The architect's side.
My brother called me on a Tuesday. He said he was going to build an app — a fitness app, for Total Gym machines. He had the idea sketched out. He was going to start that night.
A few months prior, Chris bought a Total Gym off Facebook Marketplace. Loves it. Has been trying to get me to buy one for months — thinks he's Chuck Norris. He told me he decided to build an app to keep it from becoming a clothesline, and if I didn't help, he'd roundhouse kick me to the head. It seemed like the perfect project to field test AI at real architectural scale.
I said I'm in. And then, because I've been an Enterprise Architect long enough to have seen what happens when smart people move fast without a plan, I also said: we're doing this right.
What my job actually was
Chris can move fast. If you've read his posts, you know this. The velocity he's describing is real — I watched it from the other side of the room. What I brought was different: I was the person asking what happens in six months. What does this look like with a thousand users? What does this look like when we need to change the data model?
That's not slowing things down. That's making sure the thing you're building survives contact with the future. Speed without structure is how you paint yourself into corners. My job was to make sure we weren't building corners.
The decisions we argued over most — what to leave out of v1, how the progression engine should work, what a "level up" means and exactly when it fires — those are the decisions that give the app a point of view. Software with a point of view is software worth using.
What surprised me about building this way
I've been going deep on AI for a couple of years — not as a user, as a student of the architecture. Agentic patterns, context management, the failure modes that don't show up in demos. When Chris called me, I wasn't new to this.
What I didn't know was whether the velocity would hold under real architectural pressure. That's a different question. I've seen a lot of "fast" — projects that sprinted in week one and were paying down debt by week six. The question was never whether AI is capable. It was whether the decisions you make at speed compound correctly.
Eight weeks in: the debt is manageable. The velocity is real. Having architectural judgment in the room from day one — not brought in afterward to untangle it — was the difference.
That's the shape that works. The person who moves fast, paired with the person who thinks three moves ahead. Each one making the other sharper.
Apple approved.
SLED is built. Six exercises, three sets, automatic progression. A system for people who have the machine and have been hanging their clothes on it.
We just got Apple Developer approval. The next phase starts now.
I've worked with a lot of fast people. I haven't worked with one who can do what Chris does — product instinct, copy, code, and sales in the same week, from the same head. AI doesn't create that. It amplifies it.
How cool is it that I get to have my brother's back while we do it. Chris is telling his side. We're both in the rabbit hole now — subscribe and follow us down.
— Dan