Essays & insights on ai & technology.
A bigger context window is not more usable context. The usable window, the three taxes of a full window, and the context engineering fix every founder needs.
AI agent observability fails because agents return confident wrong answers, not errors. A four-trace framework for catching silent failure before customers do.
Deploying AI agents breaks when authority outruns oversight. The Blast Radius rule and the Autonomy Dial show founders how much control to give an AI agent.
AI collapsed the cost to build software. Should you replace your SaaS? The build vs buy decision for the AI era, and the ownership tail almost nobody prices.
AI agents are becoming the buyer, not just the tool. Here is how founders build software a machine customer can find, trust, and actually choose to use.
Spec-driven development flips the source of truth from code to the spec. Why the spec is now the unit of work, and how to write one an AI agent can ship.
AI platform risk, explained for founders: why AI vendors climb into their customers' markets, plus the dependency map and moat design that survive stack creep.
AI regulation is retreating in 2026, but enterprise buyers still gate revenue on trust. The AI governance stack: how to build governance as a moat, not a tax.
Loop engineering, not prompts, is the real moat in AI agents. A field guide to the loop ladder, the verification gap, and building agents that stay reliable.
Agent delegation, not model quality, decides whether your AI ships or stalls. A builder's framework for when one agent beats a swarm and how to grant autonomy.
Building something interesting? Looking for an investor who’s been in the trenches? Or just want to say hello? I’d love to hear from you.
One email when I publish. Essays on reskilling, rewiring your mindset, and building your next thing. Free, no spam, unsubscribe anytime.