A1FP — AI First Principles
A course for product leaders who want to reason about AI from first principles — to understand what the systems can and can’t do, where trust comes from, and how to build with them — rather than just wiring up the latest model.
Most AI product work today is integration — wiring a model into a flow and hoping the demo holds. This course is for product leaders who want the layer underneath: what these systems actually are, where their reliability comes from, and how to design products that stay trustworthy when the model is confidently wrong. If you want to reason from first principles rather than from the latest launch post, this is for you.
A working mental model of how these systems produce outputs, stripped of both hype and mysticism.
Why trust is a property of the product and the process around the model — not of the model itself.
Review and oversight patterns that still work when volume goes up 100x and the easy cases stop arriving.
The engine got faster; the steering didn’t. How to ship fast with AI without losing control of where you’re going.
I’ve shipped applied AI across a decade — from CategoryIQ’s award-winning catalog intelligence, to taking a Walmart team’s AI weekly-active use from under 20% to over 90% and putting AI behind 70%+ of the roadmap, to the targeting and measurement systems behind Zepto Ads. A1FP distils what actually generalised across all of it: the first principles, not the framework of the month.
Small cohort, limited seats. Waitlist members get first access and the early-cohort rate before it opens publicly.