The New Lock-In Lives in Memory
APIs can be swapped. Habits, approvals, and learned context are much harder to move. Why the next major AI lock-in may cut deeper than classic software dependency.
Essays on AI, knowledge work, and the transformation of organizations.
Long-form writing about how AI is changing the way we think, work, and make decisions.
The Most Expensive Conflation: Tasks and Jobs
55% of companies regret their AI layoffs. The task analysis was correct. The job decision was not. What this reveals about the difference between tasks and jobs.
The Vocabulary Gap
1.59x better quality. Zero model change. Just vocabulary.
Executive Briefing: The Invisible Input
What stands between your AI business case and a noble gas in the Persian Gulf. In not a single AI business case I've seen in the last six months does the word helium appear.
The Wedding Planner and the Artist
Why AI brings everyone to 6/10 – and why that's exactly the problem. When everyone becomes equally passable, the one who's better than passable wins.
Why AI Tools Fail – And Where the Real Lever Is
The bottleneck isn't the software. It's the knowledge that only exists in people's heads. Why institutional context is the real lever – and why most organizations are at Level 0 while believing they're at Level 2.
The Invisible Economy
What happens when your products don't exist for agents? Six companies built agent commerce infrastructure in a single week -- but the industries with the most to lose are resisting the longest.
The Fast Spiral Eats Its Young
Why AI-accelerated iteration without real feedback loops just produces garbage faster — and what drone operators taught NATO about modern knowledge work.
Who's Writing the Spec?
Why the bottleneck of knowledge work isn't production but the ability to say what you want. On Taste, Specification, and Evaluation -- the three competencies that are suddenly scarce.
Artifact Production Just Got Cheap
What happens when the cost of producing knowledge work artifacts drops to near zero? On taste, brand, and data models as the three things that remain.