Hofmeister Builds
AI looks easy. It explains itself, it answers in plain English, and it demos beautifully. So leaders assume the hard part is buying it.
It isn’t. The hard part is people actually using it. Trusting it, being trained on it, feeling safe enough to try it badly at first. That’s not an IT problem. That’s an HR competency, and it’s why most companies never get off the ground.
Harder than it looks. Not impossible. Here’s what it looks like when someone from HR builds.
THE BUILD STORIES
I don't just post what worked. Every build includes the walls, the dead ends, and the thing that survived into the next one, written from the timestamped records.
Recursos → A Spanish-language AI starter page, built with a community health promoter, for people who have never used these tools.
WHAT'S HERE
AI at Work
Five sentences you already say at work, and how each one works on an AI. Written for someone who has never opened one of these. What I’m working on in it now is the first step, so a beginner actually tries something instead of reading about trying it.
Builds
Every project so far, each one with the raw log of how it actually went: the walls, the dead ends, what carried into the next build. The one I’m in the middle of is a test of three AI models on the same real task, graded.
Kid Lab
Robot athletics, designed by my son. He has final say on everything in there. Basketball, the warped wall, and boxing all work. Soccer just went in, with a goalkeeper who dives.
The Lab
A 3D experiment: the builds orbiting in a starfield, with a sunrise you drive by scrolling. It’s where I try interface ideas that don’t belong on a normal page yet.