I build AI-native software, run Merchant Protocol, and ship Sulla Desktop. I think out loud about where AI is going for working businesses — without the hype, without the course-grad nonsense. 25 years deep in production code; betting the next 25 on where this all lands.
Two real businesses, one platform thesis. Both ship every week. Both have actual customers. The case study is me.
Open-source multi-agent AI platform — workflow orchestration, agent comms, vault, marketplace, multi-model routing. The bet I'm making on where AI for working businesses lands in the next 24 months. Already shipped, growing community, customers in production.
Visit sulladesktop.com →Free skills you point your Claude at to make it dramatically more productive. Same playbooks I use to ship paid client work. Scaffolding tool flagship.
playbooks.sulladesktop.com →Custom software development. AI as the build engine, an open-source platform as the foundation, a finished system you own. 25 years of knowing where the hard 20% lives.
merchantprotocol.com →Five themes: AI-native dev, where the market is going, build-in-public on Sulla, custom dev war stories, anti-hype takes. One of these every few days. Sundays for the longer pieces.
Model providers can't credibly route to cheaper competitors — it cannibalizes their margin. The structural moat is real, but it's a window, not a permanent condition.
It's not that AI tools fail. It's that the long tail of integration work outlasts the developer's patience. AI doesn't fix that — it gives you infinite grinding capacity to push through.
The marketplace was the hardest piece of the platform to ship — not because of code, but because of taxonomy. Workflows vs. recipes vs. functions vs. agents.
25 years in production code. 15 of them deep in PHP / WooCommerce / Magento on real businesses with real revenue at stake. Merchant Protocol pays the bills today; Sulla Desktop is what I think comes next.
I write because the AI conversation is dominated by people who don't ship and silenced by people who do, and that imbalance is bad for everyone trying to make decisions about real businesses. I'd rather be useful than impressive.
Christian. Husband. Coeur d'Alene, Idaho. I take the work seriously and the rest of it lightly.