What's actually happening inside AI rollouts. What works, what stalls, and what it costs.
Featured
Nov 9, 2026
•
4 min read
Nov 1, 2026
2 min read
Oct 11, 2026
All posts
A year without the news cycle taught me something I tell every client about AI.
Your best AI users are building tools nobody agreed to maintain. Handing them to engineering has a cost, too.
An agent that notices what you dropped is the first kind of AI that could reach the employees who open it twice a week. The question headed your way is what to give it access to.
I'm willing to try more ideas when a failed experiment doesn't take my evening with it. Imagine that happening in the physical world.
Oct 4, 2026
Two of the oldest shortcuts for judging people are breaking at the same time.
Sep 28, 2026
We turn expertise in one field into authority over everything nearby.
Sep 21, 2026
3 min read
I heard it building AI companies in 2010 and 2017. The forecasts keep changing. The confidence doesn't.
Sep 14, 2026
6 min read
If model capability tops out, the frontier labs are racing toward a commodity. Wrappers take the premium.
Sep 6, 2026
5 min read
Everyone says AI is in the first inning. The physics, the scaling data, and a measurement problem nobody talks about suggest we might be further along than that.
Aug 30, 2026
Your AI bill is exact. Its value isn't. Companies need two different scoreboards before the CFO starts asking questions.
Aug 23, 2026
A four-step test for whether an AI workflow will actually give you time back, or just move the work.
Aug 16, 2026
AI made building cheaper. It didn't touch who decides what gets built.
Aug 10, 2026
Company-wide AI licenses are overhead. The money is in narrow projects nobody calls transformation.
Aug 3, 2026
Predicting where AI will actually pay off with a simple framework.
Mar 3, 2026
Most AI tools fail not because they’re bad, but because they ask too much of busy humans.
Feb 17, 2026
20% of your organization has been using AI even before you told them to. You can learn from them.
Feb 3, 2026
LLMs are converging and changing - but the quality of your workflow remains.
Jan 20, 2026
Becoming more efficient (while still being essential).
Dec 30, 2025
Your people are optimistic about AI, and that optimism can leave a few things unsaid.
Dec 16, 2025
Making the use of AI as routine as checking email.
Nov 24, 2025
