it to adapt with integrity, not adopt with fidelity(learned from a great mind) and I use it in workshops and keynotes and newsletter pieces.
But I realized I had been talking about AI as a learning scaffold mostly in abstract terms. This project is my attempt to document what that actually looks like in a real, extended, messy, real-stakes learning process. One where I’m not the expert. One where I get things wrong in public and correct them.
If I’m going to tell teachers that productive AI use looks different from outsourcing, I should be able to point to something real. This is the something real.
The garden is the domain. The learning is the point. And honestly I’m having a lot of fun with it.
New posts go up on coffeeforthebrain.com as the season unfolds. I’ll drop a note here in Chaos Navigators when something worth your attention lands.
For now go read Post 1 if you want the backstory, Post 2 if you want the AI teaching moment. Either works as a standalone.
And if you have ever caught an AI giving you confidently wrong advice in a domain you know well I’d genuinely love to hear about it in the comments.
Happy Friday. Go outside if you can.
— Aaron