What if medicine discovery became a loop between artificial intelligence, simulation and real experiments?

The prize is not more artificial intelligence-generated molecules. It's getting more promising candidates into rigorous experiments sooner.

Finding a medicine often requires searching through an enormous number of possible molecules, then spending years discovering which ones are promising in living systems. Artificial intelligence and simulation could help narrow that search, while automated experiments return real results that improve the next round of predictions. That creates the possibility of a faster cycle between imagining a treatment and testing whether it might work. Nothing removes the need to demonstrate safety and benefit in patients, but the early stages of discovery could look very different.