Learning by Building
I've worked in digital for close to twenty years. Web design. Development. User experience. Digital operations. The work has changed constantly, but the pace of change over the last year feels different.
Every conversation seems to come back to AI.
Some people are excited. Some are terrified. Most are just trying to figure out where they fit.
I'm somewhere in the middle.
I don't think AI is replacing everyone tomorrow, but I do think the people who learn to work alongside it will have an advantage over the people who don't.
That realization led me to a strange purchase.
A used Mac Mini.
Not a course. Not another certification. Not another productivity app. A computer dedicated entirely to building and experimenting with AI systems.
My goal isn't to build the next startup. It's much simpler than that.
I want to understand how these tools actually work.
Not the polished demos.
Not the viral LinkedIn posts.
The real systems underneath.
How they connect to each other. How they fail. How they recover. And how they can actually help someone run a business.
The first project is something called OpenClaw, an open-source framework for building AI agents.
If you've never heard of it, that's okay. A week before I started, I hadn't either.
What I quickly realized is that building AI systems feels a lot like building websites twenty years ago. The possibilities are exciting, the documentation is incomplete, and most of your time is spent figuring out why something isn't working.
Oddly enough, that's part of the fun.
So that's what this Journal is about.
Not becoming an AI expert.
Not selling a course.
Not pretending I have all the answers.
Just documenting what happens when a digital professional decides to learn by building.
The Mac Mini is plugged in.
The first agent is running.
And almost nothing is working exactly as planned.
Which is probably a good sign.