Abstract
Structural biology is entering a new phase beyond the original breakthrough of AlphaFold 2, with two emerging frontiers poised to redefine the field: the prediction of full protein conformational landscapes and the routine de novo design of high-affinity protein binders. New generative approaches aim to approximate Boltzmann-weighted ensembles at a fraction of the cost of molecular dynamics, enabling direct estimation of free energy differences and state populations. In parallel, AI-guided design platforms are transforming protein binder development into a scalable engineering discipline with high experimental success rates. While the Critical Assessment of Structure Prediction (CASP) pushes for conformational landscape prediction among other frontiers, experiment-based contests ensure that protein design keeps evolving. Coupled with increasingly accessible deployment platforms and increasingly powerful general reasoning AI agents, these advances suggest that dynamic ensemble prediction and programmable protein design may constitute two “AlphaFold moments” now ongoing in biology.