Why executives need a dark factory
September 2026
A small AI production capability can advance work while your attention is elsewhere. Building one teaches you what to delegate, how to judge the result, and what to commission next.
Read the piece →Latest essay
Why AI writing sounds like AI
6 min read
Default AI prose becomes generic when the assignment withholds purpose, evidence, trade-offs, and judgment. A practical editorial method puts those stakes back into the work.
Read the essay →Recently featured
Do more with more
September 2026
Recovered capacity can support learning, better management, and a more sustainable working life. Decide which benefits deserve protection.
Read the piece →SaaS isn't dead. The dashboard is yours.
September 2026
A custom operating view can connect useful services around a decision. Its value depends on source fidelity, permissions, and an owner for maintenance.
Read the piece →Earlier essays
Stand by is not silence
September 2026
Protect attention with an acknowledged pause: preserve the question, assign an owner, set a return time, and make earlier escalation explicit.
Read the piece →Shape the work before you stamp it
September 2026
Change the representation of an unsettled problem to expose its assumptions. Standardize the useful form only after it has survived meaningful review.
Read the piece →Route the work by terrain
September 2026
Choose AI capability for each step by its ambiguity, consequences, and verification cost. Project prestige is a poor guide to model spend.
Read the piece →Build the harness the work has earned
September 2026
Add tools and coordination when repeated work justifies them. Start with a small experiment, then invest where its actual limits become visible.
Read the piece →AI literacy
September 2026
A practical test of AI literacy: understand how an answer was produced, check its source, and know what the decision still requires.
Read the piece →AI•DOOM Assessment: the first cut of an operating maturity diagnostic
August 2026
How the first AI•DOOM Assessment used operating questions to start a conversation, and why its original scores were never a verdict on an executive.
Read the piece →The cost of waiting to start with AI
August 2026
A five-year model separating calendar-time improvement in AI from the operating practice an executive builds only after starting.
Read the piece →How we chose Quiet Turn
August 2026
The Quiet Turn naming search changed when brand meaning and domain availability became separate decisions.
Read the piece →What a model can't decide
August 2026
Use AI to examine assumptions and alternatives while keeping the objectives, evidence standards, and responsibility for consequential choices explicit.
Read the piece →The meeting that disappeared
August 2026
A hypothetical status meeting shows how to test whether a summary can replace reporting while preserving the discussion the team still needs.
Read the piece →Buying back attention
August 2026
Time saved on preparation becomes useful attention only when leaders account for review costs and protect space for consequential work.
Read the piece →