The article "AlphaFold two years on: validation and impact" by Oleg Kovalevskiy, Juan Mateos-Garcia, and Kathryn Tunyasuvunakool discusses the widespread adoption and impact of AlphaFold2 in structural biology. Since its initial release in 2021, AlphaFold2 has been used to accelerate experimental structure determination, enable new computational studies, and build new tools and workflows. The article highlights how AlphaFold2 has improved molecular replacement techniques in X-ray crystallography and cryo electron microscopy (cryo-EM), and how it has been integrated into various software suites for macromolecular crystallography. It also explores AlphaFold2's role in predicting protein-protein interactions and its applications in protein design. The article reviews recent studies that compare AlphaFold2 predictions with experimental results, emphasizing the importance of confidence metrics in interpreting predictions. Finally, it discusses ongoing efforts to validate AlphaFold2 and its limitations, including the need to model protein-DNA and RNA complexes, predict all functional states of proteins, and improve predictions for antigen-antibody interactions and orphan proteins.The article "AlphaFold two years on: validation and impact" by Oleg Kovalevskiy, Juan Mateos-Garcia, and Kathryn Tunyasuvunakool discusses the widespread adoption and impact of AlphaFold2 in structural biology. Since its initial release in 2021, AlphaFold2 has been used to accelerate experimental structure determination, enable new computational studies, and build new tools and workflows. The article highlights how AlphaFold2 has improved molecular replacement techniques in X-ray crystallography and cryo electron microscopy (cryo-EM), and how it has been integrated into various software suites for macromolecular crystallography. It also explores AlphaFold2's role in predicting protein-protein interactions and its applications in protein design. The article reviews recent studies that compare AlphaFold2 predictions with experimental results, emphasizing the importance of confidence metrics in interpreting predictions. Finally, it discusses ongoing efforts to validate AlphaFold2 and its limitations, including the need to model protein-DNA and RNA complexes, predict all functional states of proteins, and improve predictions for antigen-antibody interactions and orphan proteins.