A few pieces, working together
When founders hear “AI stack,” it can sound like a deep technical topic. In practice, the underlying picture is simpler than it sounds — a small number of components that fit together in fairly standard ways.
Understanding that picture lets you make sensible decisions about what to use, when, and how to combine things — even if you’re not the one writing the code.
This topic walks through the main pieces of a modern AI stack and how they relate to each other.
Models: the brain
At the heart of any AI product is a model — the underlying system that takes inputs (often text, sometimes images or audio) and produces outputs.
Most modern products don’t train their own models. They use models built by major AI labs, accessed through APIs. You send a request, the mod...
