Notes from the studio.
Practical thinking on product, design, and building AI-first. Real teardowns, the small fixes that move the numbers, and what AI-first actually looks like in the work. No theory, no fluff.
AI RolloutAI isn't stalling because teams don't care
Most companies don't have an AI curiosity problem. They have a governance problem. Why rollouts stall at the leadership-confidence gate, and the move that unsticks them.
AI StrategyYou don't have an AI strategy, you have experiments
Isolated AI wins quietly become tool sprawl: duplicate workflows, no standards, operator overwhelm. Structure has to come before scale.
AI AdoptionStop training teams on AI. Package the workflow
Teams don't need abstract AI training. They need recurring work turned into reusable, reviewable workflows. Adoption follows the work.
Operating ModelEveryone wants AI. But who owns it?
AI rollout stalls when everyone experiments but no one owns the operating model. It is an operating-model decision, not a tooling one.
AutomationDrowning teams ask: hire or automate?
In ops-heavy teams the real question isn't whether AI is interesting. It is whether it beats the next hire. Here is how to decide.
AI EraYour customers don't want your AI assistant
Most AI features ship to applause and die three weeks later. You can't out-LLM the frontier labs, so stop trying. A guest note on the one position that is actually defensible.
AI EraThe anatomy of an AI agent
Everyone is talking about agents. Far fewer can say what one actually is. The five parts that turn a model into an agent that does real work.
StrategyWhy founders stop seeing their own product
You've used it a thousand times. That's exactly why you can't see the friction a new user hits in the first thirty seconds — and why nobody on payroll points at it.
OperationsSoonThe smallest fixes move the biggest numbers
The highest-ROI product work is usually embarrassingly small. Three real examples of what quietly caps growth — and why they survive every sprint.
AI EraSoonAI won't take your job. Someone using AI will
I watched it happen this month. The skill wasn't the gap — the operating model was. And it follows you from career to career until you fix it.