FILTER · TAG
4entries · clear ✕
I committed a 15 MB knowledge graph and forced every agent to consult it before reading code, on the promise it would save tokens. So I ran a controlled test on a real task. It saved about 5%—but only by returning a worse, incomplete answer. Here is what I measured and why I pulled the whole thing out.
Read →A response to Sequoia's "Services as the New Software." The next durable businesses won't sell another piece of software that races the next model. They'll deliver the actual service, measured in real outcomes like recovered revenue, audit findings, and savings found.
Read →How I'm rebuilding my coding workflow around one /specs directory—complete architecture, a feature directory, an implementation roadmap, and a few RFCs—borrowing how the Claude team writes specs, with Coolify or Cloudflare Pages + Workers handling preview deployments.
Read →
After eleven years across hotel operations, revenue, and distribution, I no longer believe the PMS is the center of a hotel—or that it should be. Open Hotel Protocols is my proposal for what comes next.
Read →