Free resources
Four visual modules. Every idea is a diagram first, with panels to open for why it matters to a product leader, how the mechanics work, or a refresher on the basics.
fig. 1 · four modules, 47 lessons
Modules
What a model actually is, how a chat assistant gets built, and the four ways you can turn one into a product. Every idea is a picture first; open the panels when you want the mechanics or a refresher on the underlying tech.
Read the primerHow a model learns from data, how you know whether it is any good, and how it gets from a notebook into a product. This is the ring underneath LLMs, and it still runs most of the AI in production. Sections 04 and 09 carry the diagnostic habits from Andrew Ng's Stanford CS229.
Read the primerWhen AI agents write the code, the specification becomes the work. This module shows the loop, what a spec has to contain, where people stay in charge, and what a product leader's week looks like when prototypes take hours.
Read the primerFollows Andrew Ng's DeepLearning.AI course 'Agentic AI' (Coursera, updated September 2026) at leadership altitude: the autonomy spectrum, why iteration beats a bigger model, the four design patterns, and the evaluation discipline that separates demos from products.
Read the primerEvery diagram uses one colour code:DataModel / computePeopleTools & systemsRisk
Suggest a topic
Tell us what would be useful and we will consider it for the library.