Mar 11, 2026 · 10 min read
Every time we've made a productive tool more efficient, we used more of it, not less. AI is no different — except this time the tool is augmenting thinking itself. That changes what happens when you restrict it, and what happens even when you don't.
Mar 10, 2026 · 9 min read
Every decade, technology makes us dramatically more productive. Every decade, GDP growth slows. These two facts should not coexist — and understanding why they do reveals the defining economic tension of our era.
Mar 09, 2026 · 18 min read
If P work is the new bottleneck, your team model has to change: who you hire, how you structure ownership, how decisions get made, and what you measure.
Mar 08, 2026 · 16 min read
When to stop checking, when to keep going, and what actually changes when AI takes over the deterministic layer. Case studies and an implementation checklist.
Mar 06, 2026 · 11 min read
Hallucinations, misunderstandings, and drift — the three types of P work failures, and exactly where to put your human checkpoints to catch each one.
Mar 04, 2026 · 10 min read
AI is offloading your deterministic work. Great. But what's left is concentrated probabilistic work — the kind that exhausts humans and compounds risk exponentially across every layer of decision-making.
Mar 03, 2026 · 9 min read
As AI masters deterministic execution, the real leverage shifts to probabilistic work — where ambiguity is the job, errors compound, and someone has to own the consequences.
Mar 02, 2026 · 11 min read
Your morning routine is D→D→D. A job interview is P. Parenting is P→P→P with no guardrails. The D/P framework isn't just for AI products — it's how everything works.
Mar 01, 2026 · 10 min read
Your AI product isn't 'AI-powered.' It's a specific combination of deterministic and probabilistic steps — and knowing which is which changes every decision you make as a PM.
Feb 28, 2026 · 9 min read
Everyone's building 'AI agents.' Most of them are just prompts with a loop. Here's what actually makes something an agent — and a walkthrough of how a prompt graduates into one.