Jumper discusses the longstanding challenge of protein structure prediction and how AlphaFold’s deep learning approach achieved unprecedented accuracy, outperforming existing methods. He highlights the research breakthroughs that enabled AlphaFold’s success, its open-source release, and the global impact on structural biology. The system not only accelerated research but also revealed unexpected capabilities, such as modeling protein interactions and aiding drug design. Jumper concludes with reflections on AI’s expanding role in scientific innovation and the future potential of general AI systems in advancing discovery.