Thoughts on GenAI, enterprise transformation, product management, and the intersection of technology and business.
Replacing Cal.com with our own booking engine took one evening and four pull requests — because the engine already existed inside our product. On reusing your own systems, deleting a subprocessor, and why the build-vs-rent calculus is broken.
One model plans and builds, a different model from a different company attacks its work, and the human keeps exactly one decision. My two-AI loop for shipping production code you can actually trust.
A worker on a cron schedule invents, scripts, and renders a 30-second video every three days for $0.37 — with zero minutes of my time. On scope engineering, guardrails before spend, and staged trust for autonomous agents.
The consulting pyramid that minted partners for decades is collapsing. McKinsey now has 20,000 AI agents working alongside 40,000 humans. This isn't disruption from the outside. It's transformation from within.
Middle managers existed to translate strategy into execution and surface blockers upward. AI agents are doing both. This isn't about eliminating middle management. It's about radically redefining it.
The Product Owner role was designed for a world where human teams executed work. Now your 'team' includes AI agents. This changes what it means to own a product.
The agile methodologies we've spent decades perfecting were designed for a world where implementation was the bottleneck. AI just collapsed that assumption. Here's what comes next.
Build vs. buy was already complex. Now there's a third option: configure agents to do what you used to build or buy. This changes the economics of capability sourcing in ways most enterprises haven't processed yet.