FAQ

FAQ

What is the distinction between the world layer and the data layer?

In the three layer model-data-world model, the boundary between the world and data layers can be understood as such: interventions in the world layer are responsible for determining the totality of all records stored across all devices in a variety of schemas. The data layer is concerned with how we store, sample, link, merge, and augment this totality of records, such that at the model layer the designer has some simple “dataset D” (oftentimes “n rows of <x,y> feature-label pairs” for supervised learning or “n tokens” for generative modeling) they can work with.