Now
Current position, current heading.
A now page is a fix on a moving chart: where the work actually is, dated, and revised as conditions change. Each week starts with one meaningful question and gets examined through the same four lenses as everything else here.
Research threads
Four lines in the water.
Human connection as the durable advantage
If AI absorbs the routine work, what exactly is left that customers and employees will not trade away — and how do you organize a business around protecting it?
Business models built for conditions that no longer exist
Most operating models quietly assume pre-AI costs of coordination, production, and expertise. I am mapping which assumptions break first.
AI-plus versus AI-first
The difference between decorating an old process with tools and redesigning the process around new conditions — and the sequencing that separates them.
Trust, transparency, and change leadership
What adoption actually costs in trust, how missteps get recovered, and how cross-checking survives when one crew member is a model.
On the desk
Open questions, under the instruments.
- How do leadership teams keep cross-checking alive when one crew member is a model?
- Which AI adoption sequences produce their failures cheapest and earliest?
- What does trust cost to rebuild after a bad automation call — and who pays it?
If one of these is your question too — or you have evidence that bears on it — that is exactly what the connection paths are for.
In the water
What I'm testing on myself first.
- Running my own research and publishing workflow with AI assistance, logging where the model is genuinely useful and where review costs more than the help.
- Converting one repeated workflow at a time into an AI-assisted process, using the same two-week field method published in the Learning Lab.
- Testing how evidence labels change the way readers challenge an argument — this site is itself the experiment.