Personal archive · 2017-2026
Working Theory.
Ideas I chose to work through in public. These are not commissioned articles, company case studies, or the practitioner publications cataloged elsewhere on this site. They are personal arguments about technology, markets, and what progress should give back to people.
01 · Recent thinking
Before the chat window.
The newest essays return to questions first raised years earlier: what the machine is actually optimizing, what evidence deserves belief, and where human judgment still matters.
Before AI Had a Chat Window
Fifteen years with machine intelligence, from model operations and data integrity to private inference, agentic development, and keeping a human being in the middle of it.
Building Trust in Digital Measurement
Why experimentation, transparent incentives, and evidence-based decisions matter more than another attribution model.
Beyond the Hype
AI performance depends on three choices advertisers still control: training data, business objectives, and the environment where decisions are deployed.
02 · Earlier questions
What should innovation give us?
Three consecutive essays from July 2017 follow one question from the manual transmission to marketing AI: if technology removes the task, does it also remove the skill?
The tool and the skill
The manual transmission disappears. The larger question is what human capability disappears with it.
Automation that teaches
The strongest AI should do more than remove work. It should leave the operator more capable than it found them.
Augmented capability
Control, autonomous scale, and learned intelligence converge into a more useful definition of progress.
How Predictive Marketing Creates Optimal Travel Experiences
A 2017 view of self-learning systems, connected customer data, and the travel experience that emerges when fragmented decisions become one journey.
Why preserve them
The dates matter. They show the questions before the answers became obvious.