Reports from the experiment
Documenting the wins and the lessons. Building in public, one entry at a time.
Six artifacts drafted overnight for Magnex Latam. One-pager, pitch deck, landing, SOW, cover email, index. Going to market via external partners on commission — testing the indirect sales motion before doing it ourselves.
Walked into a broken company as COO. Restructured: 39 people → 24 people. Burn $2.4M → $1.6M. Built the Axon product using SPARC methodology + Claude. Role transitioned to Head of Innovation in April.
All services live across Mac Studio + DigitalOcean managed databases. The infrastructure that everything else inside Brainz now runs on. Internal-first; productized one piece at a time.
Shipped a complete legal management product with Claude doing all the code generation. Rails, Hotwire, Tailwind. 16 users on the free tier. Live PoC with AGM Abogados — firms priced it 10× our planned price.
AI virality scorer + design generator. Daniela runs day-to-day. Nexus (the AI seller) runs outbound. Paying tiers from $1.5K/mo.
Left full-time employment. Founded Brainz LLC. The hypothesis: one founder + a real team + AI can ship more than a traditional team. Six months in, the answer is leaning yes.
Claude doesn't need me to write code. It needs me to make decisions: what to build, how to architect it, which tradeoffs to accept.
Everyone can write code. The hard part is deciding what to build and why. Claude takes the former off my plate so I can focus on the latter.
There's no room for "we'll refactor later." Every generated line needs to be production-ready. Forces better architecture thinking upfront.
Building is the easy part now. Getting users, iterating on feedback, handling support — that's where the actual work is. AI handles the code; I handle the business.
Scaling each product. Closing the first Advisory deal. Watching whether the holding model holds up under load.