Five years after AlphaFold 2 transformed protein structure prediction, co‑developer John Jumper reflected on its scientific impact and future directions. DeepMind’s AlphaFold family—now including Multimer and AlphaFold 3—has predicted structures for roughly 200 million proteins and reshaped workflows in structural biology, drug discovery and protein engineering. Jumper cautioned that AlphaFold’s database remains predictive, not definitive, and emphasized continued work on multimers, binding interactions, and integrating experimental uncertainty. Researchers and industry partners will continue combining AlphaFold outputs with lab validation and downstream functional assays to accelerate target identification and molecule design.