The First Diffusion-Based Forecasting Model for Music Royalties

Most forecasting models are built on streams and consumption data - proxies that tell you what's popular, not what something is worth. Ours is different. Chapter Two's model is trained directly on royalty earnings from 100,000+ catalogs and over $1B in processed royalty data, modeling the thing that actually matters: cash flow.