Blog · working notes
Things I'm working on.
Short pieces about the operating model I run across four ventures, the Claude Code skills I publish, and the patterns I keep finding underneath the work. Updated when I have something useful, not on a schedule.
15 notes · updated July 2026
How I Won Lovable Shipped Season 1 ($100K, 5,800 Builders)
How I won Lovable Shipped Season 1: the six-week GiveFeedback build, what Lovable handled vs. what discipline handled, and the near-miss.
GiveFeedback: zero to paying clients in six weeks
How I built GiveFeedback.dev solo in six weeks, beat 5,800 builders for Lovable Shipped's $100K prize, and which decisions transfer to any founder.
What Is a Fractional Operator (and Do You Need One)?
What a fractional operator is, how the role differs from a fractional CMO, CTO, or COO, when founders need one, and how to evaluate one before hiring.
How to Run Marketing With AI Agents (No Team Required)
A solo founder's working setup for running marketing with AI agents: the minimum stack by domain, what breaks first, and where judgment stays human.
Why ChatGPT Doesn't Recommend Your Product: 5 Real Reasons
Five checkable reasons ChatGPT recommends competitors instead of your product, plus a ten-minute test you can run for each one today.
How to Sell AI to Hospitals: The Buying Committee Playbook
An operator playbook for selling AI to hospitals: the five people who can veto your deal, the compliance gauntlet, and why pilots stall.
Selling AI to Hospitals for Eight Years: The Verto Story
Employee #2 to a team of 80, 100+ hospital deployments, 10M+ patient journeys. The Verto years as a case study for founders.
Is Lovable Good for Real Products? An Honest Verdict
A yes-with-boundaries verdict on whether Lovable apps are production ready, from the builder who won Lovable Shipped with a real, paying product.
Vibe Coding to Production: The Eight-Item Checklist
The eight checks that take an AI-built app from demo to paying users: auth, data ownership, migrations, cost ceilings, backups, the marketing-site trap.
Migrating Off Lovable: When It's Worth It, and How
When migrating off Lovable is actually worth it (rarely, early on), the triggers that justify it, and a staged path for code, database, auth, and hosting.
Fractional CMO for AI Startups: What Actually Changes
Why AI products break the standard CMO playbook: hype-fatigued buyers, a higher proof bar, and why receipts-first content plus AEO win.
Does llms.txt Actually Work? The Evidence in Mid-2026
No major engine has committed to llms.txt and most files never get read. What the evidence actually shows, and why I ship one on every site anyway.
Why Healthcare AI Pilots Die: An Operator Post-Mortem
Eight years selling AI into hospitals taught me why pilots stall: no expansion owner, unsigned metrics, the integration cliff. Here's the escape plan.
The app was never the moat (I rebuilt one in an afternoon to be sure)
AI rebuilt a finance tool I'd have paid for in an afternoon, so I gave it away. The lesson wasn't that software is dead. It was that the app was never the moat, and what's left is harder to clone than a feature: a person worth betting on.
Skills are the new package manager (and they work in Lovable too)
How the same Claude Code skill folder works across Lovable, Cursor, Codex, and a half-dozen other agents reading AGENTS.md at a repo root. The mental model is npm packages, for prompts.
Operator notes, monthly.
Working notes on agentic marketing, Claude Code skills, and the operating models behind four ventures. It ships when there is something worth reading.