Writing
Blog
Thinking about AI product leadership, building with AI tools, governance, and the economics of shipping AI in production.
Writing
Thinking about AI product leadership, building with AI tools, governance, and the economics of shipping AI in production.
Showing 13–24 of 77 articles

Energy, chips, systems, models, applications. Every layer matters. Only one pays compounding returns. A framework for picking yours.

Every useful agent becomes a power user of the SaaS underneath it. Install base explodes, API calls multiply, workflow gets more essential, not less.

Six production chat surfaces, a habit of breaking every AI chat in the wild, and the defence-in-depth stack that keeps your prompts contained.

Keith Rabois' barrels vs. ammunition framework reframes what AI changes about teams. More ammunition without more barrels solves the wrong problem.

Most enterprise AI teams centralise first, then decentralise. Both fail. Here's the hub-and-spoke structure that actually works.

90%+ enterprise AI tool access, most people stuck in chat. The rollout bottleneck isn't the model. It's the harness. Here's the product fix.

Product teams reflexively strip onboarding friction. Intentional friction that helps users understand why the product is for them increases conversion.

The top AI user in high-performing companies isn't engineering — it's the CMO. Here's why that's the real signal for whether AI adoption has reached the decision layer.

The best AI growth teams deliberately sacrifice short-term metrics. Restraint on pricing, error handling, and safety compounds into retention and trust.

Enterprise software encodes decades of domain knowledge across every architectural layer. Vibe coding can't shortcut what took thousands of people 25 years to accumulate.

Growth teams trained in linear markets spend 70% on small experiments. In exponential markets, that allocation captures a rounding error.

AI coding tools tripled engineering output overnight. PM and design headcount stayed flat. The ratio broke, and most orgs haven't noticed yet.