Abstract: For more than a hundred years, colloidal particles have been used to model the behavior of atoms or molecules. Recently, this approach has been reversed: Chemical reactions between molecules are now being used to model self-assembly of nanoparticles. Owing to the well-defined number and controllable spatial distribution of attractive surface regions, “patchy” particles exhibit directional interactions and assemble in clusters mimicking the symmetry of molecular structures. In particular, polymer science offers unique strategies to address the challenges in nanoparticle assembly. By using lessons of polymer physics and chemistry, we develop new paradigms for nanoparticle surface patterning and nanoparticles self-organization. To generate “colloidal molecules,” we used thermodynamically driven formation of surface-pinned micelles. A striking resemblance between block copolymers and nanoparticles enabled nanoparticle assembly in nanostructures with varying morphologies, all mapped by state diagrams. A marked similarity between step-growth polymerization and nanoparticle self-assembly enabled nanoparticle assembly with quantitative prediction of the architecture of linear, branched, and cyclic nanostructures, their aggregation number and size distribution, as well as the formation of isomers. Building on this similarity, we proposed the concept of colloidal chain stoppers, as well as random and block copolymers. This work has far-reaching implications for the molecular world (by offering simple, easy to visualize nanoscale models for polymerization reactions), and for the nano-world (by providing a polymer approach to nanostructures with structure-dependent electronic, optical, and magnetic properties).