I know the speaker mentioned that we can currently use computational methods to reduce the need for experiments, but I think just researching and creating new materials purely using computation shouldn’t be a norm. No matter what, physics is an empirical science, so theory and computation should match experimental results.